Skip to main content

"Design, implementation and validation of an automatic learning cure cycle optimisation process for the eco-efficient autoclave processing of composite materials"

Final Report Summary - OPTO-CLAVE (Design, implementation and validation of an automatic learning cure cycle optimisation process for the eco-efficient autoclave processing of composite materials)

Executive Summary:
The OPTOCLAVE project developed knowledge based optimisation software called DETA-LEARN. The software performs analysis and constrained optimisation of the autoclave manufacturing process. The output is an alternative cure profile that aims to either minimise the processing time and/or the energy usage during processing.
The main functions of the software are the following:
• Analysis of the manufacturing process taking into account the autoclave dimensions, composite part geometry, materials specifications, material state models and real temperature data.
• Constrained optimization algorithm that can be tailored to the process. The optimization can be directed to shortening the cure cycle, achieving a material specification (for example final glass transition temperature) or minimizing the energy requirements for the process.
• Estimation of energy demands for the process, in kWh

The DETA-LEARN software has optimised the autoclave manufacturing of a composite stiffened panel. The process details were supplied by the Topic Manager. The optimised profile resulted in a reduction of 12% in cure time and 16% in energy consumption.
Project Context and Objectives:
Autoclave processing is recognised in aerospace as the process that produces high performance composite structures of large size and complex shape. Composite parts manufactured in autoclave are widely used in order to benchmark the quality of composite parts produced using other manufacturing processes such as Resin Transfer Moulding (RTM) and Resin Infusion (RI).
Autoclave is generally recognised as an expensive manufacturing process both in terms of capital investment and in energy usage.
The need for the OPTOCLAVE project initiated from the challenge to control the temperature in large autoclaves. Temperature control in autoclaves is generally poor with temperature overshoots of >10°C being typically observed. The reason for this lack of accurate control is the convective heat transfer mechanism and the need to heat a large volume of air in order to cure the part. The slow transfer of heat through convection is coupled with the faster conduction mechanism from the metallic tool to the composite and the very fast autocatalytic polymerisation reaction.
The ultimate aim of the project is to provide software that can simulate the cure process inside the autoclave and provide information about the material state of the composite.
In order to achieve the above aim, a number of objectives were defined:
• Analyse the materials defined by the Topic Manager and develop material state models.
• Develop an optimisation algorithm for the autoclave process that takes into account the equipment constraints (for example, the maximum heating rate that can be achieved), the material state properties, the part geometry and shape, and customer specifications (for example a pre-defined time the part needs to stay at a specific temperature).
• Develop an auto-learning algorithm that can perform system identification from temperature signals and dielectric sensors embedded in the part.
• Include all the models and algorithms in one single software that can be readily used by the Topic Manager
• Install final version of the software to the Topic Manager site and train personnel in its use.
In addition to the above objectives, a further objective has been agreed between the Topic Manager and the consortium:
• Estimate the energy savings achieved by the optimized cure process that is calculated by the optimization algorithm.
Project Results:
The work had two phases, depicted in the two RTD Work Packages of the programme.
In the first phase, the methodology of the optimisation algorithm was developed. Also, the materials, demonstrator part and process were defined by the Topic Manager. An experimental analysis of the materials was performed in order to develop material state models that were incorporated in the final software.
In the second phase of the project, the software itself was developed using the LabView 2010 Software Development Kit (SDK). The first version of the software was used for the verification experiments that took place at the Topic Manager site. The results from the first version of the software were assessed alongside an overall review regarding the Graphical User Interface (GUI). All the comments and suggestions were incorporated in the second version of the software.
The second - and final - version of the DETA-LEARN software has been sent to the Topic Manager. The software can performed the following distinct actions:
1) System identification: Analysis of the autoclave equipment. Temperature measurements from previous runs are input. The software calculates the heat transfer function of the autoclave, which is needed for the optimisation algorithm.
2) Process Optimisation: The optimisation algorithm takes into account the process and material constraints in order to calculate an alternative cure cycle that results in reductions either in the processing time or/and the energy requirements.
3) Energy usage estimation: The energy requirements of the process is estimated in KWh.

A training session on the use of the final version of the software has been performed.
Potential Impact:
The exploitable project results are:
• The material modeling methodology: The materials modelling methodology is developed by TWI. It involves a combination of numerical methodologies for the modelling of the cure kinetics of resin system and a new methodology for the determination of the glass transition temperature advancement. The new methodology applies to systems that exhibit a phenomenological deviation from the standard Di Benedetto model for the advancement of glass transition in reactive systems. The new methodologies will be offered to TWI member companies for the characterisation of reactive systems.
• The heat transfer model development methodology and software: The heat transfer model development methodology and software have a wide application to all composites processing routes and the potential will be exploited by ADVISE and TWI (in partnership with ADVISE). Formulation of case studies will be developed with industrial partners for autoclave (prepreg and RTI routes), RTM, oven (RFI and infusion routes) in order to assess the suitability of the algorithm to predict accurately the temperature profile during the process. The software code can be adjusted accordingly and a stand-alone version can be formed. The target industrial groups include aeronautics, automotive and wind energy sectors.
• The cure profile optimisation methodology and software: The cure profile optimisation methodology and software have also a wide application to composites processing routes and the potential will be exploited by ADVISE and TWI. Involvement of material models of other popular resin systems will be made while case studies will be developed with industrial partners for autoclave (prepreg and RTI routes), RTM and oven (RFI and infusion routes). The software will be industrialised through user interface improvement and usability development. The target industrial groups include aeronautics, automotive and wind energy sectors and a license scheme will be devised for the industrial use of the tool. The link to the existing DETA SCOPE cure monitoring system of ADVISE will be made so that the cure profile optimisation software becomes part of the cure monitoring and control suite.
List of Websites:
The main contact for the project are:
Dr Mihalis Kazilas (TWI): Project Coordinator and responsible for the technical work of TWI
Dr George Maistros (ADV): Company owner and in charge of the the technical work of ADV
Dr Aggelos Poulimenos (ADV): Technical contact. Developer of the optimisation algorithm and software development for ADV
Mr Fernando Bianchetti (Alenia Aermacchi): Topic Manager of the project

The contact details of the above are given in a separate document.