Facility engineering · Clean utilities · SIP scheduling
Decision and system boundary
Do not size a clean-steam generator from average daily steam consumption. The decision is whether a candidate clean-steam system has enough instantaneous production capacity for a declared SIP schedule, and whether precedence-feasible start staggering can remove an overload without changing qualified phase durations or acceptance limits. The calculated boundary begins at the clean-steam generator outlet, includes shared-header demand and ends at each user's condensate outlet. It quantifies heat-up and exposure/hold as finite-capacity events; purge/vent, cooldown/drain, feed-water demand, primary-steam duty and generator efficiency remain external constraints until project data are supplied.
Equipment and retained-liquid sensible heat, saturated-steam latent heat, delivered dryness, transfer efficiency, heat-up loss allowance, hold heat loss, five-minute concurrency, design factor, available-capacity margin, schedule horizon and returnable condensate.
Steam quality and sampling, biological indicators where applicable, cold-point coverage, air and condensate removal, pressure vessels, relief, piping stress, trap selection, vendor transient response, redundancy philosophy and approved sterile-hold limits.
A positive capacity margin does not prove sterility, compliance or vendor suitability. A negative margin is not permission to shorten purge, heat-up or exposure. It identifies a capacity, schedule or evidence gap for formal review.
Evidence classes
EU GMP Annex 1 and FDA inspection material support the validation and steam-quality boundary; they do not supply the synthetic equipment masses or cycle values below. Noble's original SIP transport study supports treating air and condensate removal as critical phenomena outside a simple heat balance. The NIST source supplies thermophysical-property context, while the manufacturer page supplies only a documented capacity plausibility range under stated reference conditions.
Nomenclature and equations
Qsens = Σ mjcp,j(TSIP − T0). Report Q in kJ. Include retained liquid only when it is deliberately inside the modeled cycle boundary.
ms,heat = Qsensfloss/(xhfgηtr), then ṁs,heat = ms,heat/theat. Do not hide the same distribution loss in both floss and ηtr.
ṁs,hold = 3,600q̇hold/(xhfgηtr). A calibrated measured hold flow is preferable; this equation is a screening fallback.
Dk = Dbase,k + Σu,pzu,p,kṁs,u,p, where z is one only while a user's phase is active in bin k.
Greq = fdesign maxk(Dk); margink = Gavailable − fdesignDk. State separately any vendor derating or redundancy rule.
mcond,return = fretΣms,u,p. Condensate value and quality are separate from the instantaneous capacity and sterility decisions.
Reproducible worked example
Common assumptions are T0 = 25 °C, TSIP = 121 °C, stainless-steel cp = 0.50 kJ kg−1 K−1, hfg = 2,163 kJ kg−1 at the declared nominal 2 barg (approximately 0.30 MPa absolute) screening basis, x = 0.95, ηtr = 0.85, floss = 1.15, Dbase = 40 kg h−1 and fdesign = 1.10. All except the reference property are synthetic assumptions, not site measurements or qualified cycle parameters.
For V-101: Qsens = 3,400 × 0.50 × (121 − 25) = 163,200 kJ. The heat-up screen is 163,200 × 1.15/(0.95 × 2,163 × 0.85) = 107.45 kg, or 322.36 kg h−1 over 20 minutes. The hold fallback is 3,600 × 25/(0.95 × 2,163 × 0.85) = 51.53 kg h−1.
Download the clean-steam SIP phase-load, schedule, sensitivity and evidence template (CSV). The asset exposes every source, input, formula, phase, five-minute demand bin, output, owner and status.
Results and interpretation
The alternative lowers the un-factored concurrent peak by 56.4%, while the three user recipes and their 347.33 kg total phase steam remain unchanged. It extends the schedule from 75 to 95 minutes and adds 13.33 kg of synthetic base-header demand, taking total modeled steam from 397.33 to 410.67 kg. Scheduling must not change validated purge, heat-up, exposure or acceptance criteria. This is a utility trade-off, not an avoided-capital claim: a real facility must price the delay, confirm sterile-hold feasibility and test the generator's transient and turndown behavior.
One-at-a-time outputs make the sensitivity actionable. Low/Base/High dryness gives Greq = 943.79/972.21/1,023.77 kg h−1; transfer efficiency gives 901.58/972.21/1,095.97 kg h−1; heat-up loss factor gives 891.49/972.21/1,093.28 kg h−1; and base-header load gives 950.21/972.21/1,016.21 kg h−1. One/two/three simultaneous highest-duty heat-ups require 398.60/721.90/972.21 kg h−1. Generator derating changes available output to 500/450/400 kg h−1, not demand; condensate return changes returnable mass to 277.87/173.67/0 kg, not the peak. The CSV carries these outputs and capacity margins.
Run one variable at a time against the same pressure and evidence basis. For each candidate capacity, report the minimum time-bin margin and the additional schedule delay. Do not combine numerically Low or High values into an implied best/worst facility unless dependencies are explicitly modeled.
Realization and validation evidence
Provide calibrated steam-flow, pressure and temperature trends; periodic or qualified dryness-test results; feed-water state; header loss; condensate flow; trap performance; generator output, turndown, recovery and derating curves at actual conditions.
Provide approved phase definitions, cold-point data, air-removal and condensate-removal evidence, cycle acceptance, alarm handling, maximum sterile hold and the change-control boundary. Capacity outputs do not replace these records.
Provide CIP completion, reassembly, operator, production-readiness, batch-start, maintenance and sterile-hold constraints. Confirm that every staggered start is precedence-feasible before treating it as a capacity alternative.
Realization evidence is a before/after time-bin ledger using the same meters, pressure basis and cycle identifiers. Attribute a lower observed peak to scheduling only after controlling for product mix, number of users, maintenance state and cycle recipe. Keep condensate recovery, capacity margin, avoided capital and throughput value in separate ledgers; this article claims none of them as a realized saving. This screening model cannot qualify steam quality or prove sterility.
Limitations and practical checklist
The lumped heat balance omits detailed two-phase distribution, local heat transfer, piping pressure drop, non-condensable gases, trap dynamics, cold spots, detailed control response and microbial lethality. Rounded heat capacity, loss, dryness, efficiency and hfg inputs are screening values. Commercial capacity tables apply only at their documented reference conditions. A real design needs code-compliant mechanical engineering, site measurements, validation and vendor review.
Define the generator-to-user boundary, pressure basis, standby loads and condensate destination.
Classify each value as measurement, validated input, manufacturer value, reference property, assumption or model output.
Break every SIP cycle into purge/vent, heat-up, hold and cooldown/drain without changing approved phase limits.
Use pressure-specific properties and avoid counting the same heat or distribution loss twice.
Place all users on one time grid; calculate peak, design requirement and minimum capacity margin.
Test start staggering against CIP completion, production readiness, operator availability and sterile-hold limits.
Confirm steam quality, cold-point coverage, air removal and condensate removal outside this capacity model.
Obtain vendor transient, turndown, reference-condition and redundancy evidence before procurement.
Technical FAQ
No. Average consumption can hide short concurrent heat-up peaks. Put every SIP phase on a common time grid, sum the active phase rates in each bin, and apply an explicit screening design basis to the highest demand.
It should not. A schedule may shift a cycle start only within qualified readiness and hold constraints. Any change to purge, heat-up, exposure or acceptance criteria belongs in the applicable quality and validation process.
No. The screening balance also needs equipment and retained-liquid thermal mass, initial temperature, steam dryness, distribution and transfer losses, hold heat loss and, when available, measured cycle traces.
Not automatically. Condensate recovery can improve the water and heat-recovery boundary, but the generator must still meet instantaneous sterilization demand and applicable steam-quality requirements.
No. It is a utility-capacity screen, not a qualified SIP recipe. Steam quality, sampling, validation, cold-point coverage, biological indicators where applicable and cycle acceptance remain governed quality activities.
A larger or redundant configuration may be preferable when delays violate production-readiness or sterile-hold constraints, users must start together, or verified transient, derating and redundancy requirements exceed the staggered screen. Procurement still requires site and vendor data.
Primary sources
European Commission. EudraLex Volume 4, Annex 1: Manufacture of Sterile Medicinal Products. Final 2022 revision; see clean/pure-steam and sterilization-cycle controls. Official guidance.
U.S. FDA. Sterile Drug Substance Manufacturers (7/94). Inspection guidance covering saturated clean steam, cold spots and condensate accumulation. Official FDA source.
Noble PT. Modeling transport processes in sterilization-in-place. Biotechnology Progress. 1992;8(4):275–284. doi:10.1021/bp00016a003.
Harvey AH. Thermodynamic Properties of Water: Tabulation From the IAPWS Formulation 1995. NISTIR 5078. doi:10.6028/NIST.IR.5078.
Spirax Sarco. CSG-HS clean steam generation system for healthcare sterilisation, TI-P663-01. Capacity figures are used only as a documented plausibility context under stated reference conditions. Manufacturer technical source.