Production Scheduling for Ethylene Oxide Sterilization (Medical Devices)

Operations managers and production planners at ethylene oxide (EO) sterilization facilities can model their full flowshop in Schantt — parallel chambers, directional changeovers, finite aeration rooms, and QA hold delays — to produce a forward-looking schedule that respects every constraint. This guide walks through configuring a representative multi-chamber facility and its three product classes step by step.

This guide follows a fictional composite company built from industry research on ethylene oxide sterilization; all names, parameters, and figures are illustrative.

Industry context

Ethylene oxide sterilization is the predominant method for single-use medical devices that cannot withstand steam, radiation, or other terminal-sterilization processes. The process is a four-stage batch flowshop: pre-conditioning (temperature and humidity conditioning), EO exposure (the sterilant dwell phase in a sealed chamber), aeration (forced-air removal of residual gas), and finally labelling & dispatch. Each stage has distinct cycle times that vary by product class, and the overall lead time from pre-conditioning to release can span days rather than hours on account of aeration durations and sterility-assurance hold periods.

Pre-conditioning brings each load to the required temperature and relative humidity before sterilant exposure — a step that can take 18 to 24 hours depending on the product's material density and packaging. The EO exposure stage itself is the shortest per-batch step (four to six hours), but it is where the critical scheduling tradeoffs converge: which chamber runs which product class, in what sequence, and with which changeover between classes. Aeration, the longest stage, removes residual ethylene oxide gas from the sterilised devices to within regulatory limits. Dense polymer products require up to 120 hours of aeration because EO molecules diffuse out of thick-walled materials far more slowly than they do from standard thin-wall disposables.

Facilities typically operate multiple parallel chambers to meet demand, and the interaction between chamber assignment, product-sequence-dependent changeovers, and finite aeration capacity creates a scheduling problem that manual methods struggle to resolve. A mid-scale sterilization plant running 200–250 loads per month across four chambers must decide which product class runs in which chamber and in what order, then track those decisions through aeration rooms that may be occupied for days at a time by a single load. Quarterly chamber requalification windows — 24-hour blocks for temperature mapping and cycle verification — add further calendar constraints that disrupt any schedule that does not account for them in advance.

The product classes themselves carry divergent processing profiles. Standard EO products (thin-wall syringes, tubing sets) have relatively short aeration cycles and are interchangeable across all four chambers. Dense Polymer products (catheters, thick-walled components) require nearly five times the aeration duration and are restricted to Chamber-1 through Chamber-3. Kit Assembly products (procedure packs combining multiple device types) sit between the two in both chamber-cycle duration and aeration requirements. These differences mean that load composition, changeover sequencing, and aeration room planning must be coordinated together — and any planner who treats all three classes interchangeably will produce an infeasible schedule.

Cadence Sterilization Technologies runs approximately 85 people at a roughly 2,800 m² facility, working across three product classes through four production stages. A three-person planning team manages the monthly schedule.

Process overview

flowchart LR
    PC["Pre-conditioning<br/>(2 rooms)"]
    EO["EO Exposure<br/>(4 chambers)"]
    AE["Aeration<br/>(4 rooms)"]
    LD["Labelling and Dispatch<br/>(2 stations)"]
    PC --> EO
    EO --> AE
    AE --> LD

The four-stage production flow for ethylene oxide sterilization of single-use medical disposables: product batches move through pre-conditioning, EO exposure, aeration, and finally labelling & dispatch. All three product classes share the same routing — divergence is expressed in processing parameters, not stage topology.

Scheduling challenges and how Schantt handles them

The demand that drives this schedule is a set of monthly load quantities for each product class, entered as discrete jobs in Schantt. (If your facility is driven by a different demand signal — cyclical customer orders or a kanban pull system — adapt the quantity input accordingly; the scheduling logic operates on whatever job quantities you enter.) Schantt's optimizer minimizes total production time, scheduling every job forward from a common start date. For this scenario the practical horizon is a 2–4 week rolling window that accommodates the longest aeration blocks. The system offers two optimisation modes: Auto mode explores both job sequence and machine assignment across all stages simultaneously, while Semi-Auto mode lets the planner fix the job sequence and lets Schantt optimise machine assignments within that order.

What Schantt handles well

• Parallel-machine stage with automatic machine assignment — four EO chambers and four aeration rooms function as parallel batch machines; Schantt assigns each job to the machine that minimises total production time across all stages.
• Sequence-dependent directional changeovers — asymmetric changeover durations between product classes on the EO chambers (45–75 minutes per transition in either direction); Schantt favours sequences that reduce total changeover time when searching for the optimal schedule.
• Batch stages with finite capacity — pre-conditioning rooms, EO chambers, and aeration rooms each have validated batch capacities and cycle durations; Schantt calculates processing time as the number of batches needed multiplied by the cycle duration per batch.
• Calendar-aware machine availability with scheduled downtime — quarterly chamber requalification windows entered as per-machine downtime; the schedule routes work around blocked periods without manual re-planning.
• Shift-aware scheduling with calendar exceptions — distinct shift calendars for different parts of the line (2-shift for pre-conditioning and labelling, 24/7 for chambers and aeration), with team-wide exceptions for holidays and shutdowns.
• Multi-stage sequential routing with transfer times — four stages in strict sequence with material-handling delays between them; Schantt chains each downstream stage to start after the prior stage completes plus the configured transfer time.

How Schantt handles each challenge

1. Long-aeration horizon creating a downstream bottleneck.
- Aeration durations range from 24 hours (Standard EO) through 72 hours (Kit Assembly) to 120 hours (Dense Polymer). A single Dense Polymer pallet occupies an aeration room for five days — roughly twenty times its six-hour chamber cycle — making aeration room contention the dominant capacity constraint. In a manual schedule it is easy to overlook how four aeration rooms fill up when multiple long-dwell loads release from chambers on the same day, creating unplanned wait times that ripple back to the EO stage. Planners discover the blocking only when the next chamber cycle finishes and finds no empty room.
- Schantt models each aeration room as a dedicated batch-stage machine with its own processing-time parameters per product class. Because the simulation walks every job through all four stages in sequence, it never assigns a job to an EO chamber unless an aeration room will be available when the chamber cycle completes. The schedule shows each aeration assignment on its own Gantt row, grouped under the stage, so the planner can see room utilisation and contention at a glance.

2. Parallel-chamber contention with asymmetric machine eligibility.
- Four chambers are available, but Dense Polymer is validated only on Chamber-1 through Chamber-3. Chamber-4 handles only Standard EO and Kit Assembly. This restriction creates an uneven load distribution across the four chambers that a manual planner must track separately. Without tool support, the natural tendency is to overload Chambers 1–3 with Dense Polymer work while Chamber-4 sits under-utilised, or to forget the restriction and schedule Dense Polymer on Chamber-4, requiring a last-minute reassignment.
- Schantt expresses this restriction through per-machine processing times: Dense Polymer processing-time entries exist only for Chamber-1 through Chamber-3, and no entry is created for the (Dense Polymer, Chamber-4) combination. The optimisation algorithm automatically respects this — it never assigns Dense Polymer to Chamber-4 because no valid processing time exists for that pairing. Chamber-4 is free to absorb Standard EO and Kit Assembly work, balancing the load across the four-machine pool.

3. Directional sequence-dependent changeovers on EO chambers.
- Transitions between different product classes on the same chamber are asymmetric — for example, switching from Standard EO to Dense Polymer takes 60 minutes, while the reverse switch takes 75 minutes. Same-class transitions take 20 minutes. There are six distinct cross-class transition durations in each direction on each of the three fully loaded chambers. A planner sequencing jobs manually has little chance of evaluating all permutation combinations to minimise the total changeover time across four chambers over a multi-week horizon. The default approach — running long campaigns of the same class — can starve other classes of chamber access.
- Schantt models changeovers as direction-specific durations per (from product class, to product class) pair on each machine. When Auto mode explores job sequences, it accounts for every changeover duration along the sequence and searches for the ordering that minimises the total time penalty. On the schedule and Gantt, changeover blocks appear between processing bars on the chamber's row, making the time consumed visible and auditable.

4. QA hold (biological indicator incubation) as a fixed delay before release.
- After aeration, each load enters a sterility-assurance hold period — 96 hours for Standard EO, 120 hours for Dense Polymer and Kit Assembly — during which biological indicators (BIs) are incubated and evaluated. The schedule must reflect this delay before a load is available for labelling and dispatch. This guide models the QA hold as a transfer time from the Aeration stage to the Labelling & Dispatch stage. Schantt applies the delay after aeration completes and before the next stage can begin — the same mechanism used for forklift transfers between production steps. What the QA hold transfer time does not model is the sterility-assurance release decision itself: BI result evaluation, positive-result escalation, and quarantine management remain a quality function performed outside the scheduling tool. The schedule assumes the minimum validated incubation duration, and real durations may be longer if positive results or retests extend the hold.
- By expressing the hold as a per-class transfer time, every job that finishes aeration enters a scheduled waiting period, and the labelling stage receives a realistic start time that accounts for the hold. The planner can see the full pipeline — pre-conditioning through to dispatch-ready — including where jobs sit in QA hold, without manually tracking incubation calendars alongside the production schedule.

What to model in Schantt

The following table lists the first-class entities you create to model this sterilization flowshop in Schantt. The entity counts match the Cadence Sterilization Technologies dataset.

Additional configured data on detail pages includes 12 per-class routings (all three classes through all four stages), 3 transfer times (pre-conditioning to EO: 30 min, EO to aeration: 45 min, aeration to labelling: 96/120 h QA hold), 3 calendar exceptions, 4 machine downtimes for quarterly chamber requalification, and 38 changeover entries across the four chambers and the pre-conditioning and aeration rooms.

Note on batch units: the dataset uses a batch size of 1,000 to represent one pallet-equivalent unit. A full chamber load of 4 pallets is entered as 4,000 units.

Step-by-step setup

1. Create the four stages in order. Add Pre-conditioning as the first batch stage, EO Exposure as the second batch stage, Aeration as the third batch stage, and Labelling & Dispatch as the final flow stage. On each stage's detail page, configure the forward transfer time to its successor:
- Pre-conditioning to EO Exposure: 30 minutes
- EO Exposure to Aeration: 45 minutes
- Aeration to Labelling & Dispatch: 5,760 minutes (96 hours — the Standard EO QA hold minimum; adjust for your product mix)

2. Add machines to each stage. For each stage, create the machines that operate at that stage:
- Pre-conditioning: PreCond-1, PreCond-2 — assigned the Default 2-Shift (Mon–Sat) calendar
- EO Exposure: Chamber-1 through Chamber-4 — assigned the Chambers 24/7 calendar
- Aeration: Aeration-1 through Aeration-4 — assigned the Aeration 24/7 calendar
- Labelling & Dispatch: Label-1, Label-2 — assigned the Default 2-Shift (Mon–Sat) calendar

3. Create product classes and define their per-class routings. Add three product classes: Standard EO, Dense Polymer, and Kit Assembly. For each class, add a routing entry for all four stages (no stage-skipping in this scenario). On each product class's detail page, leave partial transfers disabled — a full batch moves as a single unit between stages.

4. Add one representative product per class. Create the following products and assign each to its class:
- 5 mL Luer-Lock Syringe → Standard EO
- Central Venous Catheter Set → Dense Polymer
- C-Section Procedure Pack → Kit Assembly

These three products carry the processing parameters of their class and serve as the modelled SKUs for scheduling.

5. Configure machine capacity parameters and changeovers on each machine's detail page. With the product classes created, enter the per-class processing parameters on each machine:

• Pre-conditioning rooms (batch): cycle durations — Standard EO: 1,080 min (18 h), Dense Polymer: 1,440 min (24 h), Kit Assembly: 1,200 min (20 h); batch size 1,000 for all classes
• EO chambers (batch): cycle durations — Standard EO: 240 min (4 h), Dense Polymer: 360 min (6 h), Kit Assembly: 300 min (5 h); batch size 1,000 for all. Important: For Dense Polymer, enter processing times only on Chamber-1, Chamber-2, and Chamber-3 — omit Chamber-4 to enforce the machine-eligibility restriction
• Aeration rooms (batch): cycle durations — Standard EO: 1,440 min (24 h), Dense Polymer: 7,200 min (120 h), Kit Assembly: 4,320 min (72 h); batch size 1,000 for all
• Labelling stations (flow): throughput of 2 units/hour for all three classes

On the EO chambers, add the directional changeover entries. Each chamber needs six cross-class transitions (three product-class pairs, both directions):
- Standard EO → Dense Polymer: 60 min; Dense Polymer → Standard EO: 75 min
- Standard EO → Kit Assembly: 45 min; Kit Assembly → Standard EO: 60 min
- Dense Polymer → Kit Assembly: 60 min; Kit Assembly → Dense Polymer: 75 min
- Same-class (Standard → Standard, Dense → Dense, Kit → Kit): 20 min

Chamber-4 needs only the Standard EO and Kit Assembly transitions (no Dense Polymer entries). On the pre-conditioning rooms and aeration rooms, add zero-duration changeovers between all cross-class pairs.

6. Configure calendars, exceptions, and machine downtimes. Set up the three shift calendars:
- Default 2-Shift (Mon–Sat): working days Monday through Saturday, 06:00–22:00, assigned to pre-conditioning rooms and labelling stations
- Chambers 24/7: continuous, assigned to all four chambers
- Aeration 24/7: continuous, assigned to all four aeration rooms

Add three calendar exceptions as team-wide non-working days: New Year's Day (1 January), International Workers' Day (1 May), and a year-end shutdown (25 December).

Add four machine downtimes for the staggered quarterly chamber requalification (24 hours each):
- Chamber-1: 15 February
- Chamber-2: 15 March
- Chamber-3: 15 April
- Chamber-4: 15 May

For step-by-step instructions on configuring each of these in Schantt, see the Schantt documentation.

Common mistakes

1. Using a single blanket changeover time for all EO chamber transitions. Chamber changeovers between different product classes vary significantly — the Standard-to-Dense transition takes 60 minutes, while Dense-to-Standard takes 75 minutes. If you enter one average value for all transitions, the schedule may over- or under-estimate the actual time between jobs.
Fix: Enter each directional pair as a separate changeover entry. If your facility has validated or historical durations, use those instead of the example values — precision matters more than completeness.

2. Omitting Dense Polymer's chamber restriction and accidentally scheduling it on Chamber-4. Because Dense Polymer processing times exist only for Chamber-1 through Chamber-3, any job assigned to Chamber-4 produces a scheduling gap or a re-assignment error. Forgetting the restriction leads to an infeasible schedule.
Fix: Verify that the (Dense Polymer, Chamber-4) processing-time entry is absent. If your facility has a similar restriction — a machine that cannot run certain families — model it the same way: simply omit the processing-time entry for that combination.

3. Underestimating aeration room occupancy when scheduling EO chambers. A Dense Polymer load occupies an aeration room for 120 hours — five full days. Scheduling a second or third Dense Polymer load back-to-back can fill all four aeration rooms simultaneously, creating a blocking chain that halts chamber output.
Fix: After entering processing times, review the schedule's aeration-stage Gantt to confirm that room occupancy aligns with your throughput expectations. If contention is excessive, adjust the job spacing or stagger Dense Polymer starts across the scheduling window.

4. Entering the QA hold as a separate dummy stage instead of a transfer time. Adding an explicit "QA Hold" stage with zero-capacity machines adds unnecessary modelling complexity and may conflict with the scheduling algorithm's capacity calculations.
Fix: Model the sterility-assurance hold as a transfer time from Aeration to Labelling & Dispatch. This gives Schantt the fixed delay without requiring a phantom stage, machines, or routing entries. Adjust the transfer duration per product class if your validated incubation times differ from the example values.

5. Ignoring calendar exceptions and machine downtimes when setting up the schedule. If the calendar assumes all days are working days while the facility observes holidays and performs quarterly requalification, any schedule generated will include non-productive periods that must be manually corrected.
Fix: Configure all known calendar exceptions and downtime windows before generating the first schedule. The few minutes spent entering them upfront save repeated manual adjustments to every affected job's timeline.

What a good schedule looks like

A well-configured Schantt schedule for this EO sterilization scenario resolves the tension between chamber throughput, aeration room contention, and changeover time in a single optimised view.

Before (manual spreadsheet scheduling):
- Chamber utilisation is unbalanced — Chambers 1–3 carry most of the work while Chamber-4 runs below capacity, because no system tracks the Dense Polymer restriction against machine loading
- Changeover sequences are chosen by rule of thumb (longer campaigns of one class), which reduces changeover time but delays other classes and extends their time-to-dispatch — a Dense Polymer campaign may occupy a chamber for days while Standard EO orders accumulate
- Aeration room contention is discovered only when a completed chamber cycle finds no empty room, forcing an ad hoc re-sequence that propagates delays to downstream jobs
- Quarterly requalification windows are tracked on a separate calendar and often overlooked, resulting in schedules that schedule work during blocked periods
- The planning team spends several days each cycle building and adjusting the schedule, with each revision requiring manual cross-checks against capacity and calendar constraints

After (Schantt Auto mode):
- Chamber load is balanced across all four units — Chamber-4 absorbs additional Standard EO and Kit Assembly work while Chambers 1–3 handle the Dense Polymer jobs they are validated for, and the optimisation finds the distribution that minimises total production time
- Sequence-dependent changeovers are ordered across all chambers to reduce total time spent in transitions — for example, grouping same-class jobs to use the shorter same-class changeover (20 min) and minimising the number of lengthy cross-class transitions
- Aeration room assignments are visible per job on the Gantt, and the schedule never assigns a chamber slot unless a room will be available at cycle end — eliminating downstream blocking discovered too late
- Chamber requalification windows appear as blocked periods on each chamber's timeline, and jobs are automatically scheduled around them without manual rework
- The planning team generates and revises the full schedule in minutes, with every constraint — calendars, changeovers, transfer times, machine eligibility — enforced consistently across all jobs