Skip to main content

Feature-Based Design and Modelling for Injection-molding Optimization

Final Report Summary - DES-MOLD (Feature-Based Design and Modelling for Injection-molding Optimization)

Executive Summary:
Des-MOLD system is born with the idea to solve the market needs that injection companies have nowadays. Most of these needs are present in the design stage and Des-MOLD is prepared to face these problems and give the users a solution. Des-MOLD is composed of several different modules that are responsible for reducing the time and money a company spends during a mold making and an injection piece production processes.

Des-MOLD uses past empirical industrial and design experiences and simulation data to provide recommendations on how to solve some issues at the design time. Artificial intelligence techniques such as case-based reasoning and computational argumentation will permit both the inference of quantitative and qualitative information based on a large variety of empirical data and the justification of each decision.
With all this optimizations, Des-MOLD do a reduction of the material consumption, a decrease of rejected parts and an improvement of plastic parts and moulds quality.

Des-MOLD product is a web application that allows the users to access simultaneously to a decision support system for knowing exactly how to sensorize the mold and a decision support tool based on computational argumentation techniques for improving the geometrical design.
This web application has a debate area that acts as community network, where designers can exchange experiences and information to learn from their experience, using these debates to recommend the design of a mold.

This useful tool will benefit the global injection molding value chain. Designers will design more accurate plastic products and moulds which will result in a decrease of the prototyping and process optimization phase and consequently, a decrease of material consumption. Moreover, a better quality part will be produced, as previous experiences will provide designers with new knowledge to improve the design of plastic parts and moulds.

Des-MOLD is not only useful for the design part of a mold, it incorporate algorithms that determinate the optimal setting of sensor in a mold. This is very useful because of the importance of having control of every process inside the mold and ensuring that the final piece has the best quality possible.

To achieve this goal, Des-MOLD application offers two different features, both to get a similar result but working in a different way. The first working mode consists on a CBR which indicates where to implement the sensors in the mould by using previous experience of other injection molders. And the second working mode consists of software which shows the placement of sensors in a mould through simulations. The second working mode is more precise than the first but requires the user to have simulation software. Des-MOLD also offers an offline tool able to create statistical models and classify production batches, based on the data gathered by the cavity sensors.

Project Context and Objectives:
Objective: The main strategic objective of the Des-MOLD project is to reduce the cost of injection-molding production by developing and validating a set of knowledge-based tools specifically oriented to mould makers and plastic injection companies, which will reduce the need for mock-ups of moulds, and several try-and-error trials to calibrate process control variables.

Approach: The Des-MOLD project aims to build a new generation of intelligent knowledge-based systems for added-value injection molding design, empowering the tool making community and thermoplastic injection companies by means of considering the entire manufacturing chain of a new/modified mould design. Des-MOLD uses as a main source, past empirical industrial experiences and simulation data to optimize, at the design time, the geometries of the pieces and moulds according to the desired features, material properties, and to the expected process control variables; modelling this domain and supporting reasoning about features. Artificial intelligence techniques such as case-based reasoning and computational argumentation will permit both the inference of quantitative and qualitative information based on a large variety of empirical data and the justification of each decision.

Des-MOLD is envisaged as a multidisciplinary collaborative environment for problem solving inter-company / intra-company for designing injection parts and moulds.

*Des-MOLD allows from the early stages of the design life to manage and share engineering knowledge and data throughout the entire development process;
*Des-MOLD provides recommendations as to possible solutions while also keeping track of the discussions and decisions taken.

Project Results:

MAIN S & T RESULTS:

SYSTEM´S REQUIREMENTS AND ONTOLOGY DESCRIPTION

This work was aimed at reaching five objectives:

• Analyze and identify explicitly current best practices in injection moulding optimization.
• Define functional and design requirements of the final system.
• Define a common ontology regarding the injection process and the mould that will be used to during the entire project,
• Determine the profile characterization of each potential partner regarding the participation in a possible exploitation platform
• Study confidentiality issues, which non-disclosure agreements, secure transmissions, warrants of copy right, and possible punishments.

Conclusions:

• The ontology has been extended / modified during the project. This work involves the cooperation of all the members in the project.
• The ontology covered a series of objectives:
- Allow data exchange between programs.
- Simplify the unification (or translation) of various representations.
- Facilitate communication between people.
- To build ontology for moulding design involves a set of sub-ontologies such as materials, geometry, and process. Overall terms, concepts and relations should be verified and agreed on the partners involved in the project.
• Ontology description and formalization, is not something that is close, it is something with alive that it can be changed in some terms or can appears news ontologies. So we can easily continue defining any new point we can see during the project or we can modify it if we see that is not an essential point to define.
• Seven ontologies were developed: Molds, part design, processes, materials, defects and additives, post-processing. The next figures depict the Desmold ontologies.
• The security issues have been considered in the modules developed so far in an individual way.

FEATURE EXTRACTION OF GEOMETRIES AND GENERATIVE SHAPES

This work was aimed at reaching the objective:

• To provide the rest of the processes with a geometrical model to be matched with the empirical database.
• To provide a graphical representation to the user as an output to share and validate the results of the decision tool.
• To develop and implement specific algorithms dedicated to detect the shape intention and to define the geometrical simplified representation.

Conclusions:

At this moment the demonstrator is finished and is able to:
• import CAD data
• apply the geometry recognition algorithms
• Generate an equivalent model. This model is a topological entity composed of basic geometric primitives, main characteristics and the relations between them.
• document the problems and provide this information to the rest of processes to be took into account in the decision making
• give a graphical output to the user to validate the equivalent model generated and the proposals
• Integration comparison mechanism with CBR

INJECTION PROCESS MODELLING AND SIMULATION

This work expected to reach two objectives:

• Characterization of the injection production process extracted from the best practices.
• Study current and possible future sensors to monitor the production phase.
• Linkage between geometrical features and production features.
• Virtualization of mould sensorization for optimal production.

Conclusions:

• A software module based on analytical and statistical modelling of melt flow injection simulation files has been developed and integrated in the Des-MOLD web application.
• New Innovative approach to locate cavity sensors, based on analytical models have been validated
• Four different sets of experimental data, corresponding to a four different moulds and pieces, have been analyzed and validated.
• The performance of the systems has been proved, and the conclusions that could be derived from the classification have received very good inputs form the final users.

DECISION SUPPORT TOOL DEVELOPMENT

This work was aimed at reaching the objective:

• To provide recommendation, in the form of dialectical exchanges of arguments, as to most suitable designs to meet specified goals for products (ARG)
• To integrate domain knowledge, past cases, policies into these recommendations

The knowledge representation will contribute to knowledge acquisition and extraction of:
• Ontologies
• Past cases selected from Case Based Reasoner (CBR-S)
• Debates generated through the Argumentation System (ARG-S)

All knowledge within the system is represented following and linked to the ontologies. This allows seamless integration of knowledge throughout the system’s modules.

Conclusions:

• We have developed a working prototype of ARG-S, based on and expanding an existing argumentation framework and system, and integrated it within the Des-MOLD system, with ontologies and with CBR-S.
• We have developed a number of preliminary use cases.
• We have made progress on the integration of ontologies by considering expertise of users on aspects of the ontology.

INTEGRATION AND PILOT TESTS

This work was aimed at reaching two objectives:

• Set up of a remote platform to host the web-based application.
• Design and implement the necessary data base to host and integrate the different modules.
• Integrate the different IT modules developed in the previous WPs.
• Prepare pilot test to validate separately each module of the application.

Conclusions:

• All communications interfaces between all modules of the Des-MOLD system have been defined and validated successfully during the pilot test.
• The hardware platform that supports the Des-MOLD system is configured. We have also created the database and we have made an initial configuration of the web-application.

MAIN FOREGROUND RESULTS:

• Development of effective structural feature-based modelling software for moulding optimization: Design, implementation and deployment of a software tool capable of suggesting for example alternative geometries for moulds and pieces or alternative materials in order to satisfy the desired structural features.

• Development of a software module for virtual sensorization of moulds and structural features: Once the geometry of a mould is set, it is necessary to sensorize the mold accordingly, in order to ensure the proper monitoring during the injection cycle. Then, we will establish a correlation between graphical features, structural features, defects and the final location of sensors.

• Development of a software module for simplifying and matching CAD geometries., obtaining a simplified representation i.e. a set of primitives from a design, providing the grouping of a set of primitives and to search similar simplified designs/parts

• Development of an integrated decision support modelling tool for moulding injection optimization and planning: Powerful decision support tool that links structural features of the final piece with the production process though the validation of the given geometry and the sensorization of molds according to the needs

Potential Impact:
Environmental impact on the industry -

With regard to the environmental impact, we estimate an improved sustainability of production processes and tools since Des-MOLD reduce the need for mock-up tools, and scrap pieces will be decreased thanks to the optimal designs and sensorization at the design level.

There is a reduction of energy consumption due to scrap reduction. Plastic production scraps end up either in landfill, or as a regrind used in another production. More and more stringent regulations will drastically limit the first option. Recycling of rejected parts is often the preferred solution but does have an impact on the process energy consumption since plasticizing requires much energy. Scrap reduction strategy has a direct impact on energy consumption reduction (less scraped parts are produced).

Additionally, there is also a reduction of spare parts change and metallic wastes by tool lifetime increase. The embedded system in the mold will gather and restore key information, which will be used to optimize the sustainability of the tool. The information obtained from the embedded systems will offer the possibility to recycle some of the components (many of them being still operational) of these intelligent and mechatronics molds. Monitoring tool behavior will provide key information on the tool’s wear and fatigue and should enable more efficient decisions for maintaining mold in specs with an obvious positive impact on tool life time. As a direct consequence: a reduction of consumption of steel and a reduction of metallic wastes

List of Websites:
Liceth Rebolledo (Project Coordinator)
liceth.rebolledo@eurecat.org
Eurecat Centre Tecnologic de Catalunya, Parc Tecnològic del Vallès, Av. Universitat Autònoma, 23, 08290 – Cerdanyola del Vallès (Barcelona) – Spain, Tel: +34 935 944 700 http://eurecat.org/es/