Engineers at a large consumer products company believe they have an intractable code problem with a laner. Unbalanced lanes at a downstream packer are causing problems for a large packaging line. Something is tripping the photo eye, shutting down the upstream cartoner even though the packer is starved.
Plant engineers have been focused on the code for the laner and have sent waves of programmers to fix the problem. Nothing seems to work.
Is the laner the problem, or are logic changes needed further up the line? Engineers look at four possible scenarios to solve the riddle:
- Change the Programmable Logic Controller (PLC) code for the laner
- Add a buffer after the laner
- Add a buffer before the laner
- Change the slug size
Off-Balance
Plant engineers are looking to resolve a 20% loss in throughput. Infeed from upstream unit operations is running at a consistent rate, but averages continue to languish at about 80%. So, where is the problem?
Plant engineers have long suspected logic problems to the laner. Over several months, they have brought in different teams to reprogram it, but the problem persists.
The company sought help from Haskell’s packaging experts, who work with plant engineers to develop several scenarios to simulate and emulate variables and their effects on the line.
Bug Hunt
The Haskell team creates an emulation of the line used to test and debug PLC programming. The working packaging line is modeled in a real-time environment, and the model is used to simulate a series of variables to the laner. Only negligible improvements are seen.
The team starts to think logic changes may be needed up the line.
The team tries adding buffer after the laner, but there is nothing to fill up. Engineers are already starving the case packer, so adding a buffer will provide no benefit.
Team members also consider changing slug lengths to vary lengths between each lane. Engineers accomplish this relatively easily by using emulation to create a 3D model of the line, where they can simulate numerous variables:
- They can define package dimensions of the product.
- They can control production variables such as motor, speed, sensor placements.
- They can control device logic to resolve problems in the line design or programming.
Such an attempt would typically yield dramatic data changes, but while the team realizes some benefit, it isn’t enough to balance line operations.
The Eyes Have It
The photo eye is being tripped, which shuts down the upstream cartoner. What if the model could emulate moving the photo eye upstream?
Moving the photo eye would mimic adding buffer upstream. Moving a photo eye is not very involved or expensive. However, physical limitations of the conveyor layout would require more significant mechanical and electrical changes to allow the accumulation of product upstream.
Because emulation can demonstrate the effect of product design variables on the working line, system changes can be created and tested without affecting the actual system or incurring the cost of a physical change.
The team decides to emulate adding buffer before the laner by moving a photo eye upstream in the model. This change provides significant improvements to the line: a 16% jump in average throughput of the laner and no shutdowns of the cartoner.
Conclusion
Confirming that additional buffer would smooth out operations of the line, the team now has a solution to implement. Again, physical limitations to the existing layout would require significant funding and time to allow the additional buffer. But emulation provides sound justification for the project.
