In the aerospace industry very high quality standards have to be met. For the manufacturing of carbon fibre parts this is currently solved through extended end-of-line inspection in combination with re-work processes to deal with defective parts. Also, in-situ visual...
In the aerospace industry very high quality standards have to be met. For the manufacturing of carbon fibre parts this is currently solved through extended end-of-line inspection in combination with re-work processes to deal with defective parts. Also, in-situ visual inspection is used for quality control. This is currently causing huge productivity losses during lay-up and leads to a real bottleneck in carbon fibre parts manufacturing.
The ZAero project aims at the improvement of production of large carbon fiber parts. Advanced sensor technology for in-line quality control is developed in order to acquire detailed data about the manufactured parts. In combination with a distributed data gathering system and advanced simulation models, more insight into production processes is made available. After completion of the first half of the project, all major elements of the project were successfully demonstrated: in-line sensors for carbon fiber lay-up, sensors for monitoring of curing processes, and simulation models for mechanical and logistic simulation. Sensor data and simulation results are displayed in order to support decision making in the complex production processes.
Technologies developed within the ZAero project provide important information about production processes. This makes large carbon fiber part production more efficient, helps to avoid waste, and gives the human operator the chance to focus on the important decisions. With a pioneering role in the aerospace domain, the developed technologies from the ZAero project are expected to give a boost also to other sectors.
Current and future carbon fiber part production processes were analyzed in the first year of the project. Based on this, a plant layout was created and a model for logistic simulation was implemented. Results from logistic simulations indicated that technologies which are developed within the ZAero project lead to significant improvements of different production performance indicators.
Optical sensors for in-line quality control of lay-up and infusion & curing processes were realized. These sensor systems were tightly integrated into the production facilities at Danobat and FIDAMC. A re-work mechanism was implemented that is able to directly provide feedback about location and shape of defects found in the lay-up processes. Curing sensors are able to determine flow-front and - to some degree - also the curing status in the infusion & curing process.
Simulation tools developed within the ZAero project address mechanical simulation as well as logistic simulation. Both tools were linked to sensor data via a so-called manufacturing database. This concept provides high flexibility in integration of individual components (hardware and software) in a production environment. Fast simulation approaches provide quick feedback on processes at the shop floor. This supports operators in the assessment of criticality of defects. Furthermore, logistic effects are predicted and shown as a dashboard to operators and line managers.
Advanced machine learning methods were implemented in the context of the project in order to boost defect detection and classification. The scientific impact of the project manifests in a set of publications and data sets (all of which were made available via OpenAIRE/Zenodo). The project also received positive feedback from industry in bilateral contacts and specifically at trade fairs such as JEC, where the project was also awarded the JEC World 2019 Innovation Award. A set of contacts were established and bi-lateral follow-up projects were initiated to bring ZAero technology to the shop floor.
\"The ZAero project aims at going beyond the current state of the art concerning intelligent production of large carbon fiber parts. Three major topics are involved: smart inline quality control, closing the feedback loop from individual parts \"\"as produced\"\" to mechanical simulation, and implementation of smart decision support for human operators.
Lay-up sensor technology developed within the ZAero project covers large areas of carbon fiber material at a high resolution. The large amount of data that is acquired in this way puts high demands on data processing. The ZAero project developed systems that can handle large and complex data sets. Cutting-edge machine learning methods are deployed to extract the most relevant information from the data. The use of such methods leads to a shift of responsibilities for the human operator. The focus for the human shifts from manual quality inspection of individual processes towards supervision of sensor systems and data analysis.
Mechanical simulation used to be focused on the \"\"ideal\"\" manufacturing process of a part. In real production processes, however, deviations of the real part from its design occur. The assessment of such deviations is in many cases very difficult to perform. Subsequently, a very restrictive handling of defects is often implemented in conventional production. Because of strict rules, deviations observed for real parts in production often lead to unnecessary scrapping or re-work. By implementation of real-time mechanical simulation based on the real part as manufactured, the ZAero project enables quick assessment of individual deviations during production. The boundaries of design and production phases are therefore becoming more and more blurred. Although such advanced feedback-loops are limited to specific production domains (i.e. carbon fiber parts in the aerospace industry), future production environments are expected to make much more use of the quick interaction of design and manufacturing. This will enable more iterative production processes in different domains. The ZAero project contributes to the implementation of such flexible production systems.
As a huge amount of data is collected in future production environments, it will be necessary to equip the human operator with the right tools to control and run production systems. As top-level decisions will always be in the hands of a human operator, all relevant data must be prepared and displayed to that operator. Within the ZAero project, a decision support tool was developed and demonstrated at the end of the project. The tool collects information from the manufacturing processes (sensor data), mechanical simulations for individual parts, and logistic simulation of the complete production environment. This global view enables the human operator to meet optimal decisions at the level of the produced part as well as on the level of the whole production chain.\"
More info: http://www.zaero-project.eu.