Antti Inkinen, Carlo de Beer, Johan de Geus, Marc Dudley
The original situation
DERO Groep designs and builds robotics and automated solutions for customers operating in FMCGs. As part of the build they need to produce custom made parts from their own workshop in batches of up to 200.
The problem
The problem for the workshop technician was that the parts had to be manually loaded into a CNC machine during the batch in order to finish off the parts.
Our solution for DERO
The solution lay in using collaborative robotics that could be easily integrated into the workshop environment. This was achieved by having the set-up table roll up to and be easily connected to the existing machinery.
The proposed sequence
The sequence prototyped was as follows:
- Technician pushes set-up to the machine and connects power, air and serial comms.
- Technician places the unfinished parts they want machining on the search table.
- Technician set-up and program the CNC as they normally would.
- Press start on HMI and then select ‘one-cycle’.
- The robot calibrates its position and searched for the part.
- The robot grips the part and loads it into the machine.
- The machine is started by the PLC.
- The PLC instructs the robot to remove the part.
- The robot opens the door and removes the part, then isolates it.
- Technician inspects the first finished part and can continue with the program for all the others on the table.
- When there are no more part to be found by the robot the program ends.
- The robot can be removed or used for another batch.
The major decisions
The main issue regarding this project was interfacing the relatively old existing machinery with the newer Techman Cobot. We decided early on in the project to test the machinery signals against the Siemens PLC so that if it was not feasible we could switch to an alternative solution. A minor issue with this solution is the risks associated with ‘hacking’ expensive workshop assets. We were lucky to have DERO’s own, supportive and understanding, research and development department backing recommendations that we made.
A minor issue occurred because of the orientation of the robot axis’ when trying to achieve a linear movement into the chuck position. This problem was designed out by attaching the gripper 90 degrees to the camera line sight. The knock on effect on this change meant that the robot spent more time orientating itself correctly between sequence phases, for example moving from search position to a pick position.
During the installation of the prototype we encountered issues with accuracy, namely placing the part within the machine chuck, which had to be very accurate. We had not accounted for the micro movements on the robot base during the door opening and closing sequences. To counter this issue we decided to have the robot calibrate its position against the machine whenever picking or placing was to happen. This had the unfortunate effect of increasing the cycle time. However we discussed and decided that optimisation could be achieved within another design iteration.
Improvements moving forward
The customer was very happy with the proof of concept presented at their facility and are extremely keen to advance this project to the next phase of research and development. General design recommendations for a sounder solution include:
Replacing the gripper with a higher engineered electrical gripper as opposed to pneumatic. This would allow for feedback on mis-picked parts and one less thing (air service) to have to provide for set-up.
A rigidly designed solution for rolling up the system to the machine that can be easily remove but also prevent vibrations caused by the robot movements.
The design of a quick release connector for all services and signals between the cobot system and machine.