A soap manufacturer planned to close several plants and move all of their work into one new facility. Between them, the old sites made more than 300 products. Before committing, the company's logistics manager needed an answer to one question: would the consolidated plant actually run once real demand, real schedules and real breakdowns hit it at the same time?
The plant
The new facility had to process soap and then package it, with a buffer between the two. Whether that worked depended on the equipment design and sizing, and just as much on how the plant would be run. Five things would change from week to week:
- demand, including seasonal swings
- product mix
- breakdowns and other equipment variability
- crewing patterns
- the sequence of product runs
Any one of those can be sized on paper. The hard part is what happens when a seasonal peak, an awkward run sequence and an unreliable machine land in the same week.
How we modeled it
The logistics manager brought in Simulation Dynamics, and we built a model of the entire plant in SDI Industry Pro, our discrete rate product for high-volume process plants. Three parts of the product did most of the work:
- Discrete rate flow. Material moved through the model as rates, which kept a plant with hundreds of products fast enough to run many schedules.
- An embedded database. The dynamic factors that drive throughput lived in relational tables inside the model, where the model could reach them quickly and the team could change them without rebuilding logic.
- Multi-stage sequential schedules. Production schedules were defined in stages, and the model played them out across every line.
The top level showed the whole plant, with each line's state as the schedule ran: blocked, starved, changing over, emptying, or in surge. Each line opened into a more detailed sub-model nested inside it. SDI Industry Pro is built in blocks, so an end user can work from high-level templates while a model builder works in the detailed blocks and dialogs underneath.
Why trust it
The manager chose simulation over other analysis tools for one reason: it was the only approach that carried real-world variability and showed how an event in one part of the plant affects another. The model was driven by production schedules and the plant's own operating factors, not by averages. Every result could be traced back to the line states behind it.
What we found
Running the model gave the team confidence in the consolidation plan. It showed how well the plant was balanced and gave the best buffer size between processing and packaging.
The costly finding was not about capacity. For one group of products, the planned machine assignments left material in the system too long. For those products, excess time in the system meant a quality problem. One configuration predicted severe long-term quality problems that would have required a $4 million retrofit.
We then tested other product routings and machine assignments in the model and showed how to avoid the problem.
What changed
- The retrofit was never needed. The routing change was worked out in the model, not on the plant floor.
- Simulation became policy. Because of this project, the company now requires a simulation study before committing to any major capital expenditure.
- The scope grew. The logistics manager went on to use simulation to study supply and distribution policies across the division.
Source
SDI project sheet, Soap Manufacturer Saves Over $4 Million with SDI Industry Pro. The client is not named. For the ReliaSim view of this study, see Consolidating a Soap Maker's Plants Into One.