IVS has combined Artificial Intelligence machine vision with cutting-edge robot automation to provide a state-of-the-art, flexible automated vision inspection cell. High-definition camera technology combined with LED lighting and laser projectors allows large products to be inspected at speed with high precision, reducing the need for inspection operators in automated production lines. Such solutions are critical to driving productivity within manufacturing and driving towards fully flexible manufacturing & Industry 4.0 production control.
The challenges in inspection
Kawasaki Precision Machinery (KPM) approached IVS to develop a fully automated vision inspection system for multiple parts on their dual pump assembly line. In total, six different parts, with a total of 63 variants, required precision inspection. Historically a human operator would be used to inspect of such parts, but even if 200% manual inspection was utilised, quality could not have been assured. The parts included stoppers, needle bearings, valve plates and three different shafts. The inspection criteria included measuring parts for tolerance checks, verification of the correct types and presence verification of critical build components.
A large number of parts and variants presented a common problem in the world of automated vision inspection, as most systems were designed to inspect only a single part with minimal variants. Creating flexibility in a system’s capability would generally require adding equipment and complexity to the system in the form of supplementary sensors, actuators, controllers and tooling – not to mention the additional programming and system training. All of these complexities further added exponential costs to projects, often making the single, ‘all inclusive’ solution more expensive than several ‘individual part’ solutions. Due to manufacturing floor space constraints, KPM wanted a single, highly flexible vision inspection machine to inspect all six parts and their corresponding variants automatically. KPM also planned to add more parts and variants to the inspection and assembly line in the future. This is also a common issue in the industry as automotive designs are frequently changed and improved. If the changes are minimal, the solution is generally straightforward. Slight changes in check routine programming make it possible to inspect the ‘new’ parts. However, inspecting a completely different part (i.e., gear vs shaft) can prove to be extremely difficult and costly, if not impossible. IVS was dedicated to designing and building a single system for KPM that could inspect all six parts and 63 variants as well as any future parts and variants that KPM would incorporate.
The perfect collaborative engineering
IVS did engineer vision systems in the past that involved rotary actuators and linear slides to position vision systems for inspection at different points on a part. However, it still required several cameras and actuators to inspect a single part, often allowing for no additional part variant changes in the future. Therefore, providing the ‘eyes’ for a robot to inspect the part was the latest innovation from the company. The new machine vision solutions utilised Artificial Intelligence machine vision algorithms to allow the robot vision to ‘learn’ the difference between good and bad parts. As the vision sees more samples, the variations in products are learnt by the system, and the system improves when enough samples have been seen, the system runs autonomously, making decisions on the part acceptance or failure for quality.
It was clear to the IVS team that using a collaborative robot (cobot) was the perfect solution. By integrating a six-axis robot into the vision inspection system, IVS eliminated the need for multiple cameras and the actuators that moved these cameras around. The robot could position the camera head at almost any angle in almost any position, allowing complete three-dimensional checks of the parts. The camera system was further combined with both laser projection and standard LED lighting technology, which allowed multiple analysis of both standard images and laser edge detection.
The output adjudged
Using the robot allowed a single vision machine to inspect endless parts and variants. This made the inspection system future-proof as KPM could add new parts and variants to the system capabilities at their leisure. IVS designed a complete machine that fitted in the production line upstream of a part pressing process. An operator loads the parts to be inspected on a pallet and the pallet was driven to the input side of the machine’s conveyor. When prompted, the machine moved the pallet into the inspection cell surrounded by tinted in-fill panels for operator safety. The robot positioned the camera in several places for the vision system to determine what part was present and what inspection sequence to run. Once the specific part and variation were determined, the inspection machine performed all the checks for that particular component based on its prescribed check routine.
Once complete, the factory information system was notified of the pass/fail outcome of the part and the reason for any failures. The pallet was then moved out of the inspection cell automatically on the outfeed of the conveyor. IVS had supplied the operator and engineer training to KPM, which allowed KPM employees to program and test new parts and variants in the inspection system. IVS had successfully supplied KPM with a vision inspection machine that met all their needs and more. The vision inspection machine fit within the new production line and fed directly to the downstream processes, eliminating the need to reconfigure the manufacturing line set-up. By reducing the operator intervention for the inspection process, IVS helped KPM guarantee 100% defect-free product to their customer.
Collaborating KPM with the future
The flexible vision cell is future-proof; KPM can program the machine and robot to inspect countless parts and part variants. IVS developed the check routines to be independent of the factory communication system or operator input, allowing the machine to determine what part is presented before performing the checks. This removes the potential for operator or communication system error of misinforming the machine of what part has been loaded for inspection. Operator safety has taken a key role in the design and build of the machine. Emergency stop buttons are located at multiple points on the machine, and the inspection cell is completely enclosed by tinted in-fills to protect both operator limbs and eyes. In addition, the robot is human collaborative, meaning it is safe for humans to interact with and will stop its movements if it feels an external force acting upon it. Not only does the vision machine automate the inspection process, removing the human error element, but it also decreases inspection the time decreasing overall production time, whilst at the same time guaranteeing quality. An operator would never be able to cope with the inspection of a large number of variants at a consistent level.
By designing a single machine for all the parts and variants, IVS eliminated the need for four separate machines to cover the six different parts. IVS developed a first-of-its-kind automated inspection cell that exceeds expectations, performs flawlessly and fits well within the project budget and timeframe. This is a real innovation that builds the foundations for removing operators with robot and vision technology, signalling a move to the ‘lights out’ factory of tomorrow, allowing industries to capitalise on the very latest developments in vision and robotic control for quality control.
Credits: Industrial Vision Systems (IVS)