SimWell’s customer is a company working in the food industry. Their plant features multiple production lines, one of which prepares and packages individually-wrapped products. The production cell works at a high rate, thanks to packaging robots and a conveyor system. The client wished to reproduce this system as precisely as possible, in an Arena simulation model, in order to test adjustments on operations control of their equipment, and measure accurately the impact of failures and maintenance on production.
A flexible and dynamic packaging system
The prepared products first pass through an oven, and then move on a cooling conveyor, before reaching the first packaging step: individual wrapping in pouches. The units are initially aligned in rows on the conveyor to keep a high density and quick production rate. Then, the products are picked up by one of 3 articulated robots, which lead to 3 parallel boxing lines. This redundancy allows the system to stay productive, even when one of the machines is down.
If the wrapping rate diminishes, an accumulating conveyor in the shape of an “U” is designed to stretch up to over 5 times its minimum length. Therefore, a large number of products can be accumulated during a short breakdown, without interrupting the production. This configuration is crucial, to avoid having to stop or slow down the conveyor oven, where the cooking time has to stay constant.
After being wrapped, the products are then placed into boxes, the boxes into cases, and the cases into pallets, with robots and conveyors. However, if ever one of the machine breaks, the line won’t be stopped. Some employees can palliate by executing some of the tasks manually. The completion rate will be lower, but at least the production of pouches will keep going. The simulation model developed takes into account these actions and triggers the required logic when it’s necessary.
Numerous interesting scenarios
Thanks to the Excel user interface, the analyst may change any parameter in the Arena simulation model. Up to 100 different products can be simulated. The user can define the attributes of the product:
- amount of raw material
- cooking time
- number of products per pouch
- pouches per box
- boxes per case
- cases per pallet
- dimensions of the boxes and cases
The analyst can test the feasibility of any production schedule. He lists the desired products and the case count.
Other parameters may be modified to test different scenarios:
- Reject-rate for at each production step (trashed products, and also packaging misplacements, which can be fixed)
- Setup times between product batches
- Speed and length of the conveyors
- Production rates for the robots and packaging equipment
- Equipment availability (failures and maintenance)
A useful animation and some trustworthy results
Once the modeling completed, the model was delivered to the customer so they could utilize it to optimize production. The customer was impressed by the realism of the 2D animation in Arena and with the process’ accuracy, which proved totally comparable to reality.
The model allows measuring the number of cases produced, per day, per product, taking into account the limited duration of the workday and the reject rates. Also, the model counts the rejects in each location, each product. Finally, if the production line managed to complete the requested production schedule , the time at which the final batch was completed , for each day simulated, is outputted.
To conclude, the customer now possesses a flexible model which allows them to test numerous useful scenarios :
- different production schedules before testing them in real life
- measure the impact of getting more or less failures on the different machines
- try producing, using only 2 of the 3 lines
- Increase the throughput of the production lines, before investing in equipment upgrades
Finally, since they own a few plants and other very similar production lines, this company can adapt the model to other production lines.