Scheduling Technical Ceramics Around Kiln Cycles and Firing Batches

A technical ceramics shop pays for the firing cycle, not for the parts inside it. EDGEBIC by User Solutions models each kiln as finite capacity with the cycle as the step time, carries the firing as real elapsed hours across nights and weekends, and lets you group parts by firing profile so fewer cycles are spent on half-empty loads.

The forming, the machining, and the inspection are all schedulable in the ordinary way. The kiln is what makes ceramics different, and it is usually the one resource nobody schedules.

The Cycle Is the Cost

A twenty-four hour sinter cycle costs twenty-four hours of kiln whether the chamber is full or holds three parts. That is the single most important fact about ceramics scheduling, and most plans lose it by treating firing as a per-part time.

Model the kiln as a work center whose step time is the cycle. A half-empty firing then shows in the plan as the same twenty-four hours a full one costs, which puts the loss where a planner can see it. Once it is visible, filling the kiln stops being an argument between the kiln operator and the expediter and becomes a scheduling decision made a week ahead.

Enter each kiln as its own work center with a capacity and calendar so two charges cannot be promised the same chamber. The general principle is covered in finite vs infinite capacity scheduling.

Firing Is Elapsed Time

A firing cycle does not pause at the end of a shift. Ramp, soak, and controlled cool run straight through the night, and any schedule that clips the cycle to worked hours will be wrong by a full day on every firing.

Model the cycle so it runs continuously. A thirty-hour cycle started Thursday at 2 PM finishes Friday at 8 PM, and unload is placed on the first shift after that. Debind, bisque, and sinter are three different cycles with three different elapsed windows, so model them as three steps rather than one lumped firing time.

Cool-down deserves its own treatment. On large parts and on kilns where thermal shock is a real risk, controlled cool is often longer than the soak, and it belongs on the routing as elapsed time rather than as an assumption the unload crew absorbs.

Grouping by Firing Profile

Two parts that share a firing profile and atmosphere can share a charge. Two parts that do not, cannot, and forcing them together scraps one of them.

Build a setup matrix on the kiln so each profile transition carries its real cost. A transition between two parts on the same profile costs only load and unload. A change of atmosphere from reducing to oxidizing, or a jump in peak temperature that requires a conditioning cycle, costs far more. Group your parts into four or five profile families and record family-to-family transition hours: five families gives you twenty-five cells rather than thousands of part-to-part entries, which is the difference between a table you maintain and a project nobody finishes. See how setup families reduce scheduling complexity.

With real numbers in the matrix, the sequence optimizer groups compatible firings on the constrained kiln. The multi-run layer is guaranteed never to return a schedule worse than the baseline it started from, so running it carries no downside: you compare total kiln cycles and late-job count and accept only if it wins. For how the optimizer reports its own confidence, see how a proven optimality gap builds trust.

After the Kiln

Fired technical ceramics still need diamond grinding, lapping, metallization, and inspection, and those steps have their own capacity. On many shops the grinders, not the kilns, are the real constraint, and nobody has measured it because the kiln gets all the attention.

Pool the interchangeable grinders into a work center group and bind the grinding step to the group instead of one machine. At schedule time the engine expands the group, compares each member's projected finish against live load, and places the lot on whichever grinder gets it done first, carrying that member's own effective time. Lots already on a machine stay there. See how a work center group shops a pool of machines.

Then flag whichever resource genuinely constrains you and the schedule anchors around it, placing kiln charges to feed the constraint at the rate it can absorb rather than at the rate the kiln can produce.

Yield Through the Firing

Ceramics lose parts, and they lose them late. A crack that appears in cool-down is a part you already paid to form, dry, and fire.

There is no yield setting on a routing step, so the loss is handled on the quantity side rather than inside the routing. A product carries a Yield percentage in its inventory planning fields, and replenishment planning inflates the suggested build quantity by it: at 90 percent yield, a need for 100 good pieces suggests starting 112. For a customer order you size the order quantity the same way before you release it, because the scheduler loads every step against the quantity the order actually carries. Firing a hundred to ship ninety then becomes a planning decision rather than a shop-floor surprise, and downstream grinding capacity is reserved for the quantity that will really arrive. The mechanics are in how a schedule accounts for scrap and yield loss.

A Worked Job

A 400-piece alumina insulator order through a shop with three kilns and three grinders.

The bisque landed on Kiln 2 because the prior charge shared its profile family, so the transition cost load and unload only. The sinter charge was filled with a second compatible job rather than run at forty percent, which saved a whole cycle later in the week. Grinding landed on Grinder 3, not the machine written in the routing.

When a Firing Runs Long

Operators log start, stop, and quantity at a floor station, so a cycle that ran three hours over pushes only the work that has not started. Completed operations are never moved by a reschedule, so the record of what actually fired stays intact through every replan. Over a few months, logged cycle times become the evidence your firing profiles and your setup matrix need to get more accurate.

Heritage in Long-Cycle Process Manufacturing

User Solutions has built finite capacity scheduling since 1991, more than 35 years, for environments where a shared, slow, expensive resource sets the pace: US Navy, GE, BAE Systems, and Cummins across 33 locations. The lineage behind EDGEBIC, including the RMDB heritage, drove GE Railcar on-time delivery from 30 percent to 90 percent in an operation full of long cycles and shared constrained resources. A ceramics shop losing cycles to half-empty kilns is that problem in a different material.

Where to Start

Enter your kilns as work centers with the firing cycle as the step time. Put debind, bisque, sinter, and controlled cool on the routings as elapsed steps. Group your parts into four or five firing-profile families and record the transition hours between them. Then reschedule and count how many cycles the grouping saved.

For fundamentals, what is production scheduling covers the basics and sequence-dependent setup times covers the changeover economics. A neighboring sector with the same charge-and-wait shape is heat treating furnace scheduling. The industry fit guide maps the rest and EDGEBIC is the product hub. Want to know what your average kiln fill actually is? Contact US for a demo.