JTEKT tests feasibility of three automation options

To stay one step ahead of the competition at a global level, supplier to the automotive industry JTEKT is investigating whether its production can be further improved using smart optimisation. It had three feasibility studies performed for various tasks within its production. 

Innovation and smart automation have for years been of key importance to JTEKT Torsen Europe from Strépy-Bracquegnies in the Walloon region. The company is a global market leader in the production of limited-slip differentials for lorries with four-wheel drive, and has 180 employees in this field.

The differentials it produces distribute the engine’s power between the wheels to improve the vehicles’ driving dynamics. They operate 100 percent mechanically and are made up of a series of complex steel gears with internal and external gearings. To stay competitive, further optimisation is needed. The company is therefore investigating whether it can introduce more automation on the shop floor, and consulted Sirris for help with this. It worked with Sirris to investigate the deployment of an AGV for transport and the automation of quality assurance and assembly.

AGV as a logistical aid

Initially, the question of whether it was possible to use an AGV (automated guided vehicle) equipped with a cobot to transport tools from the CNC machines to the warehouse and back was investigated. An overview of possible navigation techniques was presented. An AGV with its own map structure turned out be the most flexible. It was found that there was plenty of time to cover the distance with the specific machines that were selected.

Once the AGV has driven to the desired position, the cobot, equipped with a 2D camera, needs to pick up and put down the tools with sufficient accuracy. To facilitate this, the tool racks must be provided with a marker (e.g. a QR code) that can be referred to. A scenario was drawn up in which the various options for practical implementation were investigated, as well as the question of which AGV would be best for this job. Sirris put JTEKT in contact with potential suppliers of suitable AGVs.

Once this study was complete, JTEKT had a good overview of various AGVs on the market and an idea of the order of magnitude of the investment.

Quality control

The second automation issue related to quality control. In the production hall, the products are manually checked against various requirements in terms of dimensions, roughness, etc. The products are transported in boxes on each pallet, tested and prepared for packaging. JTEKT wanted to know whether some or all of the tasks currently performed by quality inspectors could be taken over by collaborative robots.

Since the range of tasks is very wide, a selection first had to be made. An investigation was thus performed to establish whether a cobot could carry out flatness measurements on gears by holding these under a dial gauge. This was found to be entirely possible, even though the cobot was much slower than an operator. The use of the compact 3D Ko-ga-me measuring system from Mitutoyo was thus proposed as an alternative. Tests showed that both cycle time and measurements were in line with the company’s wishes when this measuring system was used.

Sirris also investigated whether a cobot could perform go/no-go tests on calibres. Since the initial tests with a small batch were positive, the test period was extended and a much higher number of products tested. It was found that the cobot was not only able to perform the tests more accurately, it could do this at a constant quality level because the measuring procedure was identical for each test piece.

Based on this feasibility study, JTEKT decided to purchase a Ko-ga-me and a KUKA LBR iiwa cobot to assist with quality tests. This is currently being implemented.

Assembly

A third test case covered assembly that is currently still being performed manually on two production lines. This related to housings comprising various components that had to be fitted together using a tongue-and-groove system. JTEKT wanted to know whether a cobot could assist with this part of its production. Sirris set up a test rig with two cobots – a UR10 (Universal Robots) for the pick & place operations and an LBR iiwa 14 (KUKA) for force-sensitive operations.

The LBR iiwa 14 was found to be sensitive enough to do the job but did not achieve the desired maximum cycle time. To resolve this shortcoming, a setup was tested in which four assembly operations were performed at the same time. This meant that one cobot no longer needed to wait for the other and it was possible to work within the maximum cycle time.

Using this information, JTEKT can proceed to the decision-making phase.