Final Report Summary - FLEXICAST (Robust, and FLEXible CAST iron manufacturing)
Executive Summary:
FLEXICAST (www.flexicast-euproject.com) is a four-year research & development collaborative project of the European Commission's Seventh Framework Programme (FoF.NMP.2012-7-Innovative technologies for casting, material removing and forming processes). FLEXICAST started on 1 Nov 2012 and has ended 31 Oct 2016. The project presented knowledge-based technologies with the aim to follow the way to transform the conventional (batch-by-batch) foundry process into a flexible (mould-by-mould) process.
The challenges that FLEXICAST project had successfully addressed to cast iron manufacturing processes can be split into three large blocks:
1- Get a high quality cast iron melt, which means:
• Melting at the minimum temperature and moulds filling at the maximum possible temperature.
• Keeping the temperatures as constant as possible
• Obtain the best quality of the cast iron melt, according to the type of cast iron (grey or nodular)
2- Get a maximum quality of casted iron pieces, which means:
• Starting from a high quality cast iron melt and a pre-established pieces geometry, the aim is to fill moulds and to solidify / cool the pieces in such a way that the obtained pieces should contain the minimum possible defects (from all points of view: internal sanity, mechanical performance, minimum deformations, etc.).
3- Get an automatic casting process based on robotics for foundry processes. This is the area in which there was a dire need for development and in which most SME’s are interested, which means:
• Ensuring job security and employment levels.
• Raising the technological base and take an important technological leap forward, particularly for industries based on conventional technologies (foundries).
• Leading the improvements needed in the workplace environment to a greater job skills based on knowledge and know how.
One of the most important outputs is a new melting shop, denominated "cast iron manufacturing cell". Conceptually, each component of cast iron manufacturing cell has been designed independently, following a development methodology, the so-called "step by step". The cell is basically composed by a hybrid heating system (induction and plasma), an automated pouring unit, a multiple inoculation system, a nodulising system (converter) as well as a network of sensors and instruments necessary to assess the quality of the cast iron. Cast iron manufacturing cell is a DEMO pilot plant in real industrial setting, which has allowed us to evaluate the progress of the project results in comparison with the state-of-art solutions, and we consider that it can be a starting point of a future lean foundry manufacturing system.
Another important result was the new knowledge and know-how acquired by all partners’ staff. Efforts have been put in the definition of a plan for internal and external exploitation of the project results. “Take-up” activities to promote the early or broad application of state-of-the-art cast iron manufacturing technologies have been implemented. These included the assessment, trial and validation of design, manufacturing and monitoring technologies, easier access to and transfer of best practices for the early use and exploitation of technologies. They have been targeted, in particular, SMEs in the cast iron foundry sector.
Project Context and Objectives:
FLEXICAST project has been crucial for the improvement of different cast iron processes and technologies in the production line, which led to significant energy and cost savings. These technologies are not only applicable to new cast iron foundry lines, but also are readily available to be retro-fitted to existing plants of cast iron.
In order to define in detail the objectives of the project, we have divided the project into 5 main pillars. These are:
1-New cast iron manufacturing cell (WP1/WP2)
2-Artificial Intelligence-based control system (WP1/WP3/WP5)
3-Automation based on robotics for foundry process improvement (WP1/WP4)
4-Demonstration (pilot plant or prototypes-WP6)
5-Life Cycle Assessment, Exploitation plan, Dissemination, and Socio-economic aspects (WP7/WP8)
Regarding the first pillar, the widespread adoption of new melting shop as an operating process was, in itself, fostering the creation of even more powerful induction-plasma power supplies, versatile melt control technology, high-power density furnaces, automated charging systems, temperature control systems, nodulizing systems (magnesium treatment), dynamic and multi-inoculation systems, and automated pouring systems. The ability to use various smaller furnaces for comparable productivity, faster melting, easier and more reliable chemistry adjustments, less oxidation of the melt, less manpower and much improved charging safety was significant.
The cell production was a model for workplace design, and has become an integral part of lean manufacturing systems.
Given the complexity to develop the cast iron manufacturing cell, it has been considered essential to develop each component independently. The cell, from a conceptual point of view, was basically composed of a hybrid heating system (induction and plasma), a pouring unit, a multiple inoculation system, as well as the sensors and instruments necessary to assess the quality of the cast iron.
The second pillar constituted the “smart control” in this project. The objective is to develop a software platform. The implementation of this platform integrates: Data gathering and Statistical Process Control Systems, Artificial Intelligence Systems (decision-making tools) and Sensors/Devices (among others: melting and pouring temperatures, chemical composition, DTA parameters, X-ray on-line inspection approach, sand parameters, quantities and velocities of inoculant agents), in order to control the new cast iron manufacturing cell.
Another main objective of this pillar was the integration of soft computing and hard computing to predict production performance in specific manufacturing processes. The experimental validation was included. The ultimate goal of process modelling was to predict the final mechanical properties. The mechanical properties (hardness, tensile strength, yield strength, and elongation) of cast iron pieces were functions of composition and microstructure. We have planned to use soft computing for implementing computationally intelligent and user-friendly features that could not be realized competitively by hard computing alone. It has been demonstrated that a high degree of interaction between soft and hard implementation economy of intelligent computing lead to better systems, improved tolerance against incomplete or uncertain sensor data, enhanced learning and adaptation capabilities.
The third pillar was related with the automation based on robotics for foundry process improvement.
The aim was the improvement of automation level of cast iron foundry through the introduction of robotized cell. The main paradigms for doing this were flexibility and easy intuitive robot programming. The basic steps were: Automation test samples cell; Multi part moulding box configuration cell, Innovative gripper for grinding and deburring for robotized solution in foundry environment, Inaccuracies compensation methods for easy grinding root programming and Control solution for grinding and deburring robots.
As already declared in the last paragraph, a first scenario has been defined in order to get more efficiency at the FLEXICAST strategic roadmap. The “moulding box configuration cell” has the aims automatize moulding box configuration. The partners have agreed that the efforts can be split in two main activities:
a) Design of the Robotized Cell For Moulding Box Configuration;
b) Development of the Robotized Cell Prototype for Moulding Box Configuration.
The technological objectives of these activities were: easy/intuitive operation programming, insertion and positioning of the cores, insertion and positioning of the filters; blowing the sprues and marking each sample with a unique serial number for quality tracking.
They were described as follows:
- 1 Requirements and line layout where the robotic cell was placed. In addition, description of the main design choices in terms of robots number and allocation tasks.
- 2 Requirements of the multi-function gripper, conceived and designed, in order to address all the possible applications.
- 3 Referencing the robots with respect to the moulding box carrousel in order to allow an accurate execution of the tasks.
- 4 Description of an innovative device developed in order to allow an easy and intuitive robot programming.
- 5 Requirements of the innovative device developed in order to check the correctness of the cores/filters placing.
-
The main goal of the fourth pillar was to demonstrate a clear breakthrough using project development in comparison with the state-of-art solutions. The main task was to construct the final demonstrators according to the specifications. It demonstrated that the hybrid manufacturing strategy and concepts were applicable to the range of processes operated by the industrial consortium partners, minimum in one open access pilot facilities. It was proposed to design and to manufacture a cast iron manufacturing cell into full-scale demonstrators (industrial). Also, it was managed the various software platform and robotic cells, being their effective performances fundamental for achieving the full commercial exploitation in the market place.
Finally, the main goal of the fifth pillar was to save and formalize the knowledge and experience accumulated (know how) through the progress of the project. This was done by a) capitalizing knowledge generated by design activity and the experience returns from using pilot applications.
b) Formalizing results of the project in guidelines and possible standards, contributing to building of the European standardization in this field. It included the following tasks: Experimental test benches and procedures. Standardization approach, best practices rules and safety rules. On the other hand, life cycle assessment has been treated in order to have ways to measure the efficiency of the environmental aspects on a number of selected processed pieces.
Another main objective of this pillar was to set up the successful exploitation of the results during the project lifetime and, in particular, after the project has concluded. The exploitation plan contained details about the market and research trends and opportunities for new areas for development and applications of the Flexicast environment.
It covered exploitation strategies for industry (including SMEs as well as large companies), research centres and universities, as well as synergies with other research projects.
In relation with dissemination activities, the continuous scientific work of all partners did lead to numerous scientific publication and media coverage. The publications covered scientific topics from metallurgic & mechanical engineering to robotics applications.
1. Relevant results publications of the research activity carried out in peer-review journal
2. Publication of relevant results coming from the research activity in Conference proceedings
3. Oral and Poster presentations of relevant results of the research activity in Int. Congresses
4. Patents
5. Workshops related with FoF:
a. Impact of the Factories of the Future PPP of relevant results of the research activity in EU projects. (4 editions)
b. Let's 2014 (1 edition)
6. Attendance to Industrial oriented Fairs either as invited guess or expositor
7. Publication articles in different newspapers or news into web-sites
A consortium of 14 partners from 5 countries has carried out the FLEXICAST project. These were:
UPC (Labson), Fundiciones de Roda, Eurecat, Cimne, Ondarlan, and Tecnalia from Spain,
CNR (Imamoter and Itia), Proservice, ARC, TSS Modena, and Comau from Italy,
Foseco from France,
InnospeXion from Denmark and
Poznan University Tech from Poland.
The total eligible budget was 9.174.941€, of which 5.700.000€ was EU contribution (62%) and 1.050 person months were dedicated.
Project Results:
In the following pages, the list of main achievements of the project is reported with a brief description of main activities. Further details are reported in the referenced deliverable documents. Deliverables and milestone description are also reported.
WP1: Specifications and Technical Coordination
WP leader: UPC-LABSON
MAIN TASKS:
This work package constituted the first step of the system engineering where the goal was to put in common, and cross partners' knowledge and expectative. It included the following tasks
1- Partner needs inventory, consisting in interviewing partners to know and formalise their needs concerning cast iron manufacturing processes.
2- Pooling the state of the art on sensors, data transmission, prognosis and proactive methods, guidelines and tools to support product and process quality management in specific manufacturing processes
3- Elaborating a common terminology, defining a common conceptual and technical terminology, in order to prepare the design and the implementation of innovative technologies for the demo applications and to prepare the knowledge feedback database.
4- Defining the application scenarios, consisting in analysing the requirements of demo applications and elaborating the technologies for each scenario.
5-Technical coordination. Given that FLEXICAST was a DEMO project and the aim was to demonstrate the validity of the project results at an industrial level, a lot of essential efforts in coordinating and monitoring the scientific progress (WP2, WP3, WP4, WP5) has been considered. It was necessary the corresponding technological implementation in an open access plant facilities. Obviously, this responsibility fell mainly on Fundiciones de Roda (FR) as the partner who had this possibility (cast iron manufacturing facilities).
One of the most important outputs of this project demonstrated that the research results and developments achieved in FLEXICAST project could be translated to old and/or new foundry industry (in general). The chosen way used was through “layout/scenario” concept. Scenario is defined as a combination of different units/developments in a “virtual model of foundry”. The first reference model of foundry was the actual FR layout. FR is a typical small–medium foundry, with all typical: departments, solutions and difficulties, which may be expected to found in a representative foundry. Current FR layout has been considered as “typical”, due to the following reasons:
- Is a typical medium-small foundry, well organized but with the similar characteristics and possible implementation problems as any other foundry.
- Use the layout of other real foundry, is not possible due to the difficulty to known exactly how it is (and no foundry is going to tell his “secrets”.
- To “invent” and ideal foundry layout is not applicable because either the difficulties of implementation are going to be already solved or artificial ones are going to be implemented.
Nine layouts/scenarios have been prepared incorporating different possible combinations.
Definition: "Reference scenario" is that foundry were none results of Flexicast has been implemented, and with a zero profit.
The objective of this definition was to make it easier to consider the impact of the project achievements when they were implemented in different scenarios (figure nº 1). For each scenario, the advantages, inconvenient and opportunity of incorporating each improvement have been described .
Figure nº 1.- Different scenarios analysed (combined results)
Figure nº 2.- Flow diagram of cast iron production line, from raw material to finished pieces
WP2: New cast iron manufacturing cell
WP Leader: F de RODA
MAIN TASKS:
WP2 contained five tasks. It has been focused on R&D activities related with the development of a new cast iron manufacturing cell.
1- Melting process (induction and plasma heating). The goal was to explore and demonstrate that a hybrid combination of different technologies can lead to an overall improvement of the efficiency, provides high heat transfer density and finally, improves cast iron quality.
2- New pouring system (high performance and low resources consuming). The goal was to install the melting system closer to the automated pouring system (heated by plasma), keeping as close as possible to the mould carrousel. Melt treatment (nodulizing) have been done into the furnace unit or the pouring unit or both of them. Successful production of high quality ductile iron depends on the successful control of magnesium. The incorporation of a system for heating and controlling melt temperature based on a plasma torch has been probably one of the most significant innovations at casting unit level that appeared in the last twenty years.
3- Nodulizing process methods. Successful production of high quality ductile iron depends on the successful control of magnesium. Melt treatment with Mg (nodulizing) have been done into the pouring unit through three alternatives: Mg vapour treatment, Initek process (converter) or wire core system. The innovative goal was to design a new nodulizing procedure based on magnesium vapour produced in a small external furnace.
4- Inoculation systems. With regard to inoculation (most appropriate type(s) and size(s) of inoculant(s), addition rates, and inoculation chain...) different options have been analysed. An inoculant is a material, added to the liquid iron just prior to moulding process, which provides a suitable phase for nucleation of graphite during the subsequent cooling. A breakthrough has been based on the multiple-dynamic inoculation. In this case, “multiple” means a combination of elements from inoculants battery, and “dynamic” means that a stream of inoculant(s) is added at a variable rate according to the melt flow.
Highlights of most significant results:
1- New manufacturing cell:
-Induction furnace. ONDARLAN, UPC
-Pouring unit (v1.0 and v2.0) F. de RODA, UPC
-Plasma heating TECNALIA
2- Analysis of lining wear in the pouring unit F de RODA, UPC
3- Measurement of melt jet temperature in pouring unit. F de RODA, PROSERVICE
4- Evaluation of heat losses in pouring unit in order to estimate the power supply needs of the plasma heating system. F. de RODA, UPC. CIMNE
5- Experimental evaluation of magnesium fading in induction furnace and in a pouring unit.
TECNALIA, F de RODA, UPC, ONDARLAN
6- Adaptation of the plasma prototype to the pilot scale. TECNALIA, F de RODA, UPC
7- Innovative prototype for nodulizing process with magnesium vapour. TECNALIA.
One of the most important results of this WP was a full characterization of plasma torch prototypes and the main outputs were the construction of the prototypes and, as a result of the trials, conclusions of energy efficiency and process working conditions, applied on the final prototype designed.
Comparing with standard heating systems, the hybrid induction-plasma and plasma heating technology offered:
• Power consumption can be reduced up to 35% depending on material and casting conditions. Average improvement value of 5-15%.
• Metal characteristics achieved were considered at an industrial acceptance level.
• Mechanical properties of produced material were at the same level than the ones melted using only induction heating.
The overview assessment of the plasma heating technology had the following conclusions:
• Allowed to heat only a desired portion of molten metal which is going to be cast.
• Simple design. Easy operational work.
• Accurate casting temperature regulation.
• Improvement of energy efficiency due to more efficient heat transfer to the melt
Finally, anodic inoculation trials at pilot plant have been carried out. Assessment of the effect of the quality of the crystalline graphite on the molten metal quality measured with DTA technology.
Most significant and measurable results were:
• Improvement of the metallurgical quality through DTA analysis.
• Electric ideal conditions: Power 20 to 70 kW. Intensity form 400 to 480 A.
• Physical ideal conditions: HP electrodes submerged at least 50 mm. Optimal 80 mm.
• Savings in inoculant materials when process adjusted at pilot plant trials.
• Anodic inoculation system. See results in the deliverable.
Main conclusions of the assessment and design of anodic inoculation process and equipment were:
• Feasibility tests of anodic inoculation successful.
• Ideal electrodes to be used are HP quality. For FLEXICAST selected industrial site 50 to 60 mm of diameter.
• Design of anodic inoculation device includes HPTP plasma torch equipment.
• Improved quality of the molten metal obtained with the system.
WP3: Artificial Intelligence-based control system (AIBCS)
WP Leader: PUT
MAIN TASKS
The metal casting process requires testing equipment that, jointly with properly customized computer software, supports the casting component analysis (characteristic properties). Due to the fact that this evaluation process involves the control of complex and multi-variable melting, casting and solidification factors, it has been necessary to develop more specific software. The AIBCS Platform was designed to support cast iron manufacturing processes, faulty components and unexpectedly higher scrap rates, even chemical analysis, pouring temperatures and so on.
WP3 and WP5 have constituted the “smart control and knowledge platform” in the project.
This platform has been integrated by: data gathering and statistical process control systems, artificial intelligence systems (decision-making tools), sensors/devices (among others: melting and pouring temperatures, chemical composition, DTA parameters, X-ray on-line inspection approach, sand parameters, quantities and flow rate of inoculant agents) and, soft & hard computing methodologies.
Figure nº 3.-Integrated AI-based Control Systems and partners involved
WP3 contained four tasks:
1- Real-Time Data Acquisition System (RTDAS). The goal was to monitor the cast iron manufacturing unit through the simultaneous measurement and recording of melt temperature in different locations and environmental conditions around the melt such as metallostatic head pressure, flow rates, inert gas pressure, sand parameters (moisture, permeability and compactness, etc.).
The data repository included, among others, thermal analysis parameters, x-ray imaging, environment conditions, process settings and melts chemical composition.
The time-temperature data, as well as the input and output values of the environmental conditions have been automatically saved and time stamped for traceability in order to make it feasible to generate graphical plots of various melt samples.
The assessment of the RTDAS data in the AIBCS in terms of accuracy have been determined by the deviation of measured values from sensors against the precise value (average value).
One of the first challenges that have been faced was the development of an electronic system that could be interfaced with different kinds of sensors (thermocouples, humidity sensors, pressure sensors, etc...). These sensors had a completely different output characteristic as well as the signal conditioning, so the unit should provide a reconfigurable analog front-end to support these sensors. Another required feature for the acquisition system was the capability of managing an acquisition system through a wireless communication, and monitoring the data in real-time.
Performing a reliable wireless communication in a foundry environment has been a challenging task, because there were several factors that interfered with the signal propagation, reducing the communication range. The technologies evaluated were the standard WiFi IEEE 802.11b due to its large diffusion and the potential integration with existing IT infrastructure and the IEEE 802.15.4 ZigBee modules. The results with ZigBee performed in Fundiciones de Roda proved an excellent performance.
2- Statistical process Control System (SPCS). The goal was to assess the AIBCS outcomes in terms of accuracy, reliability and timeliness. It was based on probability and statistics analysis. Once the sensor signals have been recorded by the RTDAS, they have been statistically analysed in the SPCS module. The SPCS monitors, measures, and evaluates melt quality based on pre-defined criteria.
Casting processes, considering data acquisition methodology proposed by FLEXICAST partners (in particular: PUT, PROSERVICE, FOSECO, Fund. de RODA and EURECAT), have been divided chronologically to casting related sub-processes:
• build the mould,
• pouring the metal into the mould,
• melting and metallurgical evaluation of the cast iron,
• final check of the casting quality, according customer acceptance conditions.
The following parameters have been recorded:
• Time of tapping from furnace to pouring ladle,
• Initial and final temperature of pouring the metal into mould [°C],
• Number of sample for chemical analysis,
• Time of making the test using spectrometer,
• Quantity of alloying elements: C, Si, Mn, P, S, Cr, Ni, Al, Cu,
• Equivalent carbon value CE = %C+1/3*(%Si+%P),
• Time of measurement of moulding material parameters,
• Moulding material parameters: green sand strength RW, permeability
• Humidity, bulk density, compactness, etc.
• Defects percentage [%].
An innovative aspect of the SPCS was the assistance roll in identifying abnormalities in the melt, improving castings quality through an “intelligent” system which can predict the casting properties and can provide recommendations and actions to the operator for process improvements.
PUT and F. de RODA partners did the core of this activity. The details of the work were described in the periodic reports and deliverable D3.3 -Statistical Process Control System and some published technical paper.
Figure nº 4.- AIBCS main screen (designed by PUT )
3- Case-based and Fuzzy Inductive Reasoning: One of the biggest challenges in foundry industry is the castings production according with a customer predefined quality. Possible casting defects, at different intensity levels must be referred to these predefined conditions. Procedures of analysis of defects and its causes require using software tools. Methods from the area of soft computing have been proven, being more and more useful for solving these matters. Examples of these methods are decision trees (DT) and artificial neural networks (ANNs), which can realize such analysis in a more effective way. En general, this is an approach named Data Mining.
Demanded goal function was the quality classification of final product, which qualifies the casting to one of two classes: conform to the final product or non-conform to the final product. Disposing a full range of available information (data) about technological parameters of the process (input variables) and information about the amount of faulty products regarding to selected assortment and date of production (output variables), it was possible to build a model based on that collection of teaching data.
The first approach was to develop and implement a Fuzzy Inductive Reasoning algorithm (FIR). Nevertheless first experimental tests showed that FIR was not the most suitable tool due to the large number of variables of the process and the poor time dependency. For this reason, the strategy was changed towards a machine learning approach.
A comparison of several machine-learning algorithms including Support Vector Machine, K-Nearest Neighbours and Linear Discriminant Analysis, was performed. The experimental results combining the foundry raw data with the state of the art algorithm showed again a poor performance. It was due to the large number of variables that introduce undesired noise to the system and the unbalanced class distribution.
It was important to note that all the research and development described in the periodic reports was based on real production data not designed or controlled experiments. This fact increased the difficulty of the development and off course reduce the accuracy of the systems, but allowed to achieve much more powerful, robust and real-life ready solutions.
As we have seen in the previous paragraph, applying state of the art algorithms to the raw data do not produce satisfactory prediction rates. Developing a new algorithm able to pre-process the raw data and looking for the most relevant parameters became crucial. Nevertheless due to the large number of variables, a brute force approach was not feasible. The main challenges were:
•The large number of variables degrades the performance of the machine learning algorithms, by adding noise to the system.
•Unbalanced classes: much more samples without the defect than with it.
•An optimal set of input features (if exists) must be found, but we are facing a combinational explosion. For instance, there are 2.4E+23 possible ways to pick 25 variables of a set of 100. At 1ms per combination, that would take us 7 billion years.
In order to overcome the previous challenges, we have developed the Indiana Module which aims to find a sub-optimal combination in an affordable time-frame. Facing the combinational explosion:
-With the aim to create a feature selection algorithm for pre-processing the raw data, a novel approach was considered, based on subset partitioning and feature ranking.
-The Indiana module analyses different combinations of input parameters with the aim to find a sub-optimal set of input parameters (in an affordable timeframe) that boost the performance of the classifier.
-A genetic algorithm refines the best model based on the top ranked features found in the previous stage.
The Indiana Module follows the following steps to reach a sub-optimal solution:
• Creation of sub-models by partitioning the input features space in order to rank the input features
• Pick the n-better ranked features
• Apply a genetic algorithm to obtain the best model
Indiana Module main characteristics and advantages:
• Feature selection algorithm
• Sub-optimal best model discovery by means of genetic algorithm
• Multiclass capability
• Multi-Classifier compatibility
• Automatic-weight distribution for inter-class balance
Figure nº 5. Pre-processing strategy to boost the classifier’s performance
The prototype has been developed in "Python" and did not require the installation of any other library. In the last years, python has been quite popular thanks to its growing community of developers, focusing on Data Analytics and mathematical computation.
The experimental results have shown that, thanks to pre-processing, the raw data with the Indiana module, the performance of the machine learning classifiers was boosted, increasing their accuracy by at least 50%. Overall, combining Indiana module with a Support Vector Machine offered a better compromise between accuracy and flexibility to adapt to the different iron types and defects.
Figure nº 6.-Three Dimensional projections of the raw data versus the pre-processed data for the grey iron porosity prediction. The axis corresponds to the 3 most relevant latent scores of a Partial Least Regression
Thanks to the algorithm design, the Indiana Module not only provides to the classifier a reduced set of input features, but also ranks all the features. This information can be very helpful for the foundries in order to identify those process parameters that are relevant to predict the presence of some defects. The delivered solution has been proven to offer an excellent performance, reaching more 85-90% of accuracy when predicting shrinkage and porosity for both grey and laminar iron, and improving by a large margin the state of the art solutions.
4- Holistic enabling technologies for cast iron quality validation: In this section, Differential thermal analysis (DTA) and X-ray approach for internal inspection were included.
In relation with DTA, the main goal has been the realization of a new system that overcomes all the other equivalent commercial products and, in addition, was suitable for the new foundry approach defined in the Flexicast project: shift of production process logic from batch-by-batch to mould-by-mould approach. Two independent thermal analysis systems have been installed: one focused on the preparation of base iron in melting furnaces and one focused on the final iron in the pouring station, in order to make possible the control and the study of the melt along the whole production process.
On the DTA systems has been implemented some algorithms that evaluates all these three factors and that raise focused alarm messages if some problems are found. They required a high effort due to intrinsic difficulties on the signal elaboration.
The main innovation was related with the results interpretation approach: the clock gauges helped to the operator about the result suitability through the colour areas (green/yellow/red).
As regards the final iron analysis, a different interface (and in fact an almost complete different system) has been developed. On this production process phase, the forecast of metallurgical defect tendency become one of the most important features. The interface developed was similar to that one developed for base iron (it shows the cooling curve and the most important results) but on the bottom part a series of clock gauges informs the workers about the current metallurgical status, relieving the workers of the results interpretation operation.
The defect indicators were:
• Cementite: probability of carbide formation on the casting surface;
• InverseChill: probability of carbide formation in the casting core;
• Porosity: probability of internal cavity formation due to thermal conductivity issues;
• Macroshrinkage: probability of internal cavity formation due to uncompensated liquid shrinkage;
• Flotation of graphite: probability of graphite segregation in the casting.
The other indicators provide some other important information about the iron properties:
• HypoEutecticHyper: it indicates the position of iron in the Fe-C state equilibrium diagram;
• Graphite expansion: it indicates the level of the graphite formed during the solidification;
• Nodulizing: it shows if the iron is nodular or grey.
One of the most important milestones of this WP was that DTA had to interpret the results in function of the boundary conditions (i.e. the specific casting features, like thickness, thermal module and so on). New module "casting designer" has been implemented to achieve this result.
Figure nº 7.- DTA display
Finally, in order to make an easier usage of DTA system for the foundry workers, three remote displays have been implemented and installed.
In relation with internal sanity, the first stage of know-how may be, of course, to consider a simultaneous specification of such defects generated during the metallurgical manufacturing processes, such as: gas bubbles, porosities, inclusions of foreign inclusions, segregation of the components, stratification, incomplete fusions, surface and internal cracks.
X-ray technology was highly requested by manufacturers (and also by the end users), for defects revelation (Internal sanity). In this sense, the consortium wanted to highlight InnospeXion's efforts relative to the development, design, construction and testing of a novel X-ray system for cast iron quality control. Flexicast consortium would like to emphasize that this equipment took in consideration the ASTM E1030-05 (2011) requirements. The objectives of the task were:
• Radiation cabinet design
• X-ray source acquisition and tests
• Optimize imaging technology (control of scattered radiation, etc.)
• Development of manipulation system (with FR), incl. manipulation system drawings and description to FR for acceptance
• Develop counting detector technology – interfacing & software development – for micro porosity determination & quantification
• PC configuration
• Software development for control (viewing software only, no automatic interpretation)
• Cabinet construction
• Installation of hardware, manipulation system, test of control and motion system etc.
• Testing of samples (imaging tests)
• Testing of samples (radiation measurement tests, porosity determination)
In Flexicast project, the focus was on iron cast samples with a thickness (in the X-ray imaging direction) up to 50 mm. The samples had different geometries, from planar to complex shapes (figure nº 8). Also, the samples used in this project for demonstration purposes were: railway guides (planar), control valve; seal ring and pump housing (complex). In addition to the cast samples, there was also a need of inspecting the “cores”, e.g. the sand – epoxy precursors for the cast part. These “cores” were crucial to inspect since their integrity determines the presence of major defects in the cast parts.
Figure nº 8.- X-ray castings inspection
Flexicast consortium partners have shown that a high degree of interaction between soft and hard implementation economy of intelligent computing lead to better systems, improved tolerance against incomplete or uncertain sensor data, enhanced learning and adaptation capabilities. See technical paper D8.6-Disseminationa activities.
Figure nº 9.- Temporal evolution and Normal distribution of porosity (%) and Rm of GJS500-7
for annual periods (2012-2015)
The AIBCS outcomes (from 2014 to 2015) were compared to the traditional process parameters commonly used for analysis, control and prediction of characteristics in the casting process (before 2012). The qualitative results shown that the Integrated AIBCS has a high future potential value.
There was no doubt that figure nº 9 demonstrated improve results achievement in FLEXICAST project. The incorporation of new methodologies and advanced technological developments were successful. Properly, the integration of the different instruments made this AIBCS a successful methodology from holistic point of view.
WP4: Automation based on robotics
WP Leader: CNR-ITIA
MAIN TASKS
WP4 contained two basic tasks. It was focused on the automation based on robotics for foundry process improvement. It included the following tasks:
1-Develop the affordable systems having the versatility to perform an automatic test samples cell and a multi-part moulding box configuration cell.
2-Develop a deburring robotized cell, which almost totally eliminate the problems in fettling and finishing of castings made in batch quantities.
A huge number of automatized solutions for each phase of cast iron production processes were present on the market, but, although this variety, few of these devices were installed in the real plant, mainly for only one reason: the low flexibility of these devices. In particular, for its programming procedure: too long and complex, that makes them feasible for large batch pieces production. Usually these machines presented very high productivity but with long reconfiguration time.
On the contrary medium cast iron foundry, presented high variability of pieces type in their daily production that require high and fast configurability of the plant. For this reason, machines with high productivity but with low flexibility, could not represent a suitable solution for medium dimensions cast iron industry. The design of robotized cell, has considered flexibility as a guideline for FLEXICAST project.
The first action was a preliminary analysis of the actual state-of-art of automation in the cast iron foundry. The second step, in charge of CNR, was the design of new robotized solutions in order to overcome the limit of actual adopted. The main paradigms for doing this were flexibility and easy intuitive robot programming. The use of industrial robots not only for standard operations but also for machining applications have been required a deep analysis of robot structure and radical improvement of the standard control unit. The WP4 has been structured in 5 main items:
1- Automation of test samples cell (COMAU);
2- Multi-part moulding box configuration cell (CNR-ITIA);
3- Innovative gripper for grinding and deburring for robotized solution in foundry environment (CNR-IMAMOTER);
4- Methods for
(i) Easy deburring and grinding programming and
(ii) Part-programme compensation of the mould production inaccuracies (CNR-ITIA);
5- Control solutions for grinding and deburring robot (CNR-ITIA).
Cell #1: TEST SAMPLE CELL
The detailed description of the layout of the Cell #1 was reported in D4.3.
The complete cycle consist on a sequence of pick and place of the mushroom sample and different stations that perform the different technological phases.
The list of tasks was reported below for sake of clarity:
- Cooling the sample with the core
- Braking the core
- Cooling the sample without the core
- Sanding (light polishing)
- Cutting the feeder
- Polishing the medal (one side, three stages)
- Insertion of the part in the Spectrometer/ Microscopy analysis.
Cell #2: MOULDING BOX CELL
The R&D work has been focused on two layers:
- To develop an easy and intuitive device for the lead-through-programming that is robust and usable in the foundry environment
- To develop the whole programming framework for the moulding box configuration (figure nº 10).
The tool has been calibrated by exploiting the standard COMAU programming modalities. The accuracy of the modality (+/- 0.3mm) grants the task requirements. For a detailed description of the cell refers to D4.4
Cell #3: GRINDING AND DEBURRING CELL
For a detailed description of the cell (figure nº 10) refers to D4.4 & D4.5
The programming phase realized can be divided into four main stages:
1. Measure of the Master Piece (e.g. piece manually deburred fixed in the desk);
a. Acquisition of areas for identification of the rigid roto-translation
b. Acquisition of areas for the identification of the deformation (afterwards the deburring trajectories will be defined in these areas)
2. Measure of the deburring trajectories in the Master Piece;
Acquisition of the deburring path
3. Measure of each Slaves Sample;
a. Acquisition of areas for identification of the rigid roto-translation.
b. Acquisition of areas for the identification of the deformation
4. Deformation of the deburring trajectories for each Slave Sample.
Execution of the deburring path on the Slave Sample taking into account the roto-translation and the deformation
Figure nº 10.- Flexicast robotic cells (#2 and #3)
The results obtained from the kinematic calibration procedure described in D4.5 have been evaluated using an optical measurement system; in this case a NDI Polaris Spectra. This instrument allows the tracking of a passive marker with accuracy up to 0.35 mm of RMS at the maximum distance in the workspace.
WP5: Simulation and optimization techniques
WP Leader: CIMNE/PUT
MAIN TASKS:
WP5 contained two tasks, focused on the development of autonomous computational intelligent systems. The prime objective was the integration of soft and hard computing to predict production performance in specific manufacturing processes. The second one was the experimental validation.
1- Thermo-fluid dynamic modelling. This task has been split in two parts. One devoted to flow analysis in melting systems and the other was focused to mould filling and solidification processes.
The aim of the first part was to show the numerical simulations and results carried out in flow analysis in melting systems related with cast iron manufacturing cell, in particular:
- Laminar barrier in furnace unit (figure nº 11)
- Pouring flow from furnace and running (tundish) unit
Figure nº 11: Gas flow for laminar barrier
A common problem, in most furnaces, was that metal melt inside the furnace reacts quickly with an open atmosphere. Elements in the casting like silicon, aluminium, chromium, etc. react with the oxygen forming slag. This slag can waste part of the casting by causing rejection of part of it. In order to avoid this particular problem, a laminar barrier of inert gas has been added. In this activity, a free Open-Foam code has been used. It is a C++ toolbox for the development of customized numerical solvers, and pre/post-processing utilities for the solution of continuum mechanics problems and also computational fluid dynamics (CFD).
As a main goal of this activity, 7-9 m/s gas flow was the optimal barrier configuration which protected the melt. Figure nº 11 shows the streamlines and the magnitude of the velocity.
Other challenge of the project was to estimate the flow rates with cast iron pouring units. Numerical simulations have been performed for both water and cast iron. Laboratory experimental tests with water were used to validate meshing’s and numerical schemes. The numerical results with cast iron have been used for the flow rate estimation in a real industrial process. Another alternative analysed was related with the experimental measurements carried out in a non-submerged vertical free liquid jet. Specific experimental assessment and quantification of the quality of the jet have been done (Jet flow rate measurement method). Then, a new technique to vertical free jets named Visual Photographic Technique (VPT) was proposed. In contrast to the PIV technique, VPT has a lack of the velocity profile/field measurements. However, in the present work, several mathematical models were proposed to allow VPT estimate the flow velocity and flow rate of a non-submerged vertical liquid jet. It is based on the DGHT (Dynamic Generalized Hough Transform) algorithm. This technique required following steps (see figure nº 12):
-Binarization of a grey-level image from a PiV measurement through a threshold operator
-Noise removal by size and salt-and-pepper operators
-Edge detection
-Determination of interpolation (connectivity) points to be used in the DGHT algorithm
-Computation of ellipse parameters through the DGHT comparison of reconstructed ellipse
Figure nº 13 shows a comparison between estimated and experimental results.
Figure nº 12.- DGHT algorithm steps
Figure nº 13.- Comparison between estimated and experimental results
The ultimate goal of process modelling was to predict the final mechanical properties. The mechanical properties (hardness, tensile strength, yield strength, and elongation) of cast iron pieces are functions of composition and microstructure. The graphite shape, graphite structure, graphite amount, carbide content, and matrix structure (pearlite, ferrite) all affect the mechanical properties of ductile iron castings.
Research activity carried out deal with the numerical simulation and experimental validation of the cast iron pieces filling and solidification process. The model pieces were two types: specific Flexicast reference pieces and simple geometrical pieces (like samples). This work (numerical and experimental tests) included:
A-Shrinkage test
Comparison of different HTC
Analysis of results of experimental-simulation tests
B- Castability of Cast Iron
Status quo based on cited references
Castability trails
Computer simulation of castability trials
C- Virtual mould (figures nº 15 and 16)
Casting of reference piece -2525
Description of simulations for versions of 2525 casting
Version 1 (classic sand risers)
Version 2 (2 exo-risers)
Destructive testing realized by Fundiciones de RODA
Preparation of simulation in Procast code.
Simulation Procast MICRO
Solidification results
Simulation Procast MICRO – micro model results
Additional validation tests
Figure nº 14.- Experimental and numerical results of simple geometrical pieces tests.
Figure nº 15.- Experimental and numerical results of specific Flexicast ref. 2525 DCV piece.
As an example, figure nº 15 shows the PROCAST micro model simulation (2525 DCV piece)
Figure nº 16.- Micro model results
These activities were proposed and realized in F de Roda facilities, in collaboration with PUT team, directed by Dr. Zenon Ignaszak.
Also, as planned in the DoW, all reference pieces have been experimentally characterized to assess their mechanical performance (fatigue, wear and deformation). TSS-Modena, F de RODA, CNR- Imamoter and UPC-Labson have carried out these works.
WP6: Demonstration activities. Prototypes
WP Leader: F de RODA
MAIN TASKS:
The main goal of WP6 was the final DEMO prototypes construction, according to WP2 and WP3 and WP4 requirements. It included the following main cells:
-New cast iron manufacturing cell.
-Robot cells for cast iron automation process
The specific activities related with DEMO were:
1-Assembly, installation and validation of cast iron manufacturing cell in F. Roda facilities:
-Induction furnace (and auxiliary equipment)
-Pouring units (and auxiliary equipment)
-Plasma equipment
-Multi-inoculation system
2-Assembly, installation and validation of the following equipment in F. Roda facilities:
-DTA systems and corresponding software.
-Robotized Cell Prototype for Moulding Box
3- Assembly, installation and validation of X-ray equipment and cabinet prototypes
4- Assembly, installation and validation of converter equipment prototype
5- Assembly, installation and validation of the cast iron manufacturing robotic cells: sample cell and deburring cell. These have been installed and validated in CNR facilities.
Figure nº 17 shows some DEMO prototypes related with FLEXICAST cast iron manufacturing cell in F de RODA facilities.
Related to Mg vapour prototype (figure nº 18), the tasks carried out were similar to the ones of the plasma torch and anodic inoculation device. All these activities have been carried out in Tecnalia facilities.
Trials have been successful, and feasibility of magnesium fading compensation in cast iron using magnesium vapour device is demonstrated at industrial relevant environment.
One important aspect of the plasma device operating was the availability for common maintenance actions without generating any operative trouble. When the plasma torch is at the working stage on the vessel allows the full tilting of the unit (Figure nº 19). The tilting of the unit can be due to maintenance operations (cleaning, emptying, etc.) or due to actions related to safety (rapid emptying, need of taking out the iron from the pouring vessel stopper area ...).
Related to the use of plasma-induction hybrid heating devices at foundry sector, these results indicated that a sensible reduction of energy was obtained when using a plasma torch as melting device (single or hybrid technique). Metal losses ratio using scraps improves respect to other melting technologies (due to inert atmosphere which reduce oxidations and the more fast and efficient heat transfer).
Figure nº 17.- Synthesis of the DEMO activities realized during 2015-16
Figure nº 18.- Mg vapour tests
Figure nº 19.- General view of plasma torch device mounted on the big tundish
In relation with the deburring robot cell (figure nº 10), setup was based on a COMAU C5G NJ220-Foundry industrial robot, a single point laser displacement sensor (LMI Gocator-1170) and a common pneumatic spindle. The three control strategies implemented were:
• Deburring maintaining the execution velocity constant
• Deburring slowing down the velocity according to the measured forces
• Deburring through the human mimetic control strategy
A more detailed analysis has addressed the evaluation of the accuracy of macro misalignment registration algorithm from a statistical point of view. In this case, a MWP was acquired through the laser scanner measurement device, thus the same WP was used as “fake” SWP in order to have two WPs equal from the geometrical point of view where the only difference is the location on the workbench. In this way it is possible to exclude the effect of local deformations. The SWP was evaluated in 30 different positions by acquiring, for each different position, the relative point cloud. In addition, 3 different deburring trajectories were defined, as straight lines, on the MWP.
Trajectories lie in 3 different planes of the WP and each plane is orthogonal w.r.t. the other ones. The acquisition of each single point of the deburring trajectories was done exploiting the laser point sensor instead of the deburring tool. This approach allows to a have a direct measure of the distance between the laser scanner measurement device frame and the WP surfaces. The measures acquired on MWP and on each SWP (figure nº 20), after the registration of macro misalignment, should provide the same distance between WP surface and laser scanner measurement device frame.
The real differences between laser scanner measurement device measures provide estimation on the “global accuracy” that can be reached by applying the macro misalignment registration algorithm. . Global accuracy is a quantity that takes into account the measuring system accuracy, the robot positioning accuracy, and the misalignment registration algorithm accuracy. From a technological point of view, the threshold of 1mm of RMSE, is evaluated as sufficient from expert workers.
Figure nº 20
Panels (A-B) for a real MWP and a real SWP placed in a significative different position.
Panels (C-D) for the raw point clouds of MWP and SWP simply overlayed.
Panels (E-F) for the SWP point cloud aligned on MWP point cloud.
A video of one of the experiments has been published online in the ITIA-CNR Youtube channel at:
https://www.youtube.com/watch?v=ciRWDrcLReQ.
WP7: Cast iron technology knowledge. Best practice rules
WP Leader: F de RODA
MAIN TASKS:
The goal of this WP7, which constituted a permanent support activity of/for the project, was to save and formalize knowledge and experience accumulated along the progress of the project. It included: Experimental test benches and procedures, pre-standardization approach, and best practices rules and safety. The main results were summarized in specific knowledge deliverables (D7.2 D7.3 D7.4 D7.5 and D7.6).
On the other hand, life cycle assessment was developed in order to have ways to measure the efficiency of the environmental aspects.
In order to perform the LCA analysis it was necessary to know, in detail, both the input and output data of the whole manufacturing process and relative environmental emissions (figure nº 21).
Figure nº 21.- Block diagram of the cast iron manufacturing process
The input and output data of materials and resources consumed (water and electricity) have been obtained in collaboration among IMAMOTER C.N.R. and partners of the FLEXICAST project. The meetings among the involved partners have been necessary to identify the useful data to collect in order to perform the LCA analysis of the whole manufacturing process. Then the data have been elaborated and implemented within the GaBi software.
A complete Life Cycle Assessment was developed to compare the impact, in term of Eco Indicators, of the solutions introduced (and developed) during the project. The LCA presented in these pages is probably the first example of a complete environmental impact analysis of the cast iron production cycle. Life cycle assessment was included in deliverable D7.7.
WP8: Exploitation Plan. Dissemination and training
WP Leader: EURECAT/CIMNE
MAIN TASKS:
It included the following tasks: (i) Exploitation plan, (ii) Socio-economic aspects and (iii) Dissemination and training activities.
The activities carried out under the scope of the FLEXICAST exploitation plan were:
1.-FLEXICAST exploitable results
2.-Specific Products (results) Portfolio
3.-Exploitation plan partner by partner
4.-Future Plan (Short term) for each exploitable result
5.-Exploitation plans by scenarios
6.-Exploitation plan of F. Roda (as a Demo partner)
Also, all Flexicast partners consortium met in Modena (TSS facilities) in order to assist a seminar tutored by Mr. Dario Mazella from ESIC (Brussels). It was a great help for all partners in order to fulfil the Flexicast exploitation plan. The Exploitation plan was summarized in D8.4.
The economic impact methodology chosen was to evaluate the possible impact in the European foundries who decided to incorporate the FLEXICAST achievements, and for that the impact in several “type” foundries incorporating several combinations of FLEXICAST achievements has been considered. (What here are presented as “scenarios”). Some major results are exposed in the coming pages.
In relation with dissemination activities, the continuous scientific work of all partners did lead to numerous scientific publication and media coverage. The publications covered scientific topics from metallurgic & mechanical engineering to robotics applications.
1. Relevant results publications of the research activity carried out in peer-review journal
(20 papers)
2. Publication of relevant results coming from the research activity in Conference proceedings
(9 publ.)
3. Oral/ Poster presentations of relevant results of the research activity in Intern. Congresses
(26 oral presentations)
4. Patents (1)
5. Workshops related with FoF:
a. Impact of the Factories of the Future PPP of relevant results of the research activity in EU projects. (4 editions)
b. Let's 2014 (1 edition)
6. Attendance to Industrial oriented Fairs either as invited guess or expositor
7. Publication articles in different newspapers or e-news web-sites
Also, all these dissemination and training activities are exposed in the coming pages.
WP9: Financial and administrative management
WP Leader: UPC-LABSON
MAIN TASKS:
The goals of the project management and coordination team were to:
• Facilitate communication and integration between the partners and disseminate information about the project to the wider community
• Identify and resolve disputes between partners
• Keep the project on track and ensure all deliverables, periodic and final reports are produced on time
• Set up a project office with finance and administration support
• Carry out annual audits of all partners
• Liaise with the EC
The deliverable D9.1 summarizes the major tasks undertaken in WP9 for the management and coordination of the project, together with a summary of the overall project management details in terms of deliverables, dissemination outcomes and key meetings
Communication with the European Commission
UPC -LABSON has been the intermediary for any communication between the EC and partners during this reporting period.
Administration of the financial contribution
The Community financial contribution was carefully managed regarding its allocation between partners and activities taken by the consortium. All payments were made without delay and records have been kept in order to determine at any time of what portion of the Community financing has been paid to each beneficiary. The Commission was informed of the distribution of the contribution. The cost claim process was carried out through the NEF system. The planning of resources has been done so as to accommodate a realistic distribution of resources, based on the on-going project situation.
Communication and Reporting
Communication within the consortium has been ensured through the normal means such as electronic mail, teleconferences and in-person meetings. Technical documentation generated by the Project have been exchanged in electronic format and made available to partners. To monitor the status of the project, internal reports have been prepared every 6-month period. A consortium mailing list service has been created.
Legal and administrative issues
The Project Management Team (PMT) has established the rules for the access and exploitation of the background of the individual partners and of foreground.
Organization of project meetings
UPC-LABSON team and Dr. Pedro Roquet as technical manager took care of the logistics and organization of all Flexicast project meetings, ensuring that all partners took part according to their role and responsibilities.
Potential Impact:
The foreseen socio-economically impacts that may mark trends in the next future, where FLEXICAST project has contributed to highlighting, have been described in the following paragraphs.
1.4.1 Towards a generation of high qualified staff.
After four years in the Flexicast project, it has become clear that one of the greatest potential impacts was the generation of high-qualified staff.
As an example, the "knowledge" and "know how" were basic achievements (as tools) for reducing the rejects and increasing box efficiency. (In the D7.5 D7.6 D8.5 and D8.9 where it´s analysed the profitability of project achievements, the "knowledge and know how” were one of the most profitable with a lower investment. (In our case, investment means money. Educational investment is quite another thing).
1.4.2 ICT’s in the production line.
Information and communications technology (ICT) is an extended term for information technology (IT) which stresses the role of unified communications and the integration of telecommunications (wireless signals), computers as well as necessary enterprise software, middleware, storage, and audio-visual systems, which enable users to access, store, transmit, and manipulate information.
One of the most important outputs was the introduction of ICT in the Flexicast-DEMO cast iron manufacturing cell.
DTA, for example, directly enters in the first line of production helping directly in the decision- making. And even that it´s a use friendly program, knowledge in metallurgy and mathematic tools are needed.
Foundry is an industry full of sensors and data. All of them are needed in some moment to take decisions. FLEXICAST has pointed out the need of having them interconnected, centralized, and analysed, and even that computer codes may manage them, the last decision must be human, not only due to safety reasons, but also because correlations, cause-effect reasons may change along the time. Foundry staff must be capable of understanding and managing them.
1.4.3 Wireless sensors
Foundry process manages a very high quantity of data, not only those directly connected with the metallurgical process, such as chemical analysis, DTA analysis, temperature, etc., but also much others data from equipment denominated “automatic auxiliary processes”, such as sand parameters, sand temperatures, sand permeability, .., air pressure, etc.
Foundries, like other process factories, evolve towards the massive use of sensors. But there are many cases where the physical connection between sensors and computers is impossible. By other side, extremely expensive foundry equipment should be a long-period replacement, then it must be there for years and years. It needs an intensive maintenance but also needs usually an updating and/or new technology.
The research with wireless connections made during the FLEXICAST showed that this technology, considered as "impossible" at the beginning of the project, (mainly due to electrical noise problems), it is now fully operative.
Until now, all the experiments to get data about the piece cooling along the moulding box carrousel must be done out of the production line. Thanks to the wireless technology developed in FLEXICAST, now it is possible to perform the tests and monitoring in the production line, which makes the process more convenient and less disturbing.
1.4.4 Remote diagnosis and treatment support (intranet and internet)
Nowadays, an important trend in the foundry industry, is the possibility of be connected to a centre where the data is received, analysed and there is the possibility of interacting remotely with the application. That is, not only it is possible to read and have access to the data “in situ” but also it is possible to program and remotely enhance the application.
In FLEXICAST, following the actual trends, DTA and X-ray are linked via internet with Proservice and InnospeXion partners respectively, and then it is now possible:
a.- To check if there is some malfunction and repair it remotely.
b.- Update the program(s), remotely, maintaining fully actualized.
c.-Analyse the data, interpret it and assist to the end-user in order to manage and to take decisions.
Given the fact that this “data centre” is linked to many foundries, each of them, with his own or similar problems, they will be the most skill and prepared staff, able to use and interpret the results.
1.4.5 Towards a global information structure
Taking account the trends mentioned in 1.4.2 and 1.4.5 there is no doubt that the staff of the control centre will have wide and global information of the foundry processes. This knowledge and know how will have a potential impact to supply an important and profitable technical support to the foundry (or any other factory).
1.4.6 Privacy conflicts
In some circumstances, the above mentioned items, may create a question: How to profit the skills and experiences of the “control centre staff" without generating problems among foundries competitors? This is an open question. This conflict is not new, but was limited and bounded before Internet era.
1.4.7 Needs of cast iron related educational programs and continuous training courses
The FLEXICAST project pointed out the need of highly skilled people, with proper knowledge that may not be incorporated only through personal experience or in house training. Unfortunately, few universities give engineering courses incorporating cast iron knowledge and technologies.
A complementary way is acquiring knowledge by continuous training courses, incorporating not only metallurgical knowledge, but also other background specially devoted to cast iron technologies.
1.4.8 Needs to encourage young people to study and learn cast iron technologies
From both sides: foundries and end users.
FLEXICAST project points out the need of a highly skilled staff. It has already denounced as a need from EU Industry. Pointing out the fact that skilled and experienced people in the foundry industry are going to retire in a near future and no young replacement is foreseen. It is necessary to encourage young people to study and learn cast iron technologies
1.4.9 Better design and use of the cast iron products
There is a need and demand from customers of parallel services to assist them for a better design and use of the cast iron products.
During the development of WP5, in particular, when some partners worked about mechanical simulation of fatigue resistance of cast pieces under stress, working in S.N curves linked with hardness, microstructure, tensile strength, and in the importance of clear and realistic boundary conditions, FLEXICAST consortium detected a lack of skilled people capable in the field of machinery design using cast iron pieces.
In this line, design guidelines and best practice rules focused on the best design of cast iron pieces have been delivered (D7.6 and D7.5).
Figure nº 22.- In deliverable D7.6 a guide of process to follow by customer and foundry was proposed following the trend to train and give a global service to customer, profiting the new knowledge acquired during FLEXICAST
Figure nº 23.- A protocol to classify the pieces by “families”, and defining the expected characteristics (mechanical and level of defects and type), has been built up to indicate the end user and foundry the expected characteristics of a new parts, following the trend of giving global services to customer.
1.4.10 Need of increasing the pre normative and normative reports and knowledge.
Traditionally, standards are more devoted to Quality control aspects such as rejecting or accepting pieces and reports, that to help in the technological design of cast iron pieces.
Figure nº 24.- Shows that FLEXICAST research goes towards this aim to complement the information given in the ISO standards
1.4.11 Economic aspects
To consider the impact of the achievements of FLEXICAST project, it has been considered which will be the impact in a “typical” foundry that incorporates several of them. This combination of achievements installed in a foundry is what called here “scenario”. The different scenarios have been defined in figure nº 1 but, for more clarify, it is repeated below as figure nº 25.
Figure nº 25.- Different scenarios analysed (combined results)
To cover a wide espectra of possible european foundries size, the simulation has been done for 20.000 15.000 and 10.000 Ton of yearly brut production. Figures nº 26, 27 and 28 show the foreseen economic impact in a “type foundry", thanks to the implementation of different FLEXICAST project achivements.
Figure nº 26.-Cost reduction, Needed Investment, and Final Result for the scenarios (20.000 Ton)
Figure nº 27.- Cost reduction, Needed Investment, and Final Result for the scenarios (15.000 Ton)
Figure nº 28.-Cost reduction, Needed Investment, and Final Result for the scenarios (10.000 Ton)
Note, that the investment has been calculated for a 10 years period of recovery, so the result is what normally is understood as “result” prior to taxes. It is supposed that the data presented will be effective once the achievements have been fully incorporated, (not before a minimum of 1 year, as a general rule, and depending on the degree of integration).
These are the economic impacts obtained through 4-years life of FLEXICAST project. These results are in line with the trends of the European Industry and, also, with the perspectives elaborated by other prestigious institutions, such as the European Foundry Industry Association , and Deutche Industry Bank, between others.
2 USE AND DISSEMINATION OF FOREGROUND
2.1 SECTION A
2.1.1 Publications
1. Ignaszak Z., Popielarski P., Hajkowski J., J-B. Prunier, Problem of acceptability of internal porosity in semi-finished cast product as new trend - "Tolerance of damage" present in modern design office, Defect and Diffusion Forum, Vol. 326-328, pp.612-619 2012.
2. Ignaszak Z., Hajkowski J., Popielarski P. – Mechanical properties gradient existing in real castings taken into account during design of cast components, Defect and Diffusion Forum, Vol. 334-335, pp.314-321.
3. Ignaszak Z., Popielarski P., Hajkowski J., – Sensitivity of models applied in chosen simulation systems with respect to database quality for resolving the casting problems, Defect and Diffusion Forum, 2013, Vol. 336, pp.135-146.
4. R. Sika, Z. Ignaszak, M. Rogalewicz, The specificity of applications of the Statistical Process Control methods in a complex production process for example foundry, Polish Metallurgy in 2011-2014, Monograph, Publish. Akapit, Edit. Officer K. Swiatkowski, Krynica-Zdroj (Poland), 2014, pp.357-371.
5. Ignaszak Z., Codina E., Hajkowski J., Popielarski P., Roquet P., Selected aspects of the use of thermal & derivative analysis to improve macro modelling of solidification for casting iron. Comparison of chosen simulation codes, 10th International Symposium on the Science and Processing of Cast Iron – SPCI10, Mar del Plata (Argentina) 2014.
6. R. Sika, Z. Ignaszak, Scada systems and theirs connection with Statistical Process Control in Foundry, Archives of Foundry Engineering, Trzebnica (Poland), 2015, Vol.15 Special Issue 1/2015, pp.145-153.
7. Z. Ignaszak, R. Sika, M. Perzyk, A. Kochanski, J. Kozlowski, Effectiveness of SCADA systems in control of green sands properties, Archives of Foundry Engineering, Busko Zdroj (Poland), 2016, Vol.16 Issue 1/2016, pp.5-12.
8. Ignaszak Z., Popielarski P., Hajkowski J., Codina E., Methodology of Comparative Validation of Selected Foundry Simulation Codes, Archives of Foundry Engineering, vol. 15/4 s.37–44, Gliwice (Poland), 2015.
9. Ignaszak Z., Prinier J-B., Effective Laboratory Method Chromite Content Estimation in Reclaimed Sands, Archives of Foundry Engineering, Vol. 16/3 pp.162 166, Gliwice - Łódź (Poland), 2016.
10. E. Villagrossi, L. Simonib, M. Beschia, N. Pedrocchi, A. Marini, L. Molinari Tosatti, A. Visioli, A Virtual Force Sensor for Interaction Tasks, Submitted to: Robotics and Autonomous Systems
11. E. Villagrossi, C. Cenati, N. Pedrocchi, M. Beschi, L. Molinari Tosatti. Robotized Solution for Cast Iron Deburring. Submitted to: International Journal of Advanced Manufacturing Technology.
12. PJ Gamez-Montero, R Castilla, J Freire, M Khamashta and E Codina, An empirical methodology for prediction of shape and flow rate of a free-falling non-submerged liquid and casting iron stream, Advances in Mechanical Engineering, 2016, Vol. 8(9) 1–12 (2016) DOI: 10.1177/1687814016669635
13. M. Chiumenti, M. Cervera, N. Dialami, B. Wu, J. Li and C. Agelet de Saracibar, Numerical modelling of the electron beam welding and its experimental validation, Finite Elements in Analysis and Design, 121 (2016) 118-133, http://dx.doi.org/10.1016/j.finel.2016.07.003i
14. N. Lafontaine, R. Rossi, M. Cervera and M. Chiumenti, Formulación mixta estabilizada explícita de elementos finitos para sólidos compresibles y quasi-incompresible, Revista Internacional de Métodos Numéricos para Cálculo y Diseño en Ingeniería, 169 (2015), http://dx.doi.org/10.1016/j.rimni.2015.09.003
15. M. Cervera, M. Chiumenti, L. Benedetti and R. Codina. Mixed stabilized finite element methods in nonlinear solid mechanics. Part III: Compressible and incompressible plasticity, Computer Methods in Applied Mechanics and Engineering, 285 (2015) 752-775
16. M. Chiumenti, M. Cervera and R. Codina., A mixed three-field FE formulation for stress accurate analysis including the incompressible limit, Computer Methods in Applied Mechanics and Engineering, 283 (2015) 1095-1116,
17. C. Agelet de Saracibar, M. Chiumenti, M. Cervera, N. Dialami and A. Seret, Computational modelling and sub-grid scale stabilization of incompressibility and convection in the numerical simulation of friction stir welding processes , Archives of Computational Methods in Engineering, 21 (2014) 3-37
18. S. Biswas, F. Sket, M. Chiumenti, I. Gutiérrez-Urrutia, J.M. Molina-Aldareguía and M.T. Pérez-Prado, Relationship between the 3D porosity and ß-phase distributions and the mechanical properties of a high pressure die cast AZ91 Mg alloy, Metallurgical and Materials Transaction A, 44(9) (2013) 4391-4403.
19. C. Agelet de Saracibar, R. López, B. Ducoeur , M. Chiumenti and B. de Meester, Un Modelo Numérico para la Simulación de Disolución de Precipitados en Aleaciones de Aluminio con Endurecimiento utilizando Redes Neuronales, Revista Int. de Métodos Numéricos para Cálculo y Diseño en Ingeniería, 29(1) (2013) 29–37.
2.1.2 Conferences proceedings
1. R.Sika Z.Ignaszak J.Kolakowski E.Codina P.Roquet SPC procedures for the manufacturing and complex unstable processes, Proceedings of Conference "Manufacturing 2014", 08-10.12.2014 Poznan, Poland.
2. P.Popielarski J.Hajkowski J.Matwiejczuk Z.Ignaszak. Comparative validation of selected simulation systems, Proceedings of Conference "Manufacturing 2014", 08-10.12.2014 Poznan, Poland.
3. Ignaszak Z., Popielarski P., Effective modelling of phenomena in over-moisture zone existing in porous sand mould using the simplified simulation systems applied in foundry, Proceedings of 10th International Conference on Diffusion in Solids and Liquids: DSL 2014, Paris, France, 23–27 June, 2014. publication in DDF by Scitec Publications Ltd (a subsidiary of Trans Tech Publications Ltd), Brandrain 6, CH-8707 Uetikon Zurich, Switzerland).
4. Ignaszak Z., Popielarski P., Effective modelling of phenomena in over-moisture zone existing in porous sand mould using the simplified simulation systems applied in foundry, Proceedings of 10th International Conference on Diffusion in Solids and Liquids: DSL 2014, Paris, France, 23–27 June, 2014. publication in DDF by Scitec Publications Ltd (a subsidiary of Trans Tech Publications Ltd), Brandrain 6, CH-8707 Uetikon Zurich, Switzerland).
5. Ignaszak Z., Popielarski P., Effective modelling of phenomena in over-moisture zone existing in porous sand mould using the simplified simulation systems applied in foundry, Proceedings of 10th International Conference on Diffusion in Solids and Liquids: DSL 2014, Paris, France, 23–27 June, 2014. publication in DDF by Scitec Publications Ltd (a subsidiary of Trans Tech Publications Ltd), Brandrain 6, CH-8707 Uetikon Zurich, Switzerland).
6. Ignaszak Z., Popielarski P., Identification of the moisture distribution in the intensively heated porous sand mould, DSL MUNICH 2015, 11th International Conference on Diffusion in Solids and Liquids 22 –26 June, 2015.
7. Ignaszak Z., Hajkowski J., Popielarski P., Gapiński B., Identification of the local properties of Al alloys high pressure die casting and cast iron products with validation of real discontinuities by computer tomography, DSL 2016 Split, 12th International Conference on Diffusion in Solids and Liquids26 – 30 June, 2016.
8. E. Villagrossi, G. Legnani, N. Pedrocchi, F. Vicentini, L. M. Tosatti, F. Abba and A. Bottero. Robot dynamic model identification through excitation trajectories minimizing the correlation influence among essential parameters. 11th International Conference on Informatics in Control, Automation and Robotics (ICINCO), pages 475-482, 2014. Doi 10.5220/0005060704750482.
9. M. Beschi, E. Villagrossi, N. Pedrocchi and L. M. Tosatti. A general analytical procedure for robot dynamic model reduction. IEEE-RSJ International Conference on Intelligent Robots and Systems (IROS), pages 4127-4132, 2015. doi 10.1109/IROS.2015.7353960.
2.1.3 Congresses
1. Ignaszak Z., 55th International Scientific Conference "Solidification and Crystallization of Metals 2014" 15th-17th September, 2014, Kielce-Mąchocice Kapitulne.
2. Ignaszak Z., Reporting Conference of Committee Metallurgy at Poland Academy of Sciences „Metallurgy 2014”, 15th-18th October 2014, Krynica-Zdrój.
3. Ignaszak Z., 10th International Symposium on the Science and Processing of Cast Iron – SPCI10, Mar del Plata (Argentina) 2014
4. Ignaszak Z., IV International Scientific and Technical Conference "Manufacturing 2014", 08 10.12.2014 Poznan, Poland.
5. Ignaszak Z., 10th International Conference on Diffusion in Solids and Liquids DSL 2014, Paris, France, 23–27 June, 2014.
6. Ignaszak Z., XV International Scientific Conference „Zapewnienie Jakości W Odlewnictwie 2015”, 20-22nd May 2015, Trzebnica.
7. Ignaszak Z., 11th International Conference on Diffusion in Solids and Liquids, DSL 2015 Munich, 22nd–26th June, 2015.
8. Ignaszak Z., 56th International Scientific Conference "Solidification and Crystallization of Metals 2015" 2nd-4th September, 2015, Busko-Zdrój.
9. Ignaszak Z., 12th International Conference on Diffusion in Solids and Liquids, DSL 2016 Split, 26th–30th June, 2016.
10. Ignaszak Z., XVI International Scientific Conference - "Optimization of Production Systems in Foundries" 14th-16th June, 2016, Łódź.
11. Ignaszak Z., 57th International Scientific Conference "Solidification and Crystallization of Metals 2016" 20th-22nd September, 2016, Kielce-Cedzyna.
12. Ignaszak Z., A General Analytical Procedure for Robot Dynamic Model Reduction, M. Beschi, E. Villagrossi, N. Pedrocchi and L. Molinari Tosatti, 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Congress Centre Hamburg, Sept 28 - Oct 2, 2015. Hamburg, Germany
13. N. Pedrocchi, E. Villagrossi, F. Vicentini and L.M. Tosatti. Robot-dynamic calibration improvement by local identification. IEEE International Conference on Robotics and Automation (ICRA), pages 5990-5997, 2014. doi 10.1109/ICRA.2014.6907742.
14. E. Villagrossi, G. Legnani, N. Pedrocchi, F. Vicentini, L. M. Tosatti, F. Abbá and A. Bottero. Robot dynamic model identification through excitation trajectories minimizing the correlation influence among essential parameters. 11th International Conference on Informatics in Control, Automation and Robotics (ICINCO), pages 475-482, 2014. Doi 10.5220/0005060704750482.
15. M. Beschi, E. Villagrossi, N. Pedrocchi and L. M. Tosatti. A general analytical procedure for robot dynamic model reduction. IEEE-RSJ International Conference on Intelligent Robots and Systems (IROS), pages 4127-4132, 2015. doi 10.1109/IROS.2015.7353960
16. A. Bonanno, M Merlin, GL Garagnani, E Codina, Numerical analysis and optimization of a lamellar cast iron hydraulic distributor. AIAS 2015-546 Associazione italiana per l’analisi delle sollecitazioni (AIAS). 44º convegno nazionale, 2-5 sept 2015. Univ. Di Messina
17. High-fidelity numerical simulation of additive manufacturing processes, M. Chiumenti, M. Cervera, C. Agelet de Saracibar, N. Dialami, W. Huang, X. Lin, L. Wei, Y. Zheng, L. Ma, COMPLAS 2015, 13th International Conference on Computational Plasticity, Barcelona, Spain, 1-3 September 2015
18. High-fidelity numerical simulation of additive manufacturing processes, M. Chiumenti, W. Huang, X. Lin, M. Cervera, C. Agelet de Saracibar, N. Dialami, M. Liang, L. Wei, MCWASP 2015, Modelling of Casting, Welding and Advanced Solidification Processes XIV, Awaji island, Hyogo, Japan, 21-26 June 2015
19. Numerical simulation of shaped metal deposition processes close to the industrial manufacturing technology, M. Chiumenti, W. Huang, X. Lin, L. Wei, M. Cervera, C. Agelet de Saracibar, N. Dialami, PLASTICITY 2015, 20th International Symposium on Plasticity and its Current Applications, Montego Bay, Jamaica, 4-9 January 2015
20. Coupled thermo-mechanical finite element technology for stress accurate analysis, M. Chiumenti, M. Cervera, R. Codina, C. Agelet de Saracibar, WCCM 2014, 11th World Congress on Computational Mechanics, Barcelona, Spain, 20-25 July 2014.
21. Stress accurate framework for the numerical simulation of FSW processes, M. Chiumenti, N. Dialami, M. Cervera, C. Agelet de Saracibar, FSWS 2014, 10th International Friction Stir Welding Symposium, China National Convention Centre, Beijing, China, 20-22 May 2014
22. Simulación numérica del proceso de soldadura por haz de electrones, M. Chiumenti, N. Dialami, M. Cervera, C. Agelet de Saracibar, CAIP 2013, 11º Congreso Interamericano de Computación Aplicada a la Industria de Procesos, Lima, Perú, 21-24 Octubre 2013
23. A novel stress-accurate FE technology for the numerical simulation of the FSW process, M. Chiumenti, N. Dialami, M. Cervera, C. Agelet de Saracibar, COMPLAS 2013, 12th International Conference on Computational Plasticity, Barcelona, Spain, 3-5 September 2013
24. A novel stress-accurate FE technology for highly non-linear analysis with incompressibility constraint. Application to the numerical simulation of the FSW process, M. Chiumenti, M. Cervera, C. Agelet de Saracibar, N. Dialami
25. Finite element modelling of EB welding process, M. Chiumenti, M. Cervera, C. Agelet de Saracibar, COUPLED PROBLEMS 2013, 5th International Conference on Coupled Problems in Science and Engineering, Ibiza, Spain, 17-19 June 2013
26. E. Codina and M. Khamashta, The cast iron technology innovations for fluid power components, MECSPE PARMA 2016, Power Drive Innovation, Parma, 18 March 2016
2.1.4 Workshops related with FoF-PPP
-Let's 2014: Bringing nanotechnology research into the European Production Industry (Bologna-Italy)
-Impact of the factories of the future PPP
(Brussels)
• 24-25 March 2014
• 29-30 April 2015
• 14-15 April 2016
•
2.1.5 Industrial fairs attendance
1. Kielce Trade Fair, 20th International Fair of Technologies for Foundry METAL, 16 18 September 2014 – PUT: participation as guest.
2. Kielce Trade Fairs - 23rd International Defence Industry Exhibition MSPO, 1 4th September, 2015 – PUT: participation as guest.
3. Kielce Trade Fair, 21st International Fair of Technologies for Foundry METAL, 20 22 September 2016 – PUT participation as guest.
4. MESIC-2015 Fair, Barcelona, Spain, July 24, 2015 – EURECAT: participation as guest
5. Innovation, Connect & Transform, ICT-2015, Lisbon, Portugal, October 20-22, 2015 – EURECAT participation as guest
6. 13th International Foundry Trade Fair with Technical Forum – GIFA, Dusseldorf, Germany, 10-20 June 2015 – UPC: participation as guest
7. 13th International Foundry Trade Fair with Technical Forum – GIFA, Dusseldorf, Germany, 10-20 June 2015 – Fundiciones de Roda: participation as guest
8. 13th International Foundry Trade Fair with Technical Forum – GIFA, Dusseldorf, Germany, 10-20 June 2015 – ONDARLAN- INDUCTOTHERM: fair expositor
9. 13th International Foundry Trade Fair with Technical Forum – GIFA, Dusseldorf, Germany, 10-20 June 2015 – EURECAT: fair expositor
10. 13th International Foundry Trade Fair with Technical Forum – GIFA, Dusseldorf, Germany, 10-20 June 2015 – FOSECO: fair expositor
11. 13th International Foundry Trade Fair with Technical Forum – GIFA, Dusseldorf, Germany, 10-20 June 2015 – PROSERVICE: fair expositor
12. MECSPE- Technologie per l'innovacione. 16 march 2016 Parma (Italy). LABSON-UPC (Prof E.Codina and Prof. M Khamashta)
2.2 Section B -PART B1
2.2.1 Patents
Ignaszak Z., Popielarski P., Hajkowski J., Poznan University of Technology, Measuring system of time and rate of casting mould cavity pouring, Application at the Polish Patent Office No. P.415693 2015.
List of Websites:
Consortium
The FLEXICAST consortium comprises 14 partners from 5 countries, having a complimentary knowledge, and covering the complete R&D chain:
2 Universities, 4 Research Centers, 4 SME’s, and 4 Industry (large)
Contact the coordinator for further information:
Dr. Ing Ind. Esteban Codina Maciá
Universitat Politecnica de Catalunya, UPC)
LABSON- Research Center
University Campus of Terrassa
E-mail: ecodina@mf.upc.edu
Tel +34 93 739 86 64
Web: www.flexicast-euproject.com
FLEXICAST (www.flexicast-euproject.com) is a four-year research & development collaborative project of the European Commission's Seventh Framework Programme (FoF.NMP.2012-7-Innovative technologies for casting, material removing and forming processes). FLEXICAST started on 1 Nov 2012 and has ended 31 Oct 2016. The project presented knowledge-based technologies with the aim to follow the way to transform the conventional (batch-by-batch) foundry process into a flexible (mould-by-mould) process.
The challenges that FLEXICAST project had successfully addressed to cast iron manufacturing processes can be split into three large blocks:
1- Get a high quality cast iron melt, which means:
• Melting at the minimum temperature and moulds filling at the maximum possible temperature.
• Keeping the temperatures as constant as possible
• Obtain the best quality of the cast iron melt, according to the type of cast iron (grey or nodular)
2- Get a maximum quality of casted iron pieces, which means:
• Starting from a high quality cast iron melt and a pre-established pieces geometry, the aim is to fill moulds and to solidify / cool the pieces in such a way that the obtained pieces should contain the minimum possible defects (from all points of view: internal sanity, mechanical performance, minimum deformations, etc.).
3- Get an automatic casting process based on robotics for foundry processes. This is the area in which there was a dire need for development and in which most SME’s are interested, which means:
• Ensuring job security and employment levels.
• Raising the technological base and take an important technological leap forward, particularly for industries based on conventional technologies (foundries).
• Leading the improvements needed in the workplace environment to a greater job skills based on knowledge and know how.
One of the most important outputs is a new melting shop, denominated "cast iron manufacturing cell". Conceptually, each component of cast iron manufacturing cell has been designed independently, following a development methodology, the so-called "step by step". The cell is basically composed by a hybrid heating system (induction and plasma), an automated pouring unit, a multiple inoculation system, a nodulising system (converter) as well as a network of sensors and instruments necessary to assess the quality of the cast iron. Cast iron manufacturing cell is a DEMO pilot plant in real industrial setting, which has allowed us to evaluate the progress of the project results in comparison with the state-of-art solutions, and we consider that it can be a starting point of a future lean foundry manufacturing system.
Another important result was the new knowledge and know-how acquired by all partners’ staff. Efforts have been put in the definition of a plan for internal and external exploitation of the project results. “Take-up” activities to promote the early or broad application of state-of-the-art cast iron manufacturing technologies have been implemented. These included the assessment, trial and validation of design, manufacturing and monitoring technologies, easier access to and transfer of best practices for the early use and exploitation of technologies. They have been targeted, in particular, SMEs in the cast iron foundry sector.
Project Context and Objectives:
FLEXICAST project has been crucial for the improvement of different cast iron processes and technologies in the production line, which led to significant energy and cost savings. These technologies are not only applicable to new cast iron foundry lines, but also are readily available to be retro-fitted to existing plants of cast iron.
In order to define in detail the objectives of the project, we have divided the project into 5 main pillars. These are:
1-New cast iron manufacturing cell (WP1/WP2)
2-Artificial Intelligence-based control system (WP1/WP3/WP5)
3-Automation based on robotics for foundry process improvement (WP1/WP4)
4-Demonstration (pilot plant or prototypes-WP6)
5-Life Cycle Assessment, Exploitation plan, Dissemination, and Socio-economic aspects (WP7/WP8)
Regarding the first pillar, the widespread adoption of new melting shop as an operating process was, in itself, fostering the creation of even more powerful induction-plasma power supplies, versatile melt control technology, high-power density furnaces, automated charging systems, temperature control systems, nodulizing systems (magnesium treatment), dynamic and multi-inoculation systems, and automated pouring systems. The ability to use various smaller furnaces for comparable productivity, faster melting, easier and more reliable chemistry adjustments, less oxidation of the melt, less manpower and much improved charging safety was significant.
The cell production was a model for workplace design, and has become an integral part of lean manufacturing systems.
Given the complexity to develop the cast iron manufacturing cell, it has been considered essential to develop each component independently. The cell, from a conceptual point of view, was basically composed of a hybrid heating system (induction and plasma), a pouring unit, a multiple inoculation system, as well as the sensors and instruments necessary to assess the quality of the cast iron.
The second pillar constituted the “smart control” in this project. The objective is to develop a software platform. The implementation of this platform integrates: Data gathering and Statistical Process Control Systems, Artificial Intelligence Systems (decision-making tools) and Sensors/Devices (among others: melting and pouring temperatures, chemical composition, DTA parameters, X-ray on-line inspection approach, sand parameters, quantities and velocities of inoculant agents), in order to control the new cast iron manufacturing cell.
Another main objective of this pillar was the integration of soft computing and hard computing to predict production performance in specific manufacturing processes. The experimental validation was included. The ultimate goal of process modelling was to predict the final mechanical properties. The mechanical properties (hardness, tensile strength, yield strength, and elongation) of cast iron pieces were functions of composition and microstructure. We have planned to use soft computing for implementing computationally intelligent and user-friendly features that could not be realized competitively by hard computing alone. It has been demonstrated that a high degree of interaction between soft and hard implementation economy of intelligent computing lead to better systems, improved tolerance against incomplete or uncertain sensor data, enhanced learning and adaptation capabilities.
The third pillar was related with the automation based on robotics for foundry process improvement.
The aim was the improvement of automation level of cast iron foundry through the introduction of robotized cell. The main paradigms for doing this were flexibility and easy intuitive robot programming. The basic steps were: Automation test samples cell; Multi part moulding box configuration cell, Innovative gripper for grinding and deburring for robotized solution in foundry environment, Inaccuracies compensation methods for easy grinding root programming and Control solution for grinding and deburring robots.
As already declared in the last paragraph, a first scenario has been defined in order to get more efficiency at the FLEXICAST strategic roadmap. The “moulding box configuration cell” has the aims automatize moulding box configuration. The partners have agreed that the efforts can be split in two main activities:
a) Design of the Robotized Cell For Moulding Box Configuration;
b) Development of the Robotized Cell Prototype for Moulding Box Configuration.
The technological objectives of these activities were: easy/intuitive operation programming, insertion and positioning of the cores, insertion and positioning of the filters; blowing the sprues and marking each sample with a unique serial number for quality tracking.
They were described as follows:
- 1 Requirements and line layout where the robotic cell was placed. In addition, description of the main design choices in terms of robots number and allocation tasks.
- 2 Requirements of the multi-function gripper, conceived and designed, in order to address all the possible applications.
- 3 Referencing the robots with respect to the moulding box carrousel in order to allow an accurate execution of the tasks.
- 4 Description of an innovative device developed in order to allow an easy and intuitive robot programming.
- 5 Requirements of the innovative device developed in order to check the correctness of the cores/filters placing.
-
The main goal of the fourth pillar was to demonstrate a clear breakthrough using project development in comparison with the state-of-art solutions. The main task was to construct the final demonstrators according to the specifications. It demonstrated that the hybrid manufacturing strategy and concepts were applicable to the range of processes operated by the industrial consortium partners, minimum in one open access pilot facilities. It was proposed to design and to manufacture a cast iron manufacturing cell into full-scale demonstrators (industrial). Also, it was managed the various software platform and robotic cells, being their effective performances fundamental for achieving the full commercial exploitation in the market place.
Finally, the main goal of the fifth pillar was to save and formalize the knowledge and experience accumulated (know how) through the progress of the project. This was done by a) capitalizing knowledge generated by design activity and the experience returns from using pilot applications.
b) Formalizing results of the project in guidelines and possible standards, contributing to building of the European standardization in this field. It included the following tasks: Experimental test benches and procedures. Standardization approach, best practices rules and safety rules. On the other hand, life cycle assessment has been treated in order to have ways to measure the efficiency of the environmental aspects on a number of selected processed pieces.
Another main objective of this pillar was to set up the successful exploitation of the results during the project lifetime and, in particular, after the project has concluded. The exploitation plan contained details about the market and research trends and opportunities for new areas for development and applications of the Flexicast environment.
It covered exploitation strategies for industry (including SMEs as well as large companies), research centres and universities, as well as synergies with other research projects.
In relation with dissemination activities, the continuous scientific work of all partners did lead to numerous scientific publication and media coverage. The publications covered scientific topics from metallurgic & mechanical engineering to robotics applications.
1. Relevant results publications of the research activity carried out in peer-review journal
2. Publication of relevant results coming from the research activity in Conference proceedings
3. Oral and Poster presentations of relevant results of the research activity in Int. Congresses
4. Patents
5. Workshops related with FoF:
a. Impact of the Factories of the Future PPP of relevant results of the research activity in EU projects. (4 editions)
b. Let's 2014 (1 edition)
6. Attendance to Industrial oriented Fairs either as invited guess or expositor
7. Publication articles in different newspapers or news into web-sites
A consortium of 14 partners from 5 countries has carried out the FLEXICAST project. These were:
UPC (Labson), Fundiciones de Roda, Eurecat, Cimne, Ondarlan, and Tecnalia from Spain,
CNR (Imamoter and Itia), Proservice, ARC, TSS Modena, and Comau from Italy,
Foseco from France,
InnospeXion from Denmark and
Poznan University Tech from Poland.
The total eligible budget was 9.174.941€, of which 5.700.000€ was EU contribution (62%) and 1.050 person months were dedicated.
Project Results:
In the following pages, the list of main achievements of the project is reported with a brief description of main activities. Further details are reported in the referenced deliverable documents. Deliverables and milestone description are also reported.
WP1: Specifications and Technical Coordination
WP leader: UPC-LABSON
MAIN TASKS:
This work package constituted the first step of the system engineering where the goal was to put in common, and cross partners' knowledge and expectative. It included the following tasks
1- Partner needs inventory, consisting in interviewing partners to know and formalise their needs concerning cast iron manufacturing processes.
2- Pooling the state of the art on sensors, data transmission, prognosis and proactive methods, guidelines and tools to support product and process quality management in specific manufacturing processes
3- Elaborating a common terminology, defining a common conceptual and technical terminology, in order to prepare the design and the implementation of innovative technologies for the demo applications and to prepare the knowledge feedback database.
4- Defining the application scenarios, consisting in analysing the requirements of demo applications and elaborating the technologies for each scenario.
5-Technical coordination. Given that FLEXICAST was a DEMO project and the aim was to demonstrate the validity of the project results at an industrial level, a lot of essential efforts in coordinating and monitoring the scientific progress (WP2, WP3, WP4, WP5) has been considered. It was necessary the corresponding technological implementation in an open access plant facilities. Obviously, this responsibility fell mainly on Fundiciones de Roda (FR) as the partner who had this possibility (cast iron manufacturing facilities).
One of the most important outputs of this project demonstrated that the research results and developments achieved in FLEXICAST project could be translated to old and/or new foundry industry (in general). The chosen way used was through “layout/scenario” concept. Scenario is defined as a combination of different units/developments in a “virtual model of foundry”. The first reference model of foundry was the actual FR layout. FR is a typical small–medium foundry, with all typical: departments, solutions and difficulties, which may be expected to found in a representative foundry. Current FR layout has been considered as “typical”, due to the following reasons:
- Is a typical medium-small foundry, well organized but with the similar characteristics and possible implementation problems as any other foundry.
- Use the layout of other real foundry, is not possible due to the difficulty to known exactly how it is (and no foundry is going to tell his “secrets”.
- To “invent” and ideal foundry layout is not applicable because either the difficulties of implementation are going to be already solved or artificial ones are going to be implemented.
Nine layouts/scenarios have been prepared incorporating different possible combinations.
Definition: "Reference scenario" is that foundry were none results of Flexicast has been implemented, and with a zero profit.
The objective of this definition was to make it easier to consider the impact of the project achievements when they were implemented in different scenarios (figure nº 1). For each scenario, the advantages, inconvenient and opportunity of incorporating each improvement have been described .
Figure nº 1.- Different scenarios analysed (combined results)
Figure nº 2.- Flow diagram of cast iron production line, from raw material to finished pieces
WP2: New cast iron manufacturing cell
WP Leader: F de RODA
MAIN TASKS:
WP2 contained five tasks. It has been focused on R&D activities related with the development of a new cast iron manufacturing cell.
1- Melting process (induction and plasma heating). The goal was to explore and demonstrate that a hybrid combination of different technologies can lead to an overall improvement of the efficiency, provides high heat transfer density and finally, improves cast iron quality.
2- New pouring system (high performance and low resources consuming). The goal was to install the melting system closer to the automated pouring system (heated by plasma), keeping as close as possible to the mould carrousel. Melt treatment (nodulizing) have been done into the furnace unit or the pouring unit or both of them. Successful production of high quality ductile iron depends on the successful control of magnesium. The incorporation of a system for heating and controlling melt temperature based on a plasma torch has been probably one of the most significant innovations at casting unit level that appeared in the last twenty years.
3- Nodulizing process methods. Successful production of high quality ductile iron depends on the successful control of magnesium. Melt treatment with Mg (nodulizing) have been done into the pouring unit through three alternatives: Mg vapour treatment, Initek process (converter) or wire core system. The innovative goal was to design a new nodulizing procedure based on magnesium vapour produced in a small external furnace.
4- Inoculation systems. With regard to inoculation (most appropriate type(s) and size(s) of inoculant(s), addition rates, and inoculation chain...) different options have been analysed. An inoculant is a material, added to the liquid iron just prior to moulding process, which provides a suitable phase for nucleation of graphite during the subsequent cooling. A breakthrough has been based on the multiple-dynamic inoculation. In this case, “multiple” means a combination of elements from inoculants battery, and “dynamic” means that a stream of inoculant(s) is added at a variable rate according to the melt flow.
Highlights of most significant results:
1- New manufacturing cell:
-Induction furnace. ONDARLAN, UPC
-Pouring unit (v1.0 and v2.0) F. de RODA, UPC
-Plasma heating TECNALIA
2- Analysis of lining wear in the pouring unit F de RODA, UPC
3- Measurement of melt jet temperature in pouring unit. F de RODA, PROSERVICE
4- Evaluation of heat losses in pouring unit in order to estimate the power supply needs of the plasma heating system. F. de RODA, UPC. CIMNE
5- Experimental evaluation of magnesium fading in induction furnace and in a pouring unit.
TECNALIA, F de RODA, UPC, ONDARLAN
6- Adaptation of the plasma prototype to the pilot scale. TECNALIA, F de RODA, UPC
7- Innovative prototype for nodulizing process with magnesium vapour. TECNALIA.
One of the most important results of this WP was a full characterization of plasma torch prototypes and the main outputs were the construction of the prototypes and, as a result of the trials, conclusions of energy efficiency and process working conditions, applied on the final prototype designed.
Comparing with standard heating systems, the hybrid induction-plasma and plasma heating technology offered:
• Power consumption can be reduced up to 35% depending on material and casting conditions. Average improvement value of 5-15%.
• Metal characteristics achieved were considered at an industrial acceptance level.
• Mechanical properties of produced material were at the same level than the ones melted using only induction heating.
The overview assessment of the plasma heating technology had the following conclusions:
• Allowed to heat only a desired portion of molten metal which is going to be cast.
• Simple design. Easy operational work.
• Accurate casting temperature regulation.
• Improvement of energy efficiency due to more efficient heat transfer to the melt
Finally, anodic inoculation trials at pilot plant have been carried out. Assessment of the effect of the quality of the crystalline graphite on the molten metal quality measured with DTA technology.
Most significant and measurable results were:
• Improvement of the metallurgical quality through DTA analysis.
• Electric ideal conditions: Power 20 to 70 kW. Intensity form 400 to 480 A.
• Physical ideal conditions: HP electrodes submerged at least 50 mm. Optimal 80 mm.
• Savings in inoculant materials when process adjusted at pilot plant trials.
• Anodic inoculation system. See results in the deliverable.
Main conclusions of the assessment and design of anodic inoculation process and equipment were:
• Feasibility tests of anodic inoculation successful.
• Ideal electrodes to be used are HP quality. For FLEXICAST selected industrial site 50 to 60 mm of diameter.
• Design of anodic inoculation device includes HPTP plasma torch equipment.
• Improved quality of the molten metal obtained with the system.
WP3: Artificial Intelligence-based control system (AIBCS)
WP Leader: PUT
MAIN TASKS
The metal casting process requires testing equipment that, jointly with properly customized computer software, supports the casting component analysis (characteristic properties). Due to the fact that this evaluation process involves the control of complex and multi-variable melting, casting and solidification factors, it has been necessary to develop more specific software. The AIBCS Platform was designed to support cast iron manufacturing processes, faulty components and unexpectedly higher scrap rates, even chemical analysis, pouring temperatures and so on.
WP3 and WP5 have constituted the “smart control and knowledge platform” in the project.
This platform has been integrated by: data gathering and statistical process control systems, artificial intelligence systems (decision-making tools), sensors/devices (among others: melting and pouring temperatures, chemical composition, DTA parameters, X-ray on-line inspection approach, sand parameters, quantities and flow rate of inoculant agents) and, soft & hard computing methodologies.
Figure nº 3.-Integrated AI-based Control Systems and partners involved
WP3 contained four tasks:
1- Real-Time Data Acquisition System (RTDAS). The goal was to monitor the cast iron manufacturing unit through the simultaneous measurement and recording of melt temperature in different locations and environmental conditions around the melt such as metallostatic head pressure, flow rates, inert gas pressure, sand parameters (moisture, permeability and compactness, etc.).
The data repository included, among others, thermal analysis parameters, x-ray imaging, environment conditions, process settings and melts chemical composition.
The time-temperature data, as well as the input and output values of the environmental conditions have been automatically saved and time stamped for traceability in order to make it feasible to generate graphical plots of various melt samples.
The assessment of the RTDAS data in the AIBCS in terms of accuracy have been determined by the deviation of measured values from sensors against the precise value (average value).
One of the first challenges that have been faced was the development of an electronic system that could be interfaced with different kinds of sensors (thermocouples, humidity sensors, pressure sensors, etc...). These sensors had a completely different output characteristic as well as the signal conditioning, so the unit should provide a reconfigurable analog front-end to support these sensors. Another required feature for the acquisition system was the capability of managing an acquisition system through a wireless communication, and monitoring the data in real-time.
Performing a reliable wireless communication in a foundry environment has been a challenging task, because there were several factors that interfered with the signal propagation, reducing the communication range. The technologies evaluated were the standard WiFi IEEE 802.11b due to its large diffusion and the potential integration with existing IT infrastructure and the IEEE 802.15.4 ZigBee modules. The results with ZigBee performed in Fundiciones de Roda proved an excellent performance.
2- Statistical process Control System (SPCS). The goal was to assess the AIBCS outcomes in terms of accuracy, reliability and timeliness. It was based on probability and statistics analysis. Once the sensor signals have been recorded by the RTDAS, they have been statistically analysed in the SPCS module. The SPCS monitors, measures, and evaluates melt quality based on pre-defined criteria.
Casting processes, considering data acquisition methodology proposed by FLEXICAST partners (in particular: PUT, PROSERVICE, FOSECO, Fund. de RODA and EURECAT), have been divided chronologically to casting related sub-processes:
• build the mould,
• pouring the metal into the mould,
• melting and metallurgical evaluation of the cast iron,
• final check of the casting quality, according customer acceptance conditions.
The following parameters have been recorded:
• Time of tapping from furnace to pouring ladle,
• Initial and final temperature of pouring the metal into mould [°C],
• Number of sample for chemical analysis,
• Time of making the test using spectrometer,
• Quantity of alloying elements: C, Si, Mn, P, S, Cr, Ni, Al, Cu,
• Equivalent carbon value CE = %C+1/3*(%Si+%P),
• Time of measurement of moulding material parameters,
• Moulding material parameters: green sand strength RW, permeability
• Humidity, bulk density, compactness, etc.
• Defects percentage [%].
An innovative aspect of the SPCS was the assistance roll in identifying abnormalities in the melt, improving castings quality through an “intelligent” system which can predict the casting properties and can provide recommendations and actions to the operator for process improvements.
PUT and F. de RODA partners did the core of this activity. The details of the work were described in the periodic reports and deliverable D3.3 -Statistical Process Control System and some published technical paper.
Figure nº 4.- AIBCS main screen (designed by PUT )
3- Case-based and Fuzzy Inductive Reasoning: One of the biggest challenges in foundry industry is the castings production according with a customer predefined quality. Possible casting defects, at different intensity levels must be referred to these predefined conditions. Procedures of analysis of defects and its causes require using software tools. Methods from the area of soft computing have been proven, being more and more useful for solving these matters. Examples of these methods are decision trees (DT) and artificial neural networks (ANNs), which can realize such analysis in a more effective way. En general, this is an approach named Data Mining.
Demanded goal function was the quality classification of final product, which qualifies the casting to one of two classes: conform to the final product or non-conform to the final product. Disposing a full range of available information (data) about technological parameters of the process (input variables) and information about the amount of faulty products regarding to selected assortment and date of production (output variables), it was possible to build a model based on that collection of teaching data.
The first approach was to develop and implement a Fuzzy Inductive Reasoning algorithm (FIR). Nevertheless first experimental tests showed that FIR was not the most suitable tool due to the large number of variables of the process and the poor time dependency. For this reason, the strategy was changed towards a machine learning approach.
A comparison of several machine-learning algorithms including Support Vector Machine, K-Nearest Neighbours and Linear Discriminant Analysis, was performed. The experimental results combining the foundry raw data with the state of the art algorithm showed again a poor performance. It was due to the large number of variables that introduce undesired noise to the system and the unbalanced class distribution.
It was important to note that all the research and development described in the periodic reports was based on real production data not designed or controlled experiments. This fact increased the difficulty of the development and off course reduce the accuracy of the systems, but allowed to achieve much more powerful, robust and real-life ready solutions.
As we have seen in the previous paragraph, applying state of the art algorithms to the raw data do not produce satisfactory prediction rates. Developing a new algorithm able to pre-process the raw data and looking for the most relevant parameters became crucial. Nevertheless due to the large number of variables, a brute force approach was not feasible. The main challenges were:
•The large number of variables degrades the performance of the machine learning algorithms, by adding noise to the system.
•Unbalanced classes: much more samples without the defect than with it.
•An optimal set of input features (if exists) must be found, but we are facing a combinational explosion. For instance, there are 2.4E+23 possible ways to pick 25 variables of a set of 100. At 1ms per combination, that would take us 7 billion years.
In order to overcome the previous challenges, we have developed the Indiana Module which aims to find a sub-optimal combination in an affordable time-frame. Facing the combinational explosion:
-With the aim to create a feature selection algorithm for pre-processing the raw data, a novel approach was considered, based on subset partitioning and feature ranking.
-The Indiana module analyses different combinations of input parameters with the aim to find a sub-optimal set of input parameters (in an affordable timeframe) that boost the performance of the classifier.
-A genetic algorithm refines the best model based on the top ranked features found in the previous stage.
The Indiana Module follows the following steps to reach a sub-optimal solution:
• Creation of sub-models by partitioning the input features space in order to rank the input features
• Pick the n-better ranked features
• Apply a genetic algorithm to obtain the best model
Indiana Module main characteristics and advantages:
• Feature selection algorithm
• Sub-optimal best model discovery by means of genetic algorithm
• Multiclass capability
• Multi-Classifier compatibility
• Automatic-weight distribution for inter-class balance
Figure nº 5. Pre-processing strategy to boost the classifier’s performance
The prototype has been developed in "Python" and did not require the installation of any other library. In the last years, python has been quite popular thanks to its growing community of developers, focusing on Data Analytics and mathematical computation.
The experimental results have shown that, thanks to pre-processing, the raw data with the Indiana module, the performance of the machine learning classifiers was boosted, increasing their accuracy by at least 50%. Overall, combining Indiana module with a Support Vector Machine offered a better compromise between accuracy and flexibility to adapt to the different iron types and defects.
Figure nº 6.-Three Dimensional projections of the raw data versus the pre-processed data for the grey iron porosity prediction. The axis corresponds to the 3 most relevant latent scores of a Partial Least Regression
Thanks to the algorithm design, the Indiana Module not only provides to the classifier a reduced set of input features, but also ranks all the features. This information can be very helpful for the foundries in order to identify those process parameters that are relevant to predict the presence of some defects. The delivered solution has been proven to offer an excellent performance, reaching more 85-90% of accuracy when predicting shrinkage and porosity for both grey and laminar iron, and improving by a large margin the state of the art solutions.
4- Holistic enabling technologies for cast iron quality validation: In this section, Differential thermal analysis (DTA) and X-ray approach for internal inspection were included.
In relation with DTA, the main goal has been the realization of a new system that overcomes all the other equivalent commercial products and, in addition, was suitable for the new foundry approach defined in the Flexicast project: shift of production process logic from batch-by-batch to mould-by-mould approach. Two independent thermal analysis systems have been installed: one focused on the preparation of base iron in melting furnaces and one focused on the final iron in the pouring station, in order to make possible the control and the study of the melt along the whole production process.
On the DTA systems has been implemented some algorithms that evaluates all these three factors and that raise focused alarm messages if some problems are found. They required a high effort due to intrinsic difficulties on the signal elaboration.
The main innovation was related with the results interpretation approach: the clock gauges helped to the operator about the result suitability through the colour areas (green/yellow/red).
As regards the final iron analysis, a different interface (and in fact an almost complete different system) has been developed. On this production process phase, the forecast of metallurgical defect tendency become one of the most important features. The interface developed was similar to that one developed for base iron (it shows the cooling curve and the most important results) but on the bottom part a series of clock gauges informs the workers about the current metallurgical status, relieving the workers of the results interpretation operation.
The defect indicators were:
• Cementite: probability of carbide formation on the casting surface;
• InverseChill: probability of carbide formation in the casting core;
• Porosity: probability of internal cavity formation due to thermal conductivity issues;
• Macroshrinkage: probability of internal cavity formation due to uncompensated liquid shrinkage;
• Flotation of graphite: probability of graphite segregation in the casting.
The other indicators provide some other important information about the iron properties:
• HypoEutecticHyper: it indicates the position of iron in the Fe-C state equilibrium diagram;
• Graphite expansion: it indicates the level of the graphite formed during the solidification;
• Nodulizing: it shows if the iron is nodular or grey.
One of the most important milestones of this WP was that DTA had to interpret the results in function of the boundary conditions (i.e. the specific casting features, like thickness, thermal module and so on). New module "casting designer" has been implemented to achieve this result.
Figure nº 7.- DTA display
Finally, in order to make an easier usage of DTA system for the foundry workers, three remote displays have been implemented and installed.
In relation with internal sanity, the first stage of know-how may be, of course, to consider a simultaneous specification of such defects generated during the metallurgical manufacturing processes, such as: gas bubbles, porosities, inclusions of foreign inclusions, segregation of the components, stratification, incomplete fusions, surface and internal cracks.
X-ray technology was highly requested by manufacturers (and also by the end users), for defects revelation (Internal sanity). In this sense, the consortium wanted to highlight InnospeXion's efforts relative to the development, design, construction and testing of a novel X-ray system for cast iron quality control. Flexicast consortium would like to emphasize that this equipment took in consideration the ASTM E1030-05 (2011) requirements. The objectives of the task were:
• Radiation cabinet design
• X-ray source acquisition and tests
• Optimize imaging technology (control of scattered radiation, etc.)
• Development of manipulation system (with FR), incl. manipulation system drawings and description to FR for acceptance
• Develop counting detector technology – interfacing & software development – for micro porosity determination & quantification
• PC configuration
• Software development for control (viewing software only, no automatic interpretation)
• Cabinet construction
• Installation of hardware, manipulation system, test of control and motion system etc.
• Testing of samples (imaging tests)
• Testing of samples (radiation measurement tests, porosity determination)
In Flexicast project, the focus was on iron cast samples with a thickness (in the X-ray imaging direction) up to 50 mm. The samples had different geometries, from planar to complex shapes (figure nº 8). Also, the samples used in this project for demonstration purposes were: railway guides (planar), control valve; seal ring and pump housing (complex). In addition to the cast samples, there was also a need of inspecting the “cores”, e.g. the sand – epoxy precursors for the cast part. These “cores” were crucial to inspect since their integrity determines the presence of major defects in the cast parts.
Figure nº 8.- X-ray castings inspection
Flexicast consortium partners have shown that a high degree of interaction between soft and hard implementation economy of intelligent computing lead to better systems, improved tolerance against incomplete or uncertain sensor data, enhanced learning and adaptation capabilities. See technical paper D8.6-Disseminationa activities.
Figure nº 9.- Temporal evolution and Normal distribution of porosity (%) and Rm of GJS500-7
for annual periods (2012-2015)
The AIBCS outcomes (from 2014 to 2015) were compared to the traditional process parameters commonly used for analysis, control and prediction of characteristics in the casting process (before 2012). The qualitative results shown that the Integrated AIBCS has a high future potential value.
There was no doubt that figure nº 9 demonstrated improve results achievement in FLEXICAST project. The incorporation of new methodologies and advanced technological developments were successful. Properly, the integration of the different instruments made this AIBCS a successful methodology from holistic point of view.
WP4: Automation based on robotics
WP Leader: CNR-ITIA
MAIN TASKS
WP4 contained two basic tasks. It was focused on the automation based on robotics for foundry process improvement. It included the following tasks:
1-Develop the affordable systems having the versatility to perform an automatic test samples cell and a multi-part moulding box configuration cell.
2-Develop a deburring robotized cell, which almost totally eliminate the problems in fettling and finishing of castings made in batch quantities.
A huge number of automatized solutions for each phase of cast iron production processes were present on the market, but, although this variety, few of these devices were installed in the real plant, mainly for only one reason: the low flexibility of these devices. In particular, for its programming procedure: too long and complex, that makes them feasible for large batch pieces production. Usually these machines presented very high productivity but with long reconfiguration time.
On the contrary medium cast iron foundry, presented high variability of pieces type in their daily production that require high and fast configurability of the plant. For this reason, machines with high productivity but with low flexibility, could not represent a suitable solution for medium dimensions cast iron industry. The design of robotized cell, has considered flexibility as a guideline for FLEXICAST project.
The first action was a preliminary analysis of the actual state-of-art of automation in the cast iron foundry. The second step, in charge of CNR, was the design of new robotized solutions in order to overcome the limit of actual adopted. The main paradigms for doing this were flexibility and easy intuitive robot programming. The use of industrial robots not only for standard operations but also for machining applications have been required a deep analysis of robot structure and radical improvement of the standard control unit. The WP4 has been structured in 5 main items:
1- Automation of test samples cell (COMAU);
2- Multi-part moulding box configuration cell (CNR-ITIA);
3- Innovative gripper for grinding and deburring for robotized solution in foundry environment (CNR-IMAMOTER);
4- Methods for
(i) Easy deburring and grinding programming and
(ii) Part-programme compensation of the mould production inaccuracies (CNR-ITIA);
5- Control solutions for grinding and deburring robot (CNR-ITIA).
Cell #1: TEST SAMPLE CELL
The detailed description of the layout of the Cell #1 was reported in D4.3.
The complete cycle consist on a sequence of pick and place of the mushroom sample and different stations that perform the different technological phases.
The list of tasks was reported below for sake of clarity:
- Cooling the sample with the core
- Braking the core
- Cooling the sample without the core
- Sanding (light polishing)
- Cutting the feeder
- Polishing the medal (one side, three stages)
- Insertion of the part in the Spectrometer/ Microscopy analysis.
Cell #2: MOULDING BOX CELL
The R&D work has been focused on two layers:
- To develop an easy and intuitive device for the lead-through-programming that is robust and usable in the foundry environment
- To develop the whole programming framework for the moulding box configuration (figure nº 10).
The tool has been calibrated by exploiting the standard COMAU programming modalities. The accuracy of the modality (+/- 0.3mm) grants the task requirements. For a detailed description of the cell refers to D4.4
Cell #3: GRINDING AND DEBURRING CELL
For a detailed description of the cell (figure nº 10) refers to D4.4 & D4.5
The programming phase realized can be divided into four main stages:
1. Measure of the Master Piece (e.g. piece manually deburred fixed in the desk);
a. Acquisition of areas for identification of the rigid roto-translation
b. Acquisition of areas for the identification of the deformation (afterwards the deburring trajectories will be defined in these areas)
2. Measure of the deburring trajectories in the Master Piece;
Acquisition of the deburring path
3. Measure of each Slaves Sample;
a. Acquisition of areas for identification of the rigid roto-translation.
b. Acquisition of areas for the identification of the deformation
4. Deformation of the deburring trajectories for each Slave Sample.
Execution of the deburring path on the Slave Sample taking into account the roto-translation and the deformation
Figure nº 10.- Flexicast robotic cells (#2 and #3)
The results obtained from the kinematic calibration procedure described in D4.5 have been evaluated using an optical measurement system; in this case a NDI Polaris Spectra. This instrument allows the tracking of a passive marker with accuracy up to 0.35 mm of RMS at the maximum distance in the workspace.
WP5: Simulation and optimization techniques
WP Leader: CIMNE/PUT
MAIN TASKS:
WP5 contained two tasks, focused on the development of autonomous computational intelligent systems. The prime objective was the integration of soft and hard computing to predict production performance in specific manufacturing processes. The second one was the experimental validation.
1- Thermo-fluid dynamic modelling. This task has been split in two parts. One devoted to flow analysis in melting systems and the other was focused to mould filling and solidification processes.
The aim of the first part was to show the numerical simulations and results carried out in flow analysis in melting systems related with cast iron manufacturing cell, in particular:
- Laminar barrier in furnace unit (figure nº 11)
- Pouring flow from furnace and running (tundish) unit
Figure nº 11: Gas flow for laminar barrier
A common problem, in most furnaces, was that metal melt inside the furnace reacts quickly with an open atmosphere. Elements in the casting like silicon, aluminium, chromium, etc. react with the oxygen forming slag. This slag can waste part of the casting by causing rejection of part of it. In order to avoid this particular problem, a laminar barrier of inert gas has been added. In this activity, a free Open-Foam code has been used. It is a C++ toolbox for the development of customized numerical solvers, and pre/post-processing utilities for the solution of continuum mechanics problems and also computational fluid dynamics (CFD).
As a main goal of this activity, 7-9 m/s gas flow was the optimal barrier configuration which protected the melt. Figure nº 11 shows the streamlines and the magnitude of the velocity.
Other challenge of the project was to estimate the flow rates with cast iron pouring units. Numerical simulations have been performed for both water and cast iron. Laboratory experimental tests with water were used to validate meshing’s and numerical schemes. The numerical results with cast iron have been used for the flow rate estimation in a real industrial process. Another alternative analysed was related with the experimental measurements carried out in a non-submerged vertical free liquid jet. Specific experimental assessment and quantification of the quality of the jet have been done (Jet flow rate measurement method). Then, a new technique to vertical free jets named Visual Photographic Technique (VPT) was proposed. In contrast to the PIV technique, VPT has a lack of the velocity profile/field measurements. However, in the present work, several mathematical models were proposed to allow VPT estimate the flow velocity and flow rate of a non-submerged vertical liquid jet. It is based on the DGHT (Dynamic Generalized Hough Transform) algorithm. This technique required following steps (see figure nº 12):
-Binarization of a grey-level image from a PiV measurement through a threshold operator
-Noise removal by size and salt-and-pepper operators
-Edge detection
-Determination of interpolation (connectivity) points to be used in the DGHT algorithm
-Computation of ellipse parameters through the DGHT comparison of reconstructed ellipse
Figure nº 13 shows a comparison between estimated and experimental results.
Figure nº 12.- DGHT algorithm steps
Figure nº 13.- Comparison between estimated and experimental results
The ultimate goal of process modelling was to predict the final mechanical properties. The mechanical properties (hardness, tensile strength, yield strength, and elongation) of cast iron pieces are functions of composition and microstructure. The graphite shape, graphite structure, graphite amount, carbide content, and matrix structure (pearlite, ferrite) all affect the mechanical properties of ductile iron castings.
Research activity carried out deal with the numerical simulation and experimental validation of the cast iron pieces filling and solidification process. The model pieces were two types: specific Flexicast reference pieces and simple geometrical pieces (like samples). This work (numerical and experimental tests) included:
A-Shrinkage test
Comparison of different HTC
Analysis of results of experimental-simulation tests
B- Castability of Cast Iron
Status quo based on cited references
Castability trails
Computer simulation of castability trials
C- Virtual mould (figures nº 15 and 16)
Casting of reference piece -2525
Description of simulations for versions of 2525 casting
Version 1 (classic sand risers)
Version 2 (2 exo-risers)
Destructive testing realized by Fundiciones de RODA
Preparation of simulation in Procast code.
Simulation Procast MICRO
Solidification results
Simulation Procast MICRO – micro model results
Additional validation tests
Figure nº 14.- Experimental and numerical results of simple geometrical pieces tests.
Figure nº 15.- Experimental and numerical results of specific Flexicast ref. 2525 DCV piece.
As an example, figure nº 15 shows the PROCAST micro model simulation (2525 DCV piece)
Figure nº 16.- Micro model results
These activities were proposed and realized in F de Roda facilities, in collaboration with PUT team, directed by Dr. Zenon Ignaszak.
Also, as planned in the DoW, all reference pieces have been experimentally characterized to assess their mechanical performance (fatigue, wear and deformation). TSS-Modena, F de RODA, CNR- Imamoter and UPC-Labson have carried out these works.
WP6: Demonstration activities. Prototypes
WP Leader: F de RODA
MAIN TASKS:
The main goal of WP6 was the final DEMO prototypes construction, according to WP2 and WP3 and WP4 requirements. It included the following main cells:
-New cast iron manufacturing cell.
-Robot cells for cast iron automation process
The specific activities related with DEMO were:
1-Assembly, installation and validation of cast iron manufacturing cell in F. Roda facilities:
-Induction furnace (and auxiliary equipment)
-Pouring units (and auxiliary equipment)
-Plasma equipment
-Multi-inoculation system
2-Assembly, installation and validation of the following equipment in F. Roda facilities:
-DTA systems and corresponding software.
-Robotized Cell Prototype for Moulding Box
3- Assembly, installation and validation of X-ray equipment and cabinet prototypes
4- Assembly, installation and validation of converter equipment prototype
5- Assembly, installation and validation of the cast iron manufacturing robotic cells: sample cell and deburring cell. These have been installed and validated in CNR facilities.
Figure nº 17 shows some DEMO prototypes related with FLEXICAST cast iron manufacturing cell in F de RODA facilities.
Related to Mg vapour prototype (figure nº 18), the tasks carried out were similar to the ones of the plasma torch and anodic inoculation device. All these activities have been carried out in Tecnalia facilities.
Trials have been successful, and feasibility of magnesium fading compensation in cast iron using magnesium vapour device is demonstrated at industrial relevant environment.
One important aspect of the plasma device operating was the availability for common maintenance actions without generating any operative trouble. When the plasma torch is at the working stage on the vessel allows the full tilting of the unit (Figure nº 19). The tilting of the unit can be due to maintenance operations (cleaning, emptying, etc.) or due to actions related to safety (rapid emptying, need of taking out the iron from the pouring vessel stopper area ...).
Related to the use of plasma-induction hybrid heating devices at foundry sector, these results indicated that a sensible reduction of energy was obtained when using a plasma torch as melting device (single or hybrid technique). Metal losses ratio using scraps improves respect to other melting technologies (due to inert atmosphere which reduce oxidations and the more fast and efficient heat transfer).
Figure nº 17.- Synthesis of the DEMO activities realized during 2015-16
Figure nº 18.- Mg vapour tests
Figure nº 19.- General view of plasma torch device mounted on the big tundish
In relation with the deburring robot cell (figure nº 10), setup was based on a COMAU C5G NJ220-Foundry industrial robot, a single point laser displacement sensor (LMI Gocator-1170) and a common pneumatic spindle. The three control strategies implemented were:
• Deburring maintaining the execution velocity constant
• Deburring slowing down the velocity according to the measured forces
• Deburring through the human mimetic control strategy
A more detailed analysis has addressed the evaluation of the accuracy of macro misalignment registration algorithm from a statistical point of view. In this case, a MWP was acquired through the laser scanner measurement device, thus the same WP was used as “fake” SWP in order to have two WPs equal from the geometrical point of view where the only difference is the location on the workbench. In this way it is possible to exclude the effect of local deformations. The SWP was evaluated in 30 different positions by acquiring, for each different position, the relative point cloud. In addition, 3 different deburring trajectories were defined, as straight lines, on the MWP.
Trajectories lie in 3 different planes of the WP and each plane is orthogonal w.r.t. the other ones. The acquisition of each single point of the deburring trajectories was done exploiting the laser point sensor instead of the deburring tool. This approach allows to a have a direct measure of the distance between the laser scanner measurement device frame and the WP surfaces. The measures acquired on MWP and on each SWP (figure nº 20), after the registration of macro misalignment, should provide the same distance between WP surface and laser scanner measurement device frame.
The real differences between laser scanner measurement device measures provide estimation on the “global accuracy” that can be reached by applying the macro misalignment registration algorithm. . Global accuracy is a quantity that takes into account the measuring system accuracy, the robot positioning accuracy, and the misalignment registration algorithm accuracy. From a technological point of view, the threshold of 1mm of RMSE, is evaluated as sufficient from expert workers.
Figure nº 20
Panels (A-B) for a real MWP and a real SWP placed in a significative different position.
Panels (C-D) for the raw point clouds of MWP and SWP simply overlayed.
Panels (E-F) for the SWP point cloud aligned on MWP point cloud.
A video of one of the experiments has been published online in the ITIA-CNR Youtube channel at:
https://www.youtube.com/watch?v=ciRWDrcLReQ.
WP7: Cast iron technology knowledge. Best practice rules
WP Leader: F de RODA
MAIN TASKS:
The goal of this WP7, which constituted a permanent support activity of/for the project, was to save and formalize knowledge and experience accumulated along the progress of the project. It included: Experimental test benches and procedures, pre-standardization approach, and best practices rules and safety. The main results were summarized in specific knowledge deliverables (D7.2 D7.3 D7.4 D7.5 and D7.6).
On the other hand, life cycle assessment was developed in order to have ways to measure the efficiency of the environmental aspects.
In order to perform the LCA analysis it was necessary to know, in detail, both the input and output data of the whole manufacturing process and relative environmental emissions (figure nº 21).
Figure nº 21.- Block diagram of the cast iron manufacturing process
The input and output data of materials and resources consumed (water and electricity) have been obtained in collaboration among IMAMOTER C.N.R. and partners of the FLEXICAST project. The meetings among the involved partners have been necessary to identify the useful data to collect in order to perform the LCA analysis of the whole manufacturing process. Then the data have been elaborated and implemented within the GaBi software.
A complete Life Cycle Assessment was developed to compare the impact, in term of Eco Indicators, of the solutions introduced (and developed) during the project. The LCA presented in these pages is probably the first example of a complete environmental impact analysis of the cast iron production cycle. Life cycle assessment was included in deliverable D7.7.
WP8: Exploitation Plan. Dissemination and training
WP Leader: EURECAT/CIMNE
MAIN TASKS:
It included the following tasks: (i) Exploitation plan, (ii) Socio-economic aspects and (iii) Dissemination and training activities.
The activities carried out under the scope of the FLEXICAST exploitation plan were:
1.-FLEXICAST exploitable results
2.-Specific Products (results) Portfolio
3.-Exploitation plan partner by partner
4.-Future Plan (Short term) for each exploitable result
5.-Exploitation plans by scenarios
6.-Exploitation plan of F. Roda (as a Demo partner)
Also, all Flexicast partners consortium met in Modena (TSS facilities) in order to assist a seminar tutored by Mr. Dario Mazella from ESIC (Brussels). It was a great help for all partners in order to fulfil the Flexicast exploitation plan. The Exploitation plan was summarized in D8.4.
The economic impact methodology chosen was to evaluate the possible impact in the European foundries who decided to incorporate the FLEXICAST achievements, and for that the impact in several “type” foundries incorporating several combinations of FLEXICAST achievements has been considered. (What here are presented as “scenarios”). Some major results are exposed in the coming pages.
In relation with dissemination activities, the continuous scientific work of all partners did lead to numerous scientific publication and media coverage. The publications covered scientific topics from metallurgic & mechanical engineering to robotics applications.
1. Relevant results publications of the research activity carried out in peer-review journal
(20 papers)
2. Publication of relevant results coming from the research activity in Conference proceedings
(9 publ.)
3. Oral/ Poster presentations of relevant results of the research activity in Intern. Congresses
(26 oral presentations)
4. Patents (1)
5. Workshops related with FoF:
a. Impact of the Factories of the Future PPP of relevant results of the research activity in EU projects. (4 editions)
b. Let's 2014 (1 edition)
6. Attendance to Industrial oriented Fairs either as invited guess or expositor
7. Publication articles in different newspapers or e-news web-sites
Also, all these dissemination and training activities are exposed in the coming pages.
WP9: Financial and administrative management
WP Leader: UPC-LABSON
MAIN TASKS:
The goals of the project management and coordination team were to:
• Facilitate communication and integration between the partners and disseminate information about the project to the wider community
• Identify and resolve disputes between partners
• Keep the project on track and ensure all deliverables, periodic and final reports are produced on time
• Set up a project office with finance and administration support
• Carry out annual audits of all partners
• Liaise with the EC
The deliverable D9.1 summarizes the major tasks undertaken in WP9 for the management and coordination of the project, together with a summary of the overall project management details in terms of deliverables, dissemination outcomes and key meetings
Communication with the European Commission
UPC -LABSON has been the intermediary for any communication between the EC and partners during this reporting period.
Administration of the financial contribution
The Community financial contribution was carefully managed regarding its allocation between partners and activities taken by the consortium. All payments were made without delay and records have been kept in order to determine at any time of what portion of the Community financing has been paid to each beneficiary. The Commission was informed of the distribution of the contribution. The cost claim process was carried out through the NEF system. The planning of resources has been done so as to accommodate a realistic distribution of resources, based on the on-going project situation.
Communication and Reporting
Communication within the consortium has been ensured through the normal means such as electronic mail, teleconferences and in-person meetings. Technical documentation generated by the Project have been exchanged in electronic format and made available to partners. To monitor the status of the project, internal reports have been prepared every 6-month period. A consortium mailing list service has been created.
Legal and administrative issues
The Project Management Team (PMT) has established the rules for the access and exploitation of the background of the individual partners and of foreground.
Organization of project meetings
UPC-LABSON team and Dr. Pedro Roquet as technical manager took care of the logistics and organization of all Flexicast project meetings, ensuring that all partners took part according to their role and responsibilities.
Potential Impact:
The foreseen socio-economically impacts that may mark trends in the next future, where FLEXICAST project has contributed to highlighting, have been described in the following paragraphs.
1.4.1 Towards a generation of high qualified staff.
After four years in the Flexicast project, it has become clear that one of the greatest potential impacts was the generation of high-qualified staff.
As an example, the "knowledge" and "know how" were basic achievements (as tools) for reducing the rejects and increasing box efficiency. (In the D7.5 D7.6 D8.5 and D8.9 where it´s analysed the profitability of project achievements, the "knowledge and know how” were one of the most profitable with a lower investment. (In our case, investment means money. Educational investment is quite another thing).
1.4.2 ICT’s in the production line.
Information and communications technology (ICT) is an extended term for information technology (IT) which stresses the role of unified communications and the integration of telecommunications (wireless signals), computers as well as necessary enterprise software, middleware, storage, and audio-visual systems, which enable users to access, store, transmit, and manipulate information.
One of the most important outputs was the introduction of ICT in the Flexicast-DEMO cast iron manufacturing cell.
DTA, for example, directly enters in the first line of production helping directly in the decision- making. And even that it´s a use friendly program, knowledge in metallurgy and mathematic tools are needed.
Foundry is an industry full of sensors and data. All of them are needed in some moment to take decisions. FLEXICAST has pointed out the need of having them interconnected, centralized, and analysed, and even that computer codes may manage them, the last decision must be human, not only due to safety reasons, but also because correlations, cause-effect reasons may change along the time. Foundry staff must be capable of understanding and managing them.
1.4.3 Wireless sensors
Foundry process manages a very high quantity of data, not only those directly connected with the metallurgical process, such as chemical analysis, DTA analysis, temperature, etc., but also much others data from equipment denominated “automatic auxiliary processes”, such as sand parameters, sand temperatures, sand permeability, .., air pressure, etc.
Foundries, like other process factories, evolve towards the massive use of sensors. But there are many cases where the physical connection between sensors and computers is impossible. By other side, extremely expensive foundry equipment should be a long-period replacement, then it must be there for years and years. It needs an intensive maintenance but also needs usually an updating and/or new technology.
The research with wireless connections made during the FLEXICAST showed that this technology, considered as "impossible" at the beginning of the project, (mainly due to electrical noise problems), it is now fully operative.
Until now, all the experiments to get data about the piece cooling along the moulding box carrousel must be done out of the production line. Thanks to the wireless technology developed in FLEXICAST, now it is possible to perform the tests and monitoring in the production line, which makes the process more convenient and less disturbing.
1.4.4 Remote diagnosis and treatment support (intranet and internet)
Nowadays, an important trend in the foundry industry, is the possibility of be connected to a centre where the data is received, analysed and there is the possibility of interacting remotely with the application. That is, not only it is possible to read and have access to the data “in situ” but also it is possible to program and remotely enhance the application.
In FLEXICAST, following the actual trends, DTA and X-ray are linked via internet with Proservice and InnospeXion partners respectively, and then it is now possible:
a.- To check if there is some malfunction and repair it remotely.
b.- Update the program(s), remotely, maintaining fully actualized.
c.-Analyse the data, interpret it and assist to the end-user in order to manage and to take decisions.
Given the fact that this “data centre” is linked to many foundries, each of them, with his own or similar problems, they will be the most skill and prepared staff, able to use and interpret the results.
1.4.5 Towards a global information structure
Taking account the trends mentioned in 1.4.2 and 1.4.5 there is no doubt that the staff of the control centre will have wide and global information of the foundry processes. This knowledge and know how will have a potential impact to supply an important and profitable technical support to the foundry (or any other factory).
1.4.6 Privacy conflicts
In some circumstances, the above mentioned items, may create a question: How to profit the skills and experiences of the “control centre staff" without generating problems among foundries competitors? This is an open question. This conflict is not new, but was limited and bounded before Internet era.
1.4.7 Needs of cast iron related educational programs and continuous training courses
The FLEXICAST project pointed out the need of highly skilled people, with proper knowledge that may not be incorporated only through personal experience or in house training. Unfortunately, few universities give engineering courses incorporating cast iron knowledge and technologies.
A complementary way is acquiring knowledge by continuous training courses, incorporating not only metallurgical knowledge, but also other background specially devoted to cast iron technologies.
1.4.8 Needs to encourage young people to study and learn cast iron technologies
From both sides: foundries and end users.
FLEXICAST project points out the need of a highly skilled staff. It has already denounced as a need from EU Industry. Pointing out the fact that skilled and experienced people in the foundry industry are going to retire in a near future and no young replacement is foreseen. It is necessary to encourage young people to study and learn cast iron technologies
1.4.9 Better design and use of the cast iron products
There is a need and demand from customers of parallel services to assist them for a better design and use of the cast iron products.
During the development of WP5, in particular, when some partners worked about mechanical simulation of fatigue resistance of cast pieces under stress, working in S.N curves linked with hardness, microstructure, tensile strength, and in the importance of clear and realistic boundary conditions, FLEXICAST consortium detected a lack of skilled people capable in the field of machinery design using cast iron pieces.
In this line, design guidelines and best practice rules focused on the best design of cast iron pieces have been delivered (D7.6 and D7.5).
Figure nº 22.- In deliverable D7.6 a guide of process to follow by customer and foundry was proposed following the trend to train and give a global service to customer, profiting the new knowledge acquired during FLEXICAST
Figure nº 23.- A protocol to classify the pieces by “families”, and defining the expected characteristics (mechanical and level of defects and type), has been built up to indicate the end user and foundry the expected characteristics of a new parts, following the trend of giving global services to customer.
1.4.10 Need of increasing the pre normative and normative reports and knowledge.
Traditionally, standards are more devoted to Quality control aspects such as rejecting or accepting pieces and reports, that to help in the technological design of cast iron pieces.
Figure nº 24.- Shows that FLEXICAST research goes towards this aim to complement the information given in the ISO standards
1.4.11 Economic aspects
To consider the impact of the achievements of FLEXICAST project, it has been considered which will be the impact in a “typical” foundry that incorporates several of them. This combination of achievements installed in a foundry is what called here “scenario”. The different scenarios have been defined in figure nº 1 but, for more clarify, it is repeated below as figure nº 25.
Figure nº 25.- Different scenarios analysed (combined results)
To cover a wide espectra of possible european foundries size, the simulation has been done for 20.000 15.000 and 10.000 Ton of yearly brut production. Figures nº 26, 27 and 28 show the foreseen economic impact in a “type foundry", thanks to the implementation of different FLEXICAST project achivements.
Figure nº 26.-Cost reduction, Needed Investment, and Final Result for the scenarios (20.000 Ton)
Figure nº 27.- Cost reduction, Needed Investment, and Final Result for the scenarios (15.000 Ton)
Figure nº 28.-Cost reduction, Needed Investment, and Final Result for the scenarios (10.000 Ton)
Note, that the investment has been calculated for a 10 years period of recovery, so the result is what normally is understood as “result” prior to taxes. It is supposed that the data presented will be effective once the achievements have been fully incorporated, (not before a minimum of 1 year, as a general rule, and depending on the degree of integration).
These are the economic impacts obtained through 4-years life of FLEXICAST project. These results are in line with the trends of the European Industry and, also, with the perspectives elaborated by other prestigious institutions, such as the European Foundry Industry Association , and Deutche Industry Bank, between others.
2 USE AND DISSEMINATION OF FOREGROUND
2.1 SECTION A
2.1.1 Publications
1. Ignaszak Z., Popielarski P., Hajkowski J., J-B. Prunier, Problem of acceptability of internal porosity in semi-finished cast product as new trend - "Tolerance of damage" present in modern design office, Defect and Diffusion Forum, Vol. 326-328, pp.612-619 2012.
2. Ignaszak Z., Hajkowski J., Popielarski P. – Mechanical properties gradient existing in real castings taken into account during design of cast components, Defect and Diffusion Forum, Vol. 334-335, pp.314-321.
3. Ignaszak Z., Popielarski P., Hajkowski J., – Sensitivity of models applied in chosen simulation systems with respect to database quality for resolving the casting problems, Defect and Diffusion Forum, 2013, Vol. 336, pp.135-146.
4. R. Sika, Z. Ignaszak, M. Rogalewicz, The specificity of applications of the Statistical Process Control methods in a complex production process for example foundry, Polish Metallurgy in 2011-2014, Monograph, Publish. Akapit, Edit. Officer K. Swiatkowski, Krynica-Zdroj (Poland), 2014, pp.357-371.
5. Ignaszak Z., Codina E., Hajkowski J., Popielarski P., Roquet P., Selected aspects of the use of thermal & derivative analysis to improve macro modelling of solidification for casting iron. Comparison of chosen simulation codes, 10th International Symposium on the Science and Processing of Cast Iron – SPCI10, Mar del Plata (Argentina) 2014.
6. R. Sika, Z. Ignaszak, Scada systems and theirs connection with Statistical Process Control in Foundry, Archives of Foundry Engineering, Trzebnica (Poland), 2015, Vol.15 Special Issue 1/2015, pp.145-153.
7. Z. Ignaszak, R. Sika, M. Perzyk, A. Kochanski, J. Kozlowski, Effectiveness of SCADA systems in control of green sands properties, Archives of Foundry Engineering, Busko Zdroj (Poland), 2016, Vol.16 Issue 1/2016, pp.5-12.
8. Ignaszak Z., Popielarski P., Hajkowski J., Codina E., Methodology of Comparative Validation of Selected Foundry Simulation Codes, Archives of Foundry Engineering, vol. 15/4 s.37–44, Gliwice (Poland), 2015.
9. Ignaszak Z., Prinier J-B., Effective Laboratory Method Chromite Content Estimation in Reclaimed Sands, Archives of Foundry Engineering, Vol. 16/3 pp.162 166, Gliwice - Łódź (Poland), 2016.
10. E. Villagrossi, L. Simonib, M. Beschia, N. Pedrocchi, A. Marini, L. Molinari Tosatti, A. Visioli, A Virtual Force Sensor for Interaction Tasks, Submitted to: Robotics and Autonomous Systems
11. E. Villagrossi, C. Cenati, N. Pedrocchi, M. Beschi, L. Molinari Tosatti. Robotized Solution for Cast Iron Deburring. Submitted to: International Journal of Advanced Manufacturing Technology.
12. PJ Gamez-Montero, R Castilla, J Freire, M Khamashta and E Codina, An empirical methodology for prediction of shape and flow rate of a free-falling non-submerged liquid and casting iron stream, Advances in Mechanical Engineering, 2016, Vol. 8(9) 1–12 (2016) DOI: 10.1177/1687814016669635
13. M. Chiumenti, M. Cervera, N. Dialami, B. Wu, J. Li and C. Agelet de Saracibar, Numerical modelling of the electron beam welding and its experimental validation, Finite Elements in Analysis and Design, 121 (2016) 118-133, http://dx.doi.org/10.1016/j.finel.2016.07.003i
14. N. Lafontaine, R. Rossi, M. Cervera and M. Chiumenti, Formulación mixta estabilizada explícita de elementos finitos para sólidos compresibles y quasi-incompresible, Revista Internacional de Métodos Numéricos para Cálculo y Diseño en Ingeniería, 169 (2015), http://dx.doi.org/10.1016/j.rimni.2015.09.003
15. M. Cervera, M. Chiumenti, L. Benedetti and R. Codina. Mixed stabilized finite element methods in nonlinear solid mechanics. Part III: Compressible and incompressible plasticity, Computer Methods in Applied Mechanics and Engineering, 285 (2015) 752-775
16. M. Chiumenti, M. Cervera and R. Codina., A mixed three-field FE formulation for stress accurate analysis including the incompressible limit, Computer Methods in Applied Mechanics and Engineering, 283 (2015) 1095-1116,
17. C. Agelet de Saracibar, M. Chiumenti, M. Cervera, N. Dialami and A. Seret, Computational modelling and sub-grid scale stabilization of incompressibility and convection in the numerical simulation of friction stir welding processes , Archives of Computational Methods in Engineering, 21 (2014) 3-37
18. S. Biswas, F. Sket, M. Chiumenti, I. Gutiérrez-Urrutia, J.M. Molina-Aldareguía and M.T. Pérez-Prado, Relationship between the 3D porosity and ß-phase distributions and the mechanical properties of a high pressure die cast AZ91 Mg alloy, Metallurgical and Materials Transaction A, 44(9) (2013) 4391-4403.
19. C. Agelet de Saracibar, R. López, B. Ducoeur , M. Chiumenti and B. de Meester, Un Modelo Numérico para la Simulación de Disolución de Precipitados en Aleaciones de Aluminio con Endurecimiento utilizando Redes Neuronales, Revista Int. de Métodos Numéricos para Cálculo y Diseño en Ingeniería, 29(1) (2013) 29–37.
2.1.2 Conferences proceedings
1. R.Sika Z.Ignaszak J.Kolakowski E.Codina P.Roquet SPC procedures for the manufacturing and complex unstable processes, Proceedings of Conference "Manufacturing 2014", 08-10.12.2014 Poznan, Poland.
2. P.Popielarski J.Hajkowski J.Matwiejczuk Z.Ignaszak. Comparative validation of selected simulation systems, Proceedings of Conference "Manufacturing 2014", 08-10.12.2014 Poznan, Poland.
3. Ignaszak Z., Popielarski P., Effective modelling of phenomena in over-moisture zone existing in porous sand mould using the simplified simulation systems applied in foundry, Proceedings of 10th International Conference on Diffusion in Solids and Liquids: DSL 2014, Paris, France, 23–27 June, 2014. publication in DDF by Scitec Publications Ltd (a subsidiary of Trans Tech Publications Ltd), Brandrain 6, CH-8707 Uetikon Zurich, Switzerland).
4. Ignaszak Z., Popielarski P., Effective modelling of phenomena in over-moisture zone existing in porous sand mould using the simplified simulation systems applied in foundry, Proceedings of 10th International Conference on Diffusion in Solids and Liquids: DSL 2014, Paris, France, 23–27 June, 2014. publication in DDF by Scitec Publications Ltd (a subsidiary of Trans Tech Publications Ltd), Brandrain 6, CH-8707 Uetikon Zurich, Switzerland).
5. Ignaszak Z., Popielarski P., Effective modelling of phenomena in over-moisture zone existing in porous sand mould using the simplified simulation systems applied in foundry, Proceedings of 10th International Conference on Diffusion in Solids and Liquids: DSL 2014, Paris, France, 23–27 June, 2014. publication in DDF by Scitec Publications Ltd (a subsidiary of Trans Tech Publications Ltd), Brandrain 6, CH-8707 Uetikon Zurich, Switzerland).
6. Ignaszak Z., Popielarski P., Identification of the moisture distribution in the intensively heated porous sand mould, DSL MUNICH 2015, 11th International Conference on Diffusion in Solids and Liquids 22 –26 June, 2015.
7. Ignaszak Z., Hajkowski J., Popielarski P., Gapiński B., Identification of the local properties of Al alloys high pressure die casting and cast iron products with validation of real discontinuities by computer tomography, DSL 2016 Split, 12th International Conference on Diffusion in Solids and Liquids26 – 30 June, 2016.
8. E. Villagrossi, G. Legnani, N. Pedrocchi, F. Vicentini, L. M. Tosatti, F. Abba and A. Bottero. Robot dynamic model identification through excitation trajectories minimizing the correlation influence among essential parameters. 11th International Conference on Informatics in Control, Automation and Robotics (ICINCO), pages 475-482, 2014. Doi 10.5220/0005060704750482.
9. M. Beschi, E. Villagrossi, N. Pedrocchi and L. M. Tosatti. A general analytical procedure for robot dynamic model reduction. IEEE-RSJ International Conference on Intelligent Robots and Systems (IROS), pages 4127-4132, 2015. doi 10.1109/IROS.2015.7353960.
2.1.3 Congresses
1. Ignaszak Z., 55th International Scientific Conference "Solidification and Crystallization of Metals 2014" 15th-17th September, 2014, Kielce-Mąchocice Kapitulne.
2. Ignaszak Z., Reporting Conference of Committee Metallurgy at Poland Academy of Sciences „Metallurgy 2014”, 15th-18th October 2014, Krynica-Zdrój.
3. Ignaszak Z., 10th International Symposium on the Science and Processing of Cast Iron – SPCI10, Mar del Plata (Argentina) 2014
4. Ignaszak Z., IV International Scientific and Technical Conference "Manufacturing 2014", 08 10.12.2014 Poznan, Poland.
5. Ignaszak Z., 10th International Conference on Diffusion in Solids and Liquids DSL 2014, Paris, France, 23–27 June, 2014.
6. Ignaszak Z., XV International Scientific Conference „Zapewnienie Jakości W Odlewnictwie 2015”, 20-22nd May 2015, Trzebnica.
7. Ignaszak Z., 11th International Conference on Diffusion in Solids and Liquids, DSL 2015 Munich, 22nd–26th June, 2015.
8. Ignaszak Z., 56th International Scientific Conference "Solidification and Crystallization of Metals 2015" 2nd-4th September, 2015, Busko-Zdrój.
9. Ignaszak Z., 12th International Conference on Diffusion in Solids and Liquids, DSL 2016 Split, 26th–30th June, 2016.
10. Ignaszak Z., XVI International Scientific Conference - "Optimization of Production Systems in Foundries" 14th-16th June, 2016, Łódź.
11. Ignaszak Z., 57th International Scientific Conference "Solidification and Crystallization of Metals 2016" 20th-22nd September, 2016, Kielce-Cedzyna.
12. Ignaszak Z., A General Analytical Procedure for Robot Dynamic Model Reduction, M. Beschi, E. Villagrossi, N. Pedrocchi and L. Molinari Tosatti, 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Congress Centre Hamburg, Sept 28 - Oct 2, 2015. Hamburg, Germany
13. N. Pedrocchi, E. Villagrossi, F. Vicentini and L.M. Tosatti. Robot-dynamic calibration improvement by local identification. IEEE International Conference on Robotics and Automation (ICRA), pages 5990-5997, 2014. doi 10.1109/ICRA.2014.6907742.
14. E. Villagrossi, G. Legnani, N. Pedrocchi, F. Vicentini, L. M. Tosatti, F. Abbá and A. Bottero. Robot dynamic model identification through excitation trajectories minimizing the correlation influence among essential parameters. 11th International Conference on Informatics in Control, Automation and Robotics (ICINCO), pages 475-482, 2014. Doi 10.5220/0005060704750482.
15. M. Beschi, E. Villagrossi, N. Pedrocchi and L. M. Tosatti. A general analytical procedure for robot dynamic model reduction. IEEE-RSJ International Conference on Intelligent Robots and Systems (IROS), pages 4127-4132, 2015. doi 10.1109/IROS.2015.7353960
16. A. Bonanno, M Merlin, GL Garagnani, E Codina, Numerical analysis and optimization of a lamellar cast iron hydraulic distributor. AIAS 2015-546 Associazione italiana per l’analisi delle sollecitazioni (AIAS). 44º convegno nazionale, 2-5 sept 2015. Univ. Di Messina
17. High-fidelity numerical simulation of additive manufacturing processes, M. Chiumenti, M. Cervera, C. Agelet de Saracibar, N. Dialami, W. Huang, X. Lin, L. Wei, Y. Zheng, L. Ma, COMPLAS 2015, 13th International Conference on Computational Plasticity, Barcelona, Spain, 1-3 September 2015
18. High-fidelity numerical simulation of additive manufacturing processes, M. Chiumenti, W. Huang, X. Lin, M. Cervera, C. Agelet de Saracibar, N. Dialami, M. Liang, L. Wei, MCWASP 2015, Modelling of Casting, Welding and Advanced Solidification Processes XIV, Awaji island, Hyogo, Japan, 21-26 June 2015
19. Numerical simulation of shaped metal deposition processes close to the industrial manufacturing technology, M. Chiumenti, W. Huang, X. Lin, L. Wei, M. Cervera, C. Agelet de Saracibar, N. Dialami, PLASTICITY 2015, 20th International Symposium on Plasticity and its Current Applications, Montego Bay, Jamaica, 4-9 January 2015
20. Coupled thermo-mechanical finite element technology for stress accurate analysis, M. Chiumenti, M. Cervera, R. Codina, C. Agelet de Saracibar, WCCM 2014, 11th World Congress on Computational Mechanics, Barcelona, Spain, 20-25 July 2014.
21. Stress accurate framework for the numerical simulation of FSW processes, M. Chiumenti, N. Dialami, M. Cervera, C. Agelet de Saracibar, FSWS 2014, 10th International Friction Stir Welding Symposium, China National Convention Centre, Beijing, China, 20-22 May 2014
22. Simulación numérica del proceso de soldadura por haz de electrones, M. Chiumenti, N. Dialami, M. Cervera, C. Agelet de Saracibar, CAIP 2013, 11º Congreso Interamericano de Computación Aplicada a la Industria de Procesos, Lima, Perú, 21-24 Octubre 2013
23. A novel stress-accurate FE technology for the numerical simulation of the FSW process, M. Chiumenti, N. Dialami, M. Cervera, C. Agelet de Saracibar, COMPLAS 2013, 12th International Conference on Computational Plasticity, Barcelona, Spain, 3-5 September 2013
24. A novel stress-accurate FE technology for highly non-linear analysis with incompressibility constraint. Application to the numerical simulation of the FSW process, M. Chiumenti, M. Cervera, C. Agelet de Saracibar, N. Dialami
25. Finite element modelling of EB welding process, M. Chiumenti, M. Cervera, C. Agelet de Saracibar, COUPLED PROBLEMS 2013, 5th International Conference on Coupled Problems in Science and Engineering, Ibiza, Spain, 17-19 June 2013
26. E. Codina and M. Khamashta, The cast iron technology innovations for fluid power components, MECSPE PARMA 2016, Power Drive Innovation, Parma, 18 March 2016
2.1.4 Workshops related with FoF-PPP
-Let's 2014: Bringing nanotechnology research into the European Production Industry (Bologna-Italy)
-Impact of the factories of the future PPP
(Brussels)
• 24-25 March 2014
• 29-30 April 2015
• 14-15 April 2016
•
2.1.5 Industrial fairs attendance
1. Kielce Trade Fair, 20th International Fair of Technologies for Foundry METAL, 16 18 September 2014 – PUT: participation as guest.
2. Kielce Trade Fairs - 23rd International Defence Industry Exhibition MSPO, 1 4th September, 2015 – PUT: participation as guest.
3. Kielce Trade Fair, 21st International Fair of Technologies for Foundry METAL, 20 22 September 2016 – PUT participation as guest.
4. MESIC-2015 Fair, Barcelona, Spain, July 24, 2015 – EURECAT: participation as guest
5. Innovation, Connect & Transform, ICT-2015, Lisbon, Portugal, October 20-22, 2015 – EURECAT participation as guest
6. 13th International Foundry Trade Fair with Technical Forum – GIFA, Dusseldorf, Germany, 10-20 June 2015 – UPC: participation as guest
7. 13th International Foundry Trade Fair with Technical Forum – GIFA, Dusseldorf, Germany, 10-20 June 2015 – Fundiciones de Roda: participation as guest
8. 13th International Foundry Trade Fair with Technical Forum – GIFA, Dusseldorf, Germany, 10-20 June 2015 – ONDARLAN- INDUCTOTHERM: fair expositor
9. 13th International Foundry Trade Fair with Technical Forum – GIFA, Dusseldorf, Germany, 10-20 June 2015 – EURECAT: fair expositor
10. 13th International Foundry Trade Fair with Technical Forum – GIFA, Dusseldorf, Germany, 10-20 June 2015 – FOSECO: fair expositor
11. 13th International Foundry Trade Fair with Technical Forum – GIFA, Dusseldorf, Germany, 10-20 June 2015 – PROSERVICE: fair expositor
12. MECSPE- Technologie per l'innovacione. 16 march 2016 Parma (Italy). LABSON-UPC (Prof E.Codina and Prof. M Khamashta)
2.2 Section B -PART B1
2.2.1 Patents
Ignaszak Z., Popielarski P., Hajkowski J., Poznan University of Technology, Measuring system of time and rate of casting mould cavity pouring, Application at the Polish Patent Office No. P.415693 2015.
List of Websites:
Consortium
The FLEXICAST consortium comprises 14 partners from 5 countries, having a complimentary knowledge, and covering the complete R&D chain:
2 Universities, 4 Research Centers, 4 SME’s, and 4 Industry (large)
Contact the coordinator for further information:
Dr. Ing Ind. Esteban Codina Maciá
Universitat Politecnica de Catalunya, UPC)
LABSON- Research Center
University Campus of Terrassa
E-mail: ecodina@mf.upc.edu
Tel +34 93 739 86 64
Web: www.flexicast-euproject.com
Documentos relacionados
- final1-final-report-fig-n-5.pdf
- final1-final-report-fig-n-24.pdf
- final1-final-report-fig-n-25.pdf
- final1-final-report-fig1-and-fig-2.pdf
- final1-final-report-fig-n-10.pdf
- final1-final-report-fig-22-and-23.pdf
- final1-final-report-fig-n-6.pdf
- final1-final-report-fig-18-and-19.pdf
- final1-final-report-fig-n-16.pdf
- final1-final-report-fig-n-9.pdf
- final1-final-report-fig-n-26.pdf
- final1-final-report-fig-4.pdf
- final1-final-report-fig-n-21.pdf
- final1-final-report-fig-n-8.pdf
- final1-final-report-fig-n-15.pdf
- final1-final-report-fig-n-14.pdf
- final1-final-report-fig-n-7.pdf
- final1-final-report-fig-n-17.pdf
- final1-final-report-fig-n-12-and-13.pdf
- final1-final-report-fig-27-and-28.pdf
- final1-final-report-flexicast-v-0-ecm.pdf
- final1-final-report-fig-n-20.pdf
- final1-final-report-n-11.pdf
- final1-flexicast-fig-n-3.pdf
