Until the end of 2018, Bartira sequenced only its critical lines, a situation that changed after the full implementation of Opcenter APS, which enabled the sequencing of all the factory's resources, taking into account its various constraints such as product-exclusive resources, resource calendars, efficiencies, tooling, and others. Since every process matures over time in terms of capability, systems, and information, the need for continuous improvement means that any investment in an APS solution takes on the character of a progressive rollout.

One year after the implementation of the first project, and with day-to-day use of the solution bringing greater visibility to the PPCP team and the shop floor, new needs were identified, leading the company to carry out another project together with NEO, this time focusing on more customized and advanced scheduling rules for sequencing through the Opcenter APS API, one of NEO's specialties and differentiators.

For each of Bartira's sectors, distinct rules were developed according to the particularities of each process. Here we present the 3 major improvements introduced in this project.

CROSS-FLOW AND PHYSICAL SPACE

In the factory's initial processes, given the physical space constraint on the rails between machines, it was imperative for Bartira that the sequencing rule developed should prioritize reducing WIP during sequencing, as well as reducing cross-flow.

For the heuristic to control and reduce inventory, the identifier of each rail in the factory and which production orders are on them at the moment of sequencing are read from the MES inventory system. With this information it will be possible in the future to manage the consumption and replenishment of rail inventory, restricting sequences destined for full rails and prioritizing those that are empty and require demand.

Each rail holds information on preferred destination machines, taking into account the distance between machine and rail. In this way it is possible to plan the following days always considering, for each machine and rail, their preferred destinations.

BATCH SPLITTING

Additionally, for all sectors a rule for splitting operations (dividing an operation into smaller batches) throughout sequencing was developed, allowing lead time to be reduced. Through a customized screen, the planner can select the minimum quantity from which a production order can be split and the minimum quantity of each division. With this data, NEO's heuristic evaluates, during sequencing, whether a specific order has the potential to be delayed if it is not split. In cases where it is identified that the order may be delayed and it meets the criteria set by the user's input, the rule splits it into as many operations as possible, respecting the minimum batch values of each division and thereby further reducing delays.

REWORK SCHEDULING

Finally, another important point in the company's day-to-day planning is rework production orders, which enter as a priority within the day to fulfill orders that are stalled awaiting completion. These orders are usually small batches, making sequencing very delicate, since simply sequencing them as a priority could cause numerous setup changes, greatly burdening the day's productivity.

To meet this need, the rule developed by NEO operates in 3 stages:

1. Optimization of similar parts: once sequencing is complete, the heuristic developed seeks to allocate rework orders during productive moments so that they do not generate setup. Since production quantities are low, adding one more production order to the queue does not lead to significant impacts on the future horizons of that resource, if well allocated.

2. In cases where rule 1 cannot act, optimization by gaps in production is sought: still with sequencing complete, the heuristic seeks to fit operations into productive moments when the machine is not working because it is waiting for previous processes to finish their batches. The fit occurs if it is possible to carry out the setup to make the part, as well as its execution within the planned idle period.

3. If there is still a need to produce rework orders that were not handled by the first 2 rules, the heuristic reschedules that day, seeking the allocation that best serves their production in terms of setup optimization and delivery dates.

After the implementation of the new rule, the flow of products across the shop floor became much smoother. The control of preferred destinations between machines and rails drastically reduced the cross-flow of materials between processes, as shown in the example below.

With the ability to split operations throughout sequencing, it was also possible to reduce resource idleness, since the splits generate better utilization and balancing of those resources.

"The implementation of all these rules took the burden off the scheduler, who used to have to perform a large part of these rules manually within the APS, bringing us closer to future autonomous scheduling. With the pandemic period, new needs arose, mainly regarding raw material control, which will be the next project we have planned with NEO. Once we gain greater maturity in the use of the new rules and with better control of these raw materials, I believe we will be ready to make this shift toward autonomous scheduling and to interface increasingly with Industry 4.0."— Fernando Canelas, PCP Manager

With the current improvements, Bartira has reduced the need for the scheduler to interfere in sequencing, making the company a future candidate to begin autonomous sequencing with automatic integrations to the factory through the ERP and its MES.