Students: Berkay Sevinc, Ben Williamson, Gerben van Dijk (2017)
Introduction
Composite materials have been used in high-end applications for decades, the high performance of these materials has in the past offset the high production cost, however as this technology matures, the efficiency of the production process also needs to advance. The market for these materials is set to increase as the price in comparison to other less advanced materials becomes even more competitive. Current typical applications include aeroplane wing lets, satellite panels and high performance cars parts. However as this development progresses and costs of production fall, more of our everyday items will start to include composite materials.
This is the firm vision of Airborne our client, which have a production and test facility in Den Haag. To work towards this vision, a completely integrated and automated production process is needed to reduce the costs associated with the production of composite materials. In order to realise this dream parts of the process that are labour intensive or slow need to be optimised. The project group is made up of Berkay Sevinc and Gerben van Dijk both of which are Mechanical Engineering students from HHS Delft, and Ben Williamson an international Mechanical Engineering student from the UK.
Problem Brief
This is where our project makes its entry, on the current production line the composite parts are cut into shapes using a 2D cutting machine, these shapes are then placed in a mould by hand and using an ultrasonic welder tacked together. To increase the production speeds and reliability our client wants a more automated process, with the use of vision software to make the current passive system more active. In order to correctly handle the products, a vacuum suction system needs to be developed, a suitable end of arm tool needs to constructed, and control systems to operate the process also need to be integrated into the current system.
Project Goals
Our client Airborne gave the project group set pieces of criteria to complete, these were made up by three main areas, as seen below. This project has the main goal of testing and developing the process to automate the production of composite materials, the final solution will not be commercially viable, but will allow for further research and understanding which at a later stage can be integrated into the current production system.
Project goals
- Detect the ply by size and shape with vision software
- Operate a network of suction cups to correctly pick the plies up
- The robot needs to know where to put the ply and in what orientation within the pattern
- The tolerance between the plies should be 0-3 mm
Our solution
The main component operating within our system is the vision program that is used to recognise the individual plies, this package was built using the COGNEX system. This was developed using the shape areas of the plies to determine which ply it is, how that ply should be picked up and how it should be placed. As the program was built different layers of criteria are checked in each image, after identifying which ply is in the system, the criteria then moves on to the coordinates of the ply, and its orientation. The photo below shows the COGNEX system as it checks the data from the image, the criteria can be seen checking each size of ply against the data base, with a score of 1 for the ply the system believes is the correct one.
In order to correctly handle and move the plies, a pneumatic control system has been developed that uses a vacuum suction cup method to lightly pick up the plies and safely place them into the mould. This system is controlled directly from the script of the robot, has been integrated with the vision program so that a correct pattern of suction cups is used in each operation. This information comes directly from the plies unique size, suction cups are operated from a pneumatic control box that houses all of the valves and switches for its operation. This system is comprised of 3 vacuum creating valves and 2 switches which use venturi valves to create suction which are connected in a network with 7 suction cups.
The output from the system developed during this project is composite plies organised and placed correctly into a predesignated pattern, therefore the most important element of the project is its ability to form the patterns made by our clients. This means a lot of care has been made to allow the right pattern to be made and within the 3mm tolerance criteria. This has required a large amount of testing and optimisation. These procedures have required the project group to spend a large amount of time calibrating the robot, this is done using a chess board sheet, which our vision package uses as a reference to gauge where each point is on our region of operation. This can be seen in the photo beneath, with each crossing point between black and white squares been used as a reference. This process allows the robot to be accurate in the pick up of objects, very important in our mission statement. The pattern placement has been developed and comprises 4 layers of plies, with each layer containing 3 plies.
Conclusion
Throughout this project the main challenge has been learning how to use all of the different programs associated with all of the new software that has been used in this minor, each member of group brought there own individual talents and knowledge, however all of us have been able to further develop our own understanding of 21st century engineering techniques. In relation to the goals we set ourselves at the start of the project, we believe we have been able to satisfy all of our criteria, and been able to produce a system that is accurate and has the ability to be developed later on to increase it’s potential contribution to creating a fully automated process for manufacturing composite materials.
