The Bagging Lines Were Tied to the Reactors
Untying Them Found Growth Nobody Had Costed

A specialty chemical manufacturer had packaging lines dedicated to individual bulk production lines, each obliged to pack off product as it was made. The question on the table was how many new bag lines to buy and how much overtime to accept. The model answered that — and then surfaced growth options the company had not known were available.

The configuration

Each bagging line was tied to a bulk production line and had to take its output as it came. That is a coupled system in the strictest sense: bulk cannot run faster than its dedicated bagger, the bagger cannot work ahead, and neither can be scheduled independently of the other.

The decision in front of the team was framed as a capital question — which alternative bag line configuration to buy, traded against the overtime cost of running the existing ones harder. That framing is reasonable, and it is also narrower than the problem.

What the model was asked

• What is the tradeoff between alternative bag line configurations and overtime cost?
• What would introducing short-term storage between bulk production and packaging be worth?
• How would production and packaging behave if they could be scheduled independently?
• What happens in the periods when demand exceeds manufacturing capacity?

The team built a model-based application over a ten-year experimental horizon, reporting average weekly hours per bag line split into production, repair, changeover, idle and overtime — so a configuration could be judged on how it actually spent its time rather than on nameplate capacity.

The answer, and the answer nobody asked for

The direct answer was a configuration recommendation with the capital-versus-overtime tradeoff quantified across demand scenarios. The more valuable result was structural: introducing bulk storage between production and packaging, with high-speed packaging lines running on their own schedule, decoupled two operations that had been welded together by the plant layout.

Those options were what the client valued most, and they were not on the original list of questions.

Read that carefully. Not cheaper, not faster — growth options at low capital cost that were not readily apparent beforehand. The model was commissioned to choose between things that were already on the table, and its return came from putting new things on it.

Why this is the same shape as the cereal case

The mechanism here is identical to the one in the cereal decoupling study, in a completely different industry: two stages that must run in lockstep, an intermediate store that lets them keep separate time, and a capacity gain that arrives without buying more of the constrained stage.

That recurrence is the argument for modelling the mechanism rather than the plant. Once a rate-based problem of this kind is properly solved, the solution tends to transfer to operations built the same way — which is also why the reblend model ended up re-implemented at a dozen similar factories.

On the evidence

The client is Rohm & Haas, later acquired by Dow Chemical, and is named here because they are named in the source record along with the quote above. The account comes from our own project documentation rather than a published paper, and the outcome is reported qualitatively — the record describes the project as “extremely successful” but does not carry a throughput or cost figure, so none is claimed here.

The work predates ReliaSim as a product. It was done using the rate-based method ReliaSim now packages.

Related: Decoupling a cereal plant · How big should a buffer be? · All case studies