Skip to main content

Adaptive Control for Metal Cutting

Final Report Summary - ADACOM (Adaptive Control for Metal cutting)

Executive summary:

The ADACOM project is a European collaborative project of the 7th Framework Programme. The aim of the project is to develop a generic modular adaptive control platform that will allow metal cutting processes to respond to changing circumstances by combining technologically advanced sensor systems, process adaptation strategies and actuator systems.

Presentation of the project and the consortium
The ADACOM project is a European collaborative project of the 7th Framework Programme within Theme 4, Nanosciences, Nanotechnologies, Materials and New Production Technologies. It is a four year project that started on October 2008 and will last until October 2012. A consortium of 12 partners coming from different fields of the technological world works on the project: 4 world leading OEMs (DAIMLER, FIAT, Heidelberg Druckmaschinen and BOSCH), 4 SMEs (LOLA, DIAD, ACTARUS, TEKS), the world leading sensor manufacturer (KISTLER) and 3 outstanding higher education centres (WZL-RWTH Aachen, Trinity College Dublin and Mondragon University) from 7 European countries.

Project Context and Objectives:

Background and context
The ADACOM project aimed to develop a generic modular adaptive control platform that will allow metal cutting processes to respond to changing circumstances by combining technologically advanced sensor systems, process adaptive control strategies and actuator systems. The consortium was a unique collaboration of some of Europe's leading industries as well as RTD performers in the field of process monitoring and control. One of the unique aspects of ADACOM is the industrial focus whereby a broad range of industry sectors including Automotive (DC, LOLA, CRF); Industrial technologies (BOSCH); Printing machine manufacturer (HDM) have defined common challenges which ADACOM targeted.

From a technological perspective the project set out to target the following goals:
1) The development of Adaptive Machining Systems for difficult metal cutting operations such as milling, turning, gun-drilling, and grinding, which consist of sensor and actuator systems for online manufacturing control related to the part quality. The major aim is to achieve an Online Quality Control system to satisfy demands for mass customisation and small batch production of gear wheel prototypes or drop-forge dies of Daimler and for the Heidelberger Druckmaschinen – applications of the above mentioned industrial sectors. In ADACOM, this aim has been achieved by separating the single components of an adaptive machining system into a modularised system. This modularisation approach promotes adaptation within the different modules to enable flexible use of the adaptive system for different machining operations in different end user sectors.

2) The next goal is the realisation of one central system for different processes, which incorporates the flexibility towards changing production situations, different machine tools and different work piece materials. This flexibility of the ADACOM-System has been reached by developing a standardisation of the Adaptive System components. Issues such as standardised interfaces, which guarantee "Plug and Produce"-production, have been addressed by standard data formats throughout the whole production sequence.

The specific Science and Technology (S&T) objectives of the ADACOM project are:
- To increase tool life by 50%
- To reduce waste such as swarf and coolants by 80% and power consumption by 50%
- To reduce machining noise to less than70dBA
- To increase machining speeds for hardened steels and aluminium
- To increase throughput by 50%
- To reduce automatic and manual finishing operations by 80%
- To reduce lead times by 20%
- To ensure best-practice cutting methodologies for new materials based on a developed intelligent adaptive system which ensures the maximum performance of machining processes
- To develop adaptive and knowledge based integrated systems for high volume as well as small batch machining operations
- To ensure stable and part quality oriented production within the automotive, electronic and printing industry.

Overview of ADACOM approach
Due to the experience of the partners and good collaboration within the consortium in defining the scope even at the proposal stage, it was already specified that the adaption mechanisms would be based on four main approaches as shown in the figure below and the interaction demonstrates that the solutions are not mutually exclusive.

- Strategy Adaptation
- Position Adaptation
- Parameter Adaptation
- Machine State Adaptation

Project Results:

1 Introduction
As the project is called ADACOM – Adaptive Control for Metal Cutting – the structure of the S & T report is based on the development of adaptive control strategies

In the first part, several adaption strategies that were identified within the project duration are presented in a clustered manner and their differences, advantages and restrictions are explained. The adaption strategies and their functionality describe the " Adaptive machining System". The functional units as identification, decision, and modification parts of the different strategies are highlighted. On this, the required hardware (WP2,3) and software (WP4) for adaptive control are presented. The hardware is divided to the control and data processing units (WP2) and the sensor and actuator systems (WP3). The software developments (WP4) are described in a clustered way of development emphasis. At the end of the report detailed information about the industrial case studies are presented that were implemented within ADACOM for the development, evaluation and demonstration of the different kinds of adaption. The industrial case studies were selected by the industrial end users of the project and are machining challenges from their productions. At the end of the project, all case studies resulted in a machining demonstrator that also was used for LCA, eco-profile and energy consumption evaluations of the different kind of machining systems. The result of the LCA, environmental and social aspects are presented in a special part of the final report.

2 Adaption System and Strategies
The inputs and outputs are differentiated to low dynamics and high dynamics. With regards to the outputs, the kind adaption is categorized: position, parameter, machine state or strategy adaption. Of course, the kinds of adaption can be combined to a more complex adaption system.

For the implementation of adaptive systems, in ADACOM different control strategies were identified, which possibility of use depends on the requirements and boundary conditions of the metal cutting process.

The strategies can be organized in two main groups:
- continuous closed loops
- iterative closed loops
In the following section these groups with their sub modes are described.

2.1 Adaptive control strategies with continuous closed loop control
The continuous closed loop control attempts to in real time automatically maintain the control variable at a specific set point based on information from the measurement system by manipulating the actuating variable through the controller. In this control strategy three different modes have been identified:
- Static control parameters
- Parameter adaption with continuous plant identification
- Parameter and process model adaption

In the control strategies, two loops are employed:
- The inner loop or feedback loop: provides information about the process state. This loop is essential for the control task and therefore is present in all strategies, it would be however be modified for processes where iteration are necessary as will be explained in the next section.
- The outer loop or adaption loop: provides the feedback for the controller parameters. This loop is used in the continuous adaptive control strategies with continuous plant identification and process model adaption.

2.2 Adaptive control strategies with iterative closed loop control
Very common adaption strategies found to be applied in the metal cutting industry implement actually not a continuous feedback, but instead an iterative closed loop control. These strategies used the real process state or control variable assess through the measurement system and evaluate it against the set point. When the control variable differs from the desired output, then the process parameters are changed and the process executed again iteratively until the output and the set point concur.

Three different iterative closed adaption strategies have been identified:
- Automatic closed loop with online measurement
- Automatic closed loop with offline measurement
- Human closed loop with offline measurement

3 ADACOM Control Hardware
The control hardware is the central unit being responsible for the recording and transformation of sensor signals to digital data, the hosting of the control software and the output of the manipulated variable and state information. For the connectivity of different kinds of sensors, the implementation of control strategies and the machine adequate output of manipulated variable, the ADACOM control hardware have to base on a flexible system architecture. The identified solution for the ADACOM project is a modularized architecture with slots for different input and output modules as also a high performance computing unit that can be programmed by user friendly software tools on normal desktop PC's. Caused by the status of prototyping and development and the unknown requirements with regards to the performance and connectivity, for the ADACOM case studies, the consortium used the integrated software and hardware platform of National Instruments in combination with some MATLAB programming tools of the mathworks.

4 Sensors and actuators (3)
In the section before, the ADACOM Hardware systems and their functionality were described. The ADACOM hardware systems include all functional hardware with exception of the sensors and actuators. For these, a separate work package was responsible to identify suitable sensor solutions, describe their properties, develop new kinds of sensor application and combination for the identification of process state and identify suitable kinds of actuate to the process.

4.1 Sensors
The ADACOM project used commercial sensors as also self-developed sensors systems for the identification of process state variables. In the following a short list of used commercial sensors is presented. Afterwards the results of ADACOM developments of sensor systems are described.
For cutting processes, the groups of piezoelectric and strain gauge based sensors are the most common ones. In the ADACOM project the most often used sensors are the piezoelectric ones. "Kistler instruments", as project partner, is an expert in developing piezoelectric based solutions as acoustic emission, accelerometers and force dynamometers.

AE Sensor:
The used AE sensors are products of the project partner Kistler. The sensor system consists of the sensor itself and a coupler. The coupler has integrated filter, amplifier and time constant components, which can be changed by plug in. By this, the signal preprocessing can be adapted to the measurement task. The system components are changeable within the Kistler product line, interchangeability with product from other manufacturers is not known.

The used accelerometers in this project are piezoelectric based sensors with IEPE standard, also called ICP standard. A common sensor setup of acceleration measurement consists of the sensor itself with integrated charging amplifier and the IEPE coupler being responsible for the power supply of the sensor. The standardization of the power supply with the IEPE standard led to the development and sell of data acquisition cards that have integrated IEPE power supplies.

Force dynamometer:
The Force measurement is done with dynamometers from the project partner Kistler. The Dynamometers exist in two main variants: static and rotating dynamometers.

The static dynamometers exist in different variation and different sizes which can be used versatile for different application such as turning, drilling or other machining operations like milling. The used sensor system setup in ADACOM is modular and consists of the dynamometer itself and a charge amplifier.

Measurement microphones are modular sensing devices consisting of the microphone itself, the microphone preamplifier and the condition amplifier.

Temperature Sensor:
The temperature measurement is done by a measurement system of our Project partner Actarus. For the measurement system different kind of sensor designs for the processes turning, milling and drilling exist. For data transfer of rotating sensors, a telemetry system is used. The systems have a pre-processing unit with analogue signal output of 0…10V. The systems are unique prints. Because of that, no interchangeability with systems of other manufacturers is able.

Chip deflector
Within the project, a special chip deflector for the gun drilling was developed to analyze the quality of the chip transportation.

4.2 Actuator systems:
In ADACOM it is not foreseen to develop new actuator systems, but consistently use existing interfaces and systems to initiate a machine reaction. The both most easy to use alternatives are the use of inputs of the NC controller and the manipulation of the NC code. The usability depends on the control task.

5 Software
The software is part of work package 4 and includes the implementation of the ADACOM adaption strategies to real ADACOM hardware systems. The development of software includes the tasks signal analysis, feature extraction, algorithm implementation, data transformation and implementation of process monitoring strategies. In general, the software tools are all developed for the ADACOM data format which is a tdms-format of NI with defined channel and channel group properties.

The signal analysis, feature extraction and development of algorithm are realized within the case study work and are connected to thematic emphasis. In the following the ADACOM development emphasis and the corresponding implementations are presented.

5.1.1 Signal analysis tool for tonal sound and vibration analysis
The emphasis "Signal analysis tool for tonal sound and vibration analysis" deals with the development and implementation of methods to analyze all kinds of vibration in manufacturing systems as also when the products are in use. Therefore a specialized software tool was programmed that includes several functionalities for time and frequency domain analysis with suitable visualization and tonal playback to support the human evaluation of the analyzed system. Within the ADACOM project the tool was used for the following tasks:
- Extraction of static frequencies and eigenfrequencies
- Analysis of speed correlated dependence of frequencies
- Extraction of speed and surface correlated frequencies
- Filter components for variable extraction and elimination of frequencies from the measurement file
- Comparison of measured frequencies with synthetic produced sound frequencies for evaluation
- Speed dependent file partition from overall measurement files
- Visualization of Results

5.1.2 Toolkit for continuous plant identification and simulation
The emphasis "Toolkit for continuous plant identification and simulation" is responsible for the basic software module of online identification that is required for the implementation of the ADACOM continuous control strategies. The identification module is part of the "Analysis system". The measurement data for the analysis are stored intermediately to a shift register, so that always the actual data are used for identification. The length of the shift register is responsible for the time that is passed in review for the analysis. The identification of plant parameters can be performed as direct identification of model parameters or a comparison model is used that is iteratively adjusted to the plant behavior by an adaption algorithm that minimizes the differences between model and real behavior. The simulation of a plant model has also the advantage of being able to predict the behavior of the plant for different manipulating variable outputs.

5.1.3 Advanced machine and process monitoring
The emphasis "Advanced machine and process monitoring" deals with latest methods in the area of machine and process monitoring. Sub emphases are:
- Handling of large data files
- High speed real time processing using FPGA
- Position correlated process monitoring
- Network based process monitoring

The extension of the system with a position correlated simulation allows the comparison of the movement to the sensor signals.
For the network based process monitoring some special aspects of data handling like transferability of the data, transfer intervals and data size are very important. For these demands an automated data preprocessing with data reduction and extraction of key features and information are indispensable. Besides this, the safety of the system and the behavior of the system if network connectivity is out of order is a very important task. Within the ADACOM project an internet based post analysis of data that were preprocessed with an high speed FPGA in the machine and afterwards analyzed on desktop PC in office environment were development emphasis.

5.1.4 Signal analysis tool for material and tool wear differentiation
The emphasis "Signal analysis tool for material and tool wear differentiation" focused on knowledge based analysis of cutting processes. Therefore, trials were performed to establish suitable measured values, sensor combinations and process monitoring strategies. The identification of different materials is influenced by changing processing properties as example the tools wear and hence the measured values are changing. Additionally some unusual events might occur that affect the identification.

5.1.5 Geometric surface analysis and NC adaption
The "Geometric surface analysis and NC adaption" is an iterative approach for process design that targets the convergence of the real work piece geometry to the theoretical specification of the product.

For the realization three development emphases were performed:
- Measurement of surface geometry and digitalization
- Geometric analysis of work piece surface regarding geometric errors
- NC-code adaption with the goal of geometric error minimization

5.1.6 GUI (graphical user interface) for manual signal processing
The development emphasis "GUI for manual signal processing" dealed with the implementation of algorithms that can be performed manually to measurement data. Therefore three tasks were primarily handled:
- Extraction of channels and signal segments,
- Operation of signal analysis algorithms,
- Visualization of results.

5.2 Data Transformation Tool
In the initial situation, the project partners have different data acquisition hard- and software systems. Hence, the data acquired from the research on the case studies are stored in the specific data format corresponding to the measurement system. This high variety of data format types hinders the possibilities to exchange data between the consortiums partners.

5.3 Process Monitoring Strategies
Furthermore the right strategy depending on used machine, manufacturing process and surface integrity limits is important. To get an idea of possible strategies from a general point of view, this chapter provides an overview of different strategies and possible applications. Some representative limit descriptions are listed in the following.

Overload and Underload
Static limits are applicable if the cutting conditions are staying almost constant over the process. Overload and Underload type limits set an alarm when the signal remains below respectively over a preset limit for a defined response time at least. This kind of limit is able to detect tool breakages and incorrect workpiece dimensions, for instance.

Work Over and Work Under
This limit type sets an alarm when the work value remains over respectively below a preset limit up to the end of the cycle. This type is also capable to detect tool breakage and incorrect workpiece dimensions.

Contact and Missing
The contact and missing alarm type sends out a message output as soon as the limit is exceeded. This message is reset when the signal remain s below the limit for a preset response time at least. The "Contact" alarm type detects contact between tool and workpiece and can be used to minimise machining times in air (GAP-Reduce). The "Missing" alarm type can detect missing tools or broken-off tools respectively.

Rising Through/Falling Through
These alarm types is set when a time defined limit is passed, but the signal does not pass through the limit in rising or falling mode. Time displaced signals occur for broken or shortened tools, missing tools or work pieces and incorrect tools or workpieces. This type represents just a specific monitoring of start and end of cut with tolerance of full cut. Chip jamming or other major changes in the monitoring signal do not result in false alarms.

Dynamic Limits
Dynamic Limits above and below of a distinct monitoring signal follow the monitor signal continuously to every load level with a limited adaption speed. They may not to be confused with signal pattern or signal tube. In case of extremely fast crossing of one of the two Dynamic Limits, they are frozen (rendered static) and total breakage, breakage, chipping, workpiece cavity, hard cut interruption, etc. are distinguished one from the other via visual comparison with the monitor signal. Sudden load changes are due to total tool breakage or tool chipping.

6 Industrial demonstration
The ADACOM project had a big focus on the demonstration of the developed adaption system. Therefore the project organization was structured to work packages and case studies. The technical work packages represented in the chapters 3 to 5 are for the development of the needed components for implementation of the ADACOM adaption strategies presented in chapter 2. The case studies use the information and results from the work packages to build up single adaption systems that are used as project demonstrators and for evaluation purposes.

In the following the 10 industrial demonstrators are presented:
Milling of chain Guides – Heidelberger Druckmaschinen AG, WZL RWTH Aachen
Optimization of noise behavior of printing presses. Focus on sound emission effect caused by periodical waviness of chain guides. Finishing of the surface by end milling of cast iron chain guides. Small batch production of 10-150 parts and enormous variety of different parts. Improvement is necessary to meet the challenges of the next printing machine generation. Increasing the present margins of printing speeds of 18.000 prints/hour and chain speeds of 4.500 mm/sec.

Grinding of printing cylinders – HDM, Trinity College Dublin, Kistler
This test case is designed to implement an adaptive manufacturing strategy to empower the operator to make decisions on machine and process condition in a complex multi-step grinding process. The approach developed provides a platform for state based condition monitoring and control of grinding process and machine tools. Impacts: Reduced time to repair, improved part quality, reduced waste material, reduced operator effort and stress, predictive maintenance.

Optimization of milling of truck engine block, for small batches production, of Compacted Graphite Iron (CGI450) components to be machined on the Grey Cast Iron (GCI) production line, without any change in machine tools, cutting tools and accessory systems. The objective of the work is to study the automatic adaptation of the machine tool working program parameters to support the change of the engine material from Grey Cast Iron (GG25) to Compact Graphite Iron (CGI450).

Drilling of CGI – FIAT RESEARCH CENTER, MONDRAGON Goi Eskola Politeknikoacnica
Optimization of drilling of truck engine block, for small batches production, of Compacted Graphite Iron (CGI450) components to be machined on the Grey Cast Iron (GCI) production line, without any change in machine tools, cutting tools and accessory systems. The objective of the work is to study the automatic adaptation of the machine tool working program parameters to support the change of the engine material from Grey Cast Iron (GG25) to Compact Graphite Iron (CGI450).

Gun Drilling – Robert Bosch GmbH, WZL RWTH Aachen
The focus is the gun drilling of high-strength steels for mass production with small tool diameters (Ø less than2 mm).
Goal is the development of a system for adaptive controlling of process parameter in order to be able to react fast concerning process disturbances. Main impacts of AC in gun-drilling should be: Rapid optimization of machining process, increased tool life, improved product quality, optimized energy consumption and reduced operator effort.

Drilling and Reaming of GG25 – Robert Bosch GmbH, MGEP Mondragon
The focus is the drilling and reaming of hydraulic control valves made of cast iron with small tool diameters (less than10mm). Goal is the development of a system for adaptive controlling of process parameter in order to be able to react concerning material and geometry variation. Main impacts of AC in drilling should be: Rapid optimisation of process and adaptation to material characteristics, increased tool life, improved product quality, optimized energy consumption and reduced operator effort.

End Milling of Hardened Ring Gears and Pinions – Daimler AG, Trinity College Dublin
This case is designed to implement an adaptive strategy to enable the low batch production of complex gear geometries manufactured from conventional hardened materials, to the same standard as conventional manufacturing techniques. This will allow the development of new ring gears and pinions with optimised tooth geometries. Anticipated impacts of AC: reduced auxiliary times, improved part quality, reduced time to first part, reduced waste material, reduced operator effort, all achieved.

Milling of superalloys – DIAD, WZL RWTH Aachen, ACTARUS
Adaptive control of the residual stresses on aeronautic super-alloys parts, by studying the correlation of milling conditions with the local heating and consequently the residual stresses. The model adaption and the appropriate parameter control is realized with a combination of measured temperature and force. The difference of measured-predicted temperature is the control variable. The set point for control is chosen according to the correlation of temperature and residual stress measurements.

Drilling of superalloys – DIAD, MONDRAGON, ACTARUS
Adaptive strategy for the control of the residual stresses on on aeronautic superalloys parts machined parts, in term of study of the correlation of drilling conditions, with the local heating and consequently the surface and sub surface defects. Main aim of the Adaptive Control model developed by MGEP is to keep temperature constant in drilling process by actuating on drilling parameters, in particular the feed rate.

Milling of composite – LOLA, TEKS, Trinity College Dublin, Actarus
This test case is designed to implement an adaptive manufacturing strategy to characterise and control the cutting of composite materials. The approach developed provides a platform for monitoring of temperature, consideration of forces and examination of tool workpiece interaction leading to chatter. Anticipated impacts: reduced machining time, improved part quality, reduced waste, reduced operator effort and stress, first of a kind temperature data on composite machining.

6.1 Milling of chain guides
During the manufacturing process for chain guides inside the delivery unit of a sheetfed printing press a characteristic topography occurs. The aim is to analyze this topography and to identify a milling strategy which reduces the periodical waviness and improves the working conditions of the printing press.

6.2 Grinding of cylinders
A significant part of the printing machine success in operation is based on precision manufacturing of large components as seen below. In today's production, HDM has to face reworking and rejection costs due to "scattering marks" on the cylinder surface. It is a highly sensitive process, which results in off-times of the machine caused by time required for diagnostics and error detection. In the grinding process of steel cylinders all previous activities have been internal meetings with partners from a broad range of experience such as internal grinding experts, sensor manufacturers and colleagues from the research and development department, hence a project team within HDM was established at the initial stages to support the technical work on the project.

Approach and development strategy
Following various face to face meetings and phone conferences between the TCD and HDM teams, a proposal for in-process-measurement was presented by TCD and discussed with experts from HDM. The proposal was developed taking into account the developed 3D-model, the support of Kistler sensors and HDM production requirements and limitations and this work was undertaken within WP2 and WP3. On site investigations took place on machine tool Schaudt PF61. Data recorded from vibration sensors installed in the machine tool show that the features of the grinding process are evident and form the basis for in-process quality assessment.

6.3 Milling of CGI
The good mechanical and thermal properties of Compact Graphite Iron (CGI) continue to motivate the development of applications of this material for internal combustion engines. The process adaption is intended to support the change of the workpiece material from GG25 (Grey Cast Iron) to CGI, at the same boundary conditions (machine and cutting tool) by the adjustment of the machining parameters dependant on the material and tool condition.

The challenge is to develop a flexible production system that is able to rapidly shift the machining conditions (adapting machine parameters) when the component material is modified. With the same boundary conditions, such as tools and machine, the systems will be able to adapt the working parameters to the workpiece's material response based on the spindle power.

Evaluation of the process monitoring and control hardware was done at WZL. Process monitoring software-machine tool communication was achieved for the actuation of the spindle and feed drive. Independent hardware for process monitoring and actuating systems were defined (chapter 4). Since difficulties in the material identification were found, the goal of power minimization was discussed as another focus for the control.

6.4 Drilling of CGI
Drilling tests were carried out in the laboratory for High Performance Cutting of the Engineering Faculty of Mondragon University. The main requirement of a sensor-actor system was the acquisition of the power consumed by the spindle and the action through parameter changes in machining program.

6.5 Gun Drilling
Only non-controlled gun drilling processes are used currently in mass production of fuel injection systems. Because of decreasing drilling hole diameters and the usage of steel materials with higher strength and higher hardness for new generation of fuel injection systems those traditional gun drilling processes reach their technical limits. In order to increase the process efficiency and to decrease the amount of process disturbances and machine breakdowns an adaptive control system for gun drilling processes with small tool diameters is needed. Additionally, the most suitable process monitoring systems have to identify in order to integrate into such an adaptive control gun drilling process.

6.6 Drilling and Reaming of GG25
The focus of the work was to develop an adaptive-control system for machining processes with small drilling and reaming tools (less thanØ 4 mm) in order to increase the process efficiency and to decrease the amount of machine downtimes. Only non-controlled cutting processes (drilling, step drilling, milling, tapping, reaming) are used currently in production of parts made of cast iron for industrial applications, for example housings of hydraulic systems. Caused by the casting process the iron material shows inhomogeneities and varying part dimensions (large fabrication tolerances). Also this manufacturing sector is characterized by the demand of high flexibility because of large part variation and small batches. Consequently, the range of tool geometries and cutting process condition is very large.

6.7 End Milling
The goal of this work is to end mill hardened spiral bevel gears with free form flank surfaces using a 5 axis machining centre. With traditional gear manufacturing methods, the purchase and tooling cost along with the expertise and time required for setup of every gear geometry type makes the gear generation method worthwhile only for mass production. The low quantities required for experimental validation purposes do not justify the expensive, complex machining solutions offered by many specialist gear manufacturers aimed at high batch production. This work is being realised alongside the development of a spiral bevel ring gear and pinion with optimised new tooth geometry. The spiral bevel gears machined during this research are designed for use within the rear differential of some Mercedes Benz vans.

Details of Technical Developments
The main strategy used in this case study involves the utilisation of three adaption modes, an iterative closed adaption mode using offline tactile measurement based on the Zeiss UPMC Carat 850 CMM to adapt flank form and position errors, an adaption loop for the homogenisation of tool wear via tool contact point adaption (TCPA) to increase process robustness and repeatability, and virtual metrology techniques to identify sources of machining error to allow for workpiece and coordinate centre adaption. The inputs to these adaption modes are based on offline measurements.

Results achieved through ADACOM
To achieve a manufacturing process which is repeatable and reproducible, it is critical to understand the sources of error in the process, so that they can be eliminated. This includes the development of suitable milling strategies on 'soft' material, and transferring these techniques to hardened steel (HRC 62), the switch from point to flank milling, the redesigning of clamping units for both pinions and ring gears, the development of the AC modes, the development of virtual metrology techniques, the development of measurement strategies and the switch to a new, more modern 5 axis machine tool, the Deckel Maho DMU 100P. These developments have had an extremely positive impact on the overall quality of the gears, rated as DIN values, the surface quality of the flanks, the flank topography, the tool life, and the running transmission errors of the gear sets. In all these areas, the gear sets produced today through the ADACOM 'state of the art' process surpass those produced via the series process, fully validating the developed machining methodologies. The time to first part has also been vastly reduced since the beginning of the project, with all first parts created through the new strategies being correct.

6.8 Face Milling of Super Alloys
During the manufacturing critical rotating parts in the aerospace industry, the resulting surface integrity is of major importance as it influences fatigue life and hence safety of the product (and consequently aeronautic transportation). Finishing cuts on turbine and compressor disks are in general a critical operation for surface integrity. Milling and drilling operations are critical and sensitive processes due to difficult cutting tip engagement, interrupted cuts and unfavorable heat dissemination. Methods for ensuring sufficient surface integrity nowadays are not flexible and productive enough to meet the challenges of the future. The aim of this case study is the definition of an adaptive strategy for the control of the residual stresses on machined parts, in term of study of the correlation of machining conditions, with the local heating and consequently the residual stresses.

Materials applied for aero-engines turbine and compressor disks are Ni based super alloys and Ti alloys. These materials have a strong tendency to strain hardening and heating generation during cutting: the consistent local heating can leave abnormal residual stresses on the aero-engines parts, decreasing their fatigue resistance at operative conditions (high temperatures). In this perspective, an adaptive machining strategy controlling residual stresses would have major impact on robustness and performance of the production processes.

Contributions of the adaptive system to the generic system include the adaptation capability of feed, speed and cooling condition. The adaptation modes used in this work include:
- Adaptation of parameters: cutting feed and speed
- Adaptation of machine state: cooling condition

The experimental activity has been carried out by following steps:
- initial trials of a first set of tools and cutting parameters, in order to ensure a sensible choice of tools and parameters.
- residual stress measurements and Process Monitoring, in order to constitute a data base for subsequent development of control strategy for case study.
- Main focus on finish milling cut. Only in a second phase the methodology has been extended to drilling, because of the complexity of this operation and the difficulty to measure the residual stresses inside the holes.

A dedicated milling test bench has been created in WZL.

The sensors configuration for milling was composed by: Temperature Measurement System Actarus implemented on the tool holder and the inserts, Acoustic Emission Sensor Kistler 8152A211 and a 3-component-Dynamometer Kistler 9255B mounted on the workpiece fixturing system.
Calibration tests were conducted on temperature sensors, confirming that the measurements were reproducible and reliable, the influence of heat source location on the cutting edge was negligible. A comparison of measurement by Actarus sensors and by thermal camera confirmed good agreement in dry cutting, whereas in case of cutting with coolant the sensitivity of the system was too low.

Determination of stable cutting parameters has been done starting from the production cycles generated from dedicated design of turbine disk. Different finishing milling radial depth and axial depths have been tested doing measurements of cutting forces and acoustic emissions. The chosen parameters, tool and workpiece given force and acoustic emission measurement yielded reasonable results during the trials.

Acquisition of process monitoring data to validate chosen models. Results for model validation have been achieved (results for different combinations of radial and axial depth of cut). For temperature prediction, available models have been taken into account and implemented in LabView. The available models use the force signal as input, furthermore assumptions for shear angle and contact length must be taken.

Creation of a data base for prediction of residual stresses. It has been decided to correlate the process monitoring signals to the residual stresses using statistical methods. In order to be close to real industrial conditions, the influence of the position on the part (tool entrance, steady cut, tool exit) has been determined randomly. Many cutting trials (by interrupted turning) have been performed in WZL, the PM signals and residual stresses measured, correlated and used for the constitution of the statistical database.

Continuous Parameter and Process Model Adaption. The force data has been used as input into the model online, whereas the modelling of residual stresses was empirically determined based on temperature, as a function of parameters. The temperature can be predicted using the data for force in an expanded Jäger model (model from Komanduri and Hou). A model structure based on temperature input and speed modification has been developed introducing a Model predictive control for temperature prediction in order to increase the reactivity of the adaption.

6.9 Drilling of Super Alloys
As mentioned, in a second phase the methodology developed for the Milling of Super Alloys has been extended to drilling operation.
A dedicated experimental set-up has been prepared in Mondragon for drilling operation.

The sensors configuration for drilling was composed by: Temperature Measurement System Actarus with thermocouple incorporated on the drills, and a 3-component-Dynamometer Kistler 9255B mounted on the workpiece fixturing system.

During the preliminary experimental activity it has been found that the cutting speed Vc is the most relevant input variable for the control of cutting temperature, see next picture:

For this reason in a preliminary phase it has been considered the possibility to modify both cutting speed, keeping constant feed rate, but then it has been verified that industries would accept easier the adaptation of feed rate than cutting speed. Therefore feed rate fn has been finally selected as the parameter to be adapted for keeping controlled the temperature.

An extensive experimental campaign it has been carried out in Mondragon for the correlation of the workpiece surface defects to the cutting conditions (parameters, cooling conditions and temperatures).

Development of adaptive control for production machine

Main aim of the Adaptive Control model developed by MGEP is to keep temperature constant in drilling process by actuating on drilling parameters.

The Adaptive Control System keeps the temperature value constant by modifying the feed rate.
A subroutine is added to the main PLC program and a C program is loaded together with the PLC program for the following tasks:
- The subroutine gets (every SCAN cycle) the value of the temperature from the analog input module and sends this value to the C program.
- The C program executes all the operations (PI controller) and specifies the new value of the feed rate. This value is sent back to the subroutine.
- The subroutine modifies the value of the feed rate.
- This cycle is repeated every SCAN cycle (4 millisecons).
The maximum and minimum feed rate must be programmed depending on the tool specification.

The AC system it has been validated in real conditions, the conclusion is that the developed technology is successful in the detection of critical cutting conditions, moreover the self adaptation of feed rate work properly. On the other hand the reduction of the only drilling feed rate doesn't seem to be enough for reducing the temperature in the cutting zone. In fact it has been found that although the feed rate is reduced, there is still heat coming to the tool (and to the workpiece) and thus the only solution (coolant flow?) to avoid material damage is to stop the process. After the end of the Project the research on this topic will continue, considering the possibility to modify also cutting rate, the opportunity apply different temperature measurement systems, investigating the heat flows in cutting behavior during adaptation and considering the adaptation of drill cooling system.

6.10 Milling of Carbon Fibre
The milling and drilling of CFRP within the ADACOM project focuses on two aspects, namely the edge trimming of light weight CFRP products, including drilling and plunge milling; and the end milling of composite moulds which can radically reduce the time for developing new parts.

Initial investigations focused on the barriers to the respective processes, namely deflection of the part during edge trimming, tool wear, and high temperatures. A dedicate test rig with special purpose vacuum chuck has been developed by LOLA and tested in TCD. Applied investigations were carried out in LOLA in the first two years of the project, and then in TEKS and in TCD. Investigations in the earlier periods focused on the measurement of temperatures and forces during the machining of CFRP samples provided by LOLA with industry parameters. This work was tightly coupled with WP3 in particular the focus on the rotating temperature measurement solution provided by Actarus. The rotating temperature sensing device is unique in the market place and development work was undertaken with particular focus on composite machining within ADACOM. The ADACOM system mode developed for the composite machining case is shown below, particularly for the case of temperature measurement and varied for the case where vibration sensors were used when chatter detection was the focus of investigations, i.e. in the pioneering work carried out by TEKS.

The sample ADACOM adaption mode shown above is focused on temperature with the novel solution by the partner Actarus developed in WP3. The novel sensor required significant characterisation work which required fundamental analysis of heat flows.

This technical work is based on temperature ranges of concern in machining composites e.g. up to 200 degrees C approx. In order to develop a characteristic of the device, detailed investigation took place on quantifying the sensor characteristics in a more scientifically rigorous manner. An experimental investigation was designed in order to have a high intensity heat source from a Laser as the point source of heat on the tools.

It can be seen that the rig allows for the tool to rotate and allows the focus of the laser to be adjusted through positioning in the Z axes, the laser spot position on the tool can be manipulated axially using the laser machine tool axes.

The experimental set-up was commissioned in the last period and investigations and analysis of the results show good agreement with analytical models and FE models of the heat flow through the tool developed in ADACOM. From this basis a software concept was developed to consider model based compensation for surface temperatures based on the ITT performance. This has high potential for application in the field.
The issue of chatter during milling was examined in detail by TEKS using the ADACOM approach and ADACOM control mode of parameter adaption with continuous plant identification.

The research undertaken is one of the first studies on the chatter issue during composite machining. Additional investigations undertaken between TCD, TEKS and LOLA focused on the measurement of energy during composite machining which supported the detailed work by DIAD and TEKS on life cycle analysis.

The case of milling of composites was important within ADACOM due the prevalence of composite applications currently and in the future, and the requirement for ADACOM to develop a generic standardised solution for most machining applications. The high end applications introduced by LOLA were valuable cases for the consortium to consider, in particular, these applications challenged the partners TCD and TEKS with novel set-up and lead to significant contributions to the field of applied research and direct application by TEKS and LOLA. The detailed temperature investigations have been reported at International conferences and the insight into chatter mechanisms in composite is also a unique contribution lead by TEKS. Therefore despite the unfortunate circumstances of the LOLA Company in the latter stages of the ADACOM project, the ADACOM investigations into milling of composites can be considered to be very successful through the additional work of WZL, TCD and TEKS.

Potential Impact:

Socio Economic Impact

The Life Cycle Assessment methodology has been applied to the Case Studies in order to compare the environmental impact of the state of the art industrial metal cutting processes with the impact of the processes improved through the application of ADACOM Adaptive Control. The first step carried out has been the LCA of Case Studies before ADACOM (task 2.2) then the LCA with ADACOM (task 2.2) and finally the comparison of the two conditions (task 5.4). A crucial point was the definition of the ADACOM LCA boundaries. Typically from raw materials in the earth to final disposal of waste materials back into the earth. In ADACOM Project the area of study has been restricted to the production phase: the Adaptive Control will not impact (on first approximation) on the product and its functionalities, but only on its production phase. This means to pass from a LCA approach "from cradle to grave" to a simplified approach "from cradle to gate" (the exit of from the factory), in general defined as an "Eco-profile". Two different LCA software have been applied, Simapro (PE International and Ecoinvent) and GABI, in consideration of their different databases, customized respectively by DIAD and TEKS. All the case studies have been analyzed except the Daimler ones, where it wasn't possible any comparison with the state of the art: in these cases the ADACOM System is adopted for a new application and it is not possible any environmental or energy assessment respect current solutions.

A Life Cycle Inventory has been produced for each case converting the all the inputs/outputs indicated by the LCA responsible in quantities of substances. Then in the Impact Assessment phase, the quantities of substances from the Inventory phase have been associated to environmental burdens using the CML 2001 (baseline) method, conform to ISO 14042.

Main environmental impacts categories, according to ISO 14042 are:
- Depletion of abiotic resources – This impact category indicator is related to extraction of minerals and fossil fuels due to inputs in the system. The Abiotic Depletion Factor (ADF) is determined for each extraction of minerals and fossil fuels (kg antimony equivalents/kg extraction) based on concentration reserves and rate of de-accumulation.
- Climate change – The characterisation model as developed by the Intergovernmental Panel on Climate Change (IPCC) is selected for development of characterisation factors. Factors are expressed as Global Warming Potential for time horizon 100 years (GWP100), in kg carbon dioxide equivalent/kg emission.
- Stratospheric Ozone depletion – The characterisation model is developed by the World Meteorological Organisation (WMO) and defines ozone depletion potential of different gasses (kg CFC-11 equivalent/ kg emission).
- Human toxicity – Characterisation factors, expressed as Human Toxicity Potentials (HTP), are calculated with USES-LCA, describing fate, exposure and effects of toxic substances for an infinite time horizon. For each toxic substance HTP's are expressed as 1,4-dichlorobenzene equivalents/ kg emissions.

In order to calculate the cumulative impact of all the substances on each single impact category, it has been carried out a Characterisation, where the substances that contribute to an impact category are multiplied with a characterization factor that expresses the relative contribution of the substance. As such it can be seen as an equivalence factor. Then it has been done a Normalization of the different impact data to create a uniform unit for all impact categories and to show the relative contribution of all impact categories to the environmental problems in a region. The calculated environmental impact of the work piece material and embedded energy was very high, therefore it had been considered the impact assessment without the work piece contribution to highlight the contribution of the process on the environment. In the cases analyzed the environmental impact of processes waste treatment wasn't negligible: it was important a proper elaboration of the waste composition and disposal/recycling scenario. The typical wastes of metal cutting processes are the mixture of chips and oil emulsion (or net oil), the emulsion (or net oil) from periodic substitution of cutting lubricant on the machine tool(s) tank, the worn cutting tools. In term of waste treatment it has considered a centrifugation of the mixture of chips and oil emulsion (or net oil), to separate metal from cutting fluid. The metal chips can be sold to foundries as a by-product of the process. The emulsion obtained from the centrifugation, as also the emulsion from the periodic substitution, must be treated using chemical additives to extract oil from emulsion. The water obtained can be recycled or reintroduced on the environment, whereas the concentrated oil can be burn at the incinerator producing energy and/or heat. In case of grinding, the metal-oil slurry requires more care in the separation metal powders-cutting fluid, but the waste treatment considered was similar.

The energy absorbed by the processes it has been elaborated in the Impact Assessment adopting an European energy mix that consider the average value of the shares of various energy sources on 27 EU Countries. This assumption is important in order to obtain from the LCA environmental impact data that are significant at European scale and not only at National level. In next table is indicated the share of energy sources for the single European countries and the EU27 mix.

Shares of various energy sources in total gross energy consumption by fuel in 2009

The ADACOM System itself has a power consumption, that has been considered comparable to a Laptop consumption (39W).

The environmental evaluation of Project results has been very positive: all the case studies showed an environmental impact reduction respect to the state of the art, when ADACOM System is applied. It is clear that in some cases, like e.g. Heidelberg cylinder grinding, the main benefits are economic, never less also in that cases an environmental impact reduction has been calculated. At the opposite the CRF case study represents a particular situation, where the ADACOM "assistance" take to very high environmental improvements.

Regarding the energy consumption evaluation, the ADACOM System application resulted beneficial for all the case studies (see next diagram).

At this purpose it is important to outline that the energy save is a key aspect for industries, because it take, at one time, reduction of production costs and environmental impacts, that are beneficial for the competitiveness and market image of the companies. Other crucial advantages of the energy save are related the high variability of energy costs. When the manufacturing industries define new production plants they consider a minimum period of usage of 10 years (sometime more), but they don't not know at that time the evolution of the energy costs during next 10 years. The reduction of power consumption represents the opportunity for the industries to reduce the economical impact of this uncertainty.

From the Social point of view, it is important to consider the relevance of ADACOM energy saves on European companies' competitiveness and consequent preservation of employment. Considering the typical production plants, the energy cost has a major impact for EU industries than for their competitors in emerging countries: all kinds of production energy saves increase the competitiveness and preserve the occupation in EU.

The LCA has been applied also to provide a quantification of the social benefit of ADACOM System application on the human health, the ecosystem quality and the save of resources save, using the LCA Eco-indicator 99 method.

The Eco-indicator 99 defines three types of damage categories:
- Human Health (unit: DALY= Disability adjusted life years; this means different disability caused by diseases are weighted). This category includes the number and duration of diseases, and life years lost due to premature death from environmental causes. The effects included are: climate change, ozone layer depletion, carcinogenic effects, respiratory effects (organics and inorganics) and ionising (nuclear) radiation.

- Ecosystem Quality (unit: PDF*m2yr; PDF= Potentially Disappeared Fraction of plant species). This category includes the effect on species diversity, especially for vascular plants and lower organisms. The effects included are: ecotoxicity, acidification, eutrophication and land use.
- Resources (unit: MJ surplus energy Additional energy requirement to compensate lower future ore grade). This category includes the surplus energy needed in future to extract lower quality mineral and fossil resources.
The application od ADACOM System results in a decrease of processes impacts on Human Health, Ecosystem Quality and Resources (consumption). Also in case of social impacts it can be observed that all the case studies show relevant advantages when ADACOM System is applied. The data obtained are summarized in next diagrams.


- An Integrated Telemetric Thermocouple Sensor for Process Monitoring of CFRP Milling Operations, K. Kerrigana, J. Thil, R. Hewison, G.E. O'Donnell, Proedia CIRP 1 2012, 449-454.
- Machine tool process monitoring and machine condition monitoring – examining data acquisition gateways for process adaption, J. Morgan, G. Eisenblätter, J. Trostel, G.E. O'Donnell International Manufacturing, Conference, 29, 2012 1-10.
- Control adaptativo del proceso de taladrado, Pedro J.Arrazola Congreso de MH y Tecnologías de Fabricacion Proceedings of XIX (2013) / every 2 years Invema Donostia-San Sebastián, 2013 .
- Monitorizacion del proceso de mecanizado Pedro J.Arrazola Seminar: Tecnological seminar on Cutting, Proceedings of Seminar, Invema, Elgoibar, 2011.

Exploitatable Results

1. Adaptive Control hardware platform
Innovation content of result:
-First time implementation of different adaptive production Strategies to machining processes represented by eight industrial case studies
-Online adaption modes for industrial production
Who will be the customer?
-Manufacturing industry in almost every sector as automotive, aeronautics, printing machines, etc.
-Machine tool manufacturers.
-Sensor and process monitoring solution providers

What benefit will it bring to the customers?
-Increase of robustness, flexibility, performance and re-configurability.
-Automatically identification and stabilisation of critical parameter of the machine tool
-Reduced energy input to process, e.g. finding optimal energy input early in process
-Reduced waste parts by getting first part right
-Productivity improvement, optimal part throughput
-Reduced operator workload by using machine intelligence
-Achievement of an online quality control system for mass customization and small batch production

2. System architecture for adaptive control of metal cutting
Innovation content of result:
-system architecture on how to implement adaptive control systems to machining processes
Who will be the customer?
-Manufacturing industry in almost every sector as automotive, aeronautics, printing machines, etc.
-Machine tool manufacturers.
-Sensor and process monitoring solution providers

What benefit will it bring to the customers?
-Increase of robustness, flexibility, performance and re-configurability.
-Automatically identification and stabilisation of critical parameter of the machine tool
-Reduced energy input to process, e.g. finding optimal energy input early in process
-Reduced waste parts by getting first part right
-Productivity improvement, optimal part throughput
-Reduced operator workload by using machine intelligence
-Achievement of an online quality control system for mass customization and small batch production

3. Process adaptive control strategies
Innovation content of result:
-Strategies how to implement adaptive control systems to machining processes
Who will be the customer?
-End users in most manufacturing sectors such as automotive, aeronautics, printing machines, etc.
-2. Machine tool manufacturers.

What benefit will it bring to the customers?
-Increase of robustness, flexibility, performance and re-configurability through automatically identification and stabilisation of critical parameter of the machine tool
-Reduced scrapped parts by getting first part right
-Productivity improvement, optimal part throughput
-Reduced operator workload/variability by using machine intelligence
-Achievement of an online quality control system for mass customization and small batch production

4. Adaptive Control hardware platform: a. Open loop adaptive platform for CFRP machining (lola case study)
Innovation content of result:
-The high quality of the waterjet surface will allow for the part to be produced to the final dimensions with very little finishing required
Who will be the customer?
-Wheelchair and other medical device manufacturers
What benefit will it bring to the customers?
-Faster process times, higher margin

5. Energy saving strategies
Innovation content of result:
-Companies with a high energy bill might want to optimize their operations for energy efficiency rather than maximum throughput. Energy efficiency is highly topical for all manufacturing in Europe currently
Who will be the customer?
-Mass producingsubcontract machinist
What benefit will it bring to the customers?
-During periods of high energy cost at peak times the machines can be optimized for high energy efficiency whereas during periods of low cost ie during night time hours the machines can be optimized for high throughput

6. Human Machine Interface for machine tool process and condition monitoring
Innovation content of result:
Companies In today's production, Heidelberg has to face reworking and rejection costs due to "chattering marks" on the cylinder surface. It is a highly sensitive process, which can result in significant downtime in order to establish the cause of the error. One focus is the process comparison and documentation of a stable process to a process with scattering marks to evaluate specific error charts. The aim is the reduction of off-times of the machine due to quantification of critical influence parameter of machine tool components. A modular Compact RIO System (based on hardware from National Instruments) has been developed and commissioned and has been installed in the grinding machine in order to record and to analyze data from accelerometers and axes position sensors in a standardized manner.

Who will be the customer?
Clients that require detailed process and condition monitoring of high end machine tools on high value components. E.g. Grinding of medical components, machining of large components such as gears for wind turbines.

What benefit will it bring to the customers?
A reduction of off-times of machines due to quantification of critical influence parameter of machine tool components.

7. New design sensors implemented on machine tool
Innovation content of result:
New kind of insert have been created, included sensor, and new tool attachment have been transformed to obtain the result with specific system of wireless data transmission.

Who will be the customer?
Everyone who mills or turns material with high precision and/or who wants to improve tool cycle life or decrease time to set production parameters

What benefit will it bring to the customers?
Less tool consumption, better quality of machining, less waste. Better control of cutting condition

List of Websites:

Documentos relacionados