Skip to main content

SemiOps Cleanroom and Facility Operations

SemiOps is an industry module for semiconductor facility and cleanroom operations. It connects operational digital twin context with cleanroom readings, SMT production records, utility measurements, maintenance records, compliance assessments, and AI-assisted review workflows.

Use this guide when a facility, manufacturing, or quality team needs to monitor cleanroom conditions, investigate equipment or process risk, and prepare reviewed actions for operations teams.

What SemiOps supports today

The current application exposes SemiOps as a set of pages and APIs for operational review:

AreaWhat users can review
DashboardCleanroom count and status, SMT line status, average OEE, current PUE, and active alerts.
CleanroomsCleanroom list, cleanroom detail, particle history, ISO assessment history, and pressure cascade status.
SMT operationsLine overview, OEE, production records, defect records, DPMO calculation inputs, and bottleneck simulation inputs.
Facility utilitiesUtility readings, PUE trend, load forecast inputs, and chiller loading optimization inputs.
MaintenanceFilter condition, equipment health, vibration assessment, and preventive maintenance schedule.
Compliance reviewISO 14644 assessment records and summary views for operational compliance evidence.
Decision centerRecent alerts, decision history, closed-loop recommendation drafts, action prioritization, what-if comparison, capacity planning, execution planning, and escalation planning.

Before you start

Prepare the tenant, asset, and source data needed for the workflow:

RequirementNotes
Tenant accessThe operator must be working in the correct customer or site tenant.
PermissionsRead workflows require semiops.view. Recording assessments, creating decision records, and preparing execution actions require semiops.edit.
Cleanroom master dataCleanrooms should include area, class target, location, and related equipment assignments.
Sensor readingsParticle readings, pressure readings, temperature, humidity, vibration, and utility measurements should be mapped to the correct cleanroom, line, or equipment.
Manufacturing recordsSMT line, OEE, production, and defect records should use consistent timestamps, line IDs, inspection types, and product context.
Maintenance recordsEquipment status, alerts, work orders, and preventive maintenance schedules should be current enough for operational review.
Data foundationUse DFS when source systems need connector setup, mapping, sync history, quality review, or governed dataset preparation.

For data preparation, start with Getting Started with DFS and Data Quality in DFS Lite.

Open SemiOps

Open the FactVerse application and go to the SemiOps workspace. The current UI routes include:

ViewRoute
Dashboard/semiops/dashboard
Cleanroom list/semiops/cleanrooms
Cleanroom detail/semiops/cleanrooms/:id
SMT overview/semiops/smt
SMT defects/semiops/smt/defects
Energy overview/semiops/energy
Utility breakdown/semiops/energy/utilities
Environment overview/semiops/environment
Environment correlation/semiops/environment/correlation
Environment prediction/semiops/environment/prediction
Soft sensors/semiops/environment/soft-sensor
Filter maintenance/semiops/maintenance/filters
Equipment health/semiops/maintenance/health
Maintenance schedule/semiops/maintenance/schedule
Vibration/semiops/vibration
Compliance/semiops/compliance
Compliance reports/semiops/compliance/reports
Decision center/semiops/decision-center
Cleanroom simulation/semiops/simulation/cleanroom
SMT simulation/semiops/simulation/smt

Review a cleanroom

  1. Open SemiOps > Cleanrooms.
  2. Select a cleanroom and review its detail page.
  3. Check particle trends by particle size and sampling period.
  4. Review the latest ISO assessment history.
  5. Check pressure gradients and pressure cascade status.
  6. Confirm related equipment assignments before using maintenance or health views.

The backend cleanroom endpoints expose list, detail, particle readings, ISO assessments, and pressure gradient data under /api/v1/semiops.

Record an ISO assessment

Use the compliance workflow when the team needs to record an ISO 14644 assessment result:

  1. Confirm particle data and cleanroom target class.
  2. Run the ISO assessment flow.
  3. Review assessed class, pass status, per-particle-size limits, means, and upper confidence limit values.
  4. Save the assessment only when the reviewer is authorized and the result has been checked.

The AI Engine ISO assessment uses particle data and target class as input. It returns the assessed class, pass status, and per-size details. Use official quality and compliance processes as the approval record.

Monitor particle and pressure conditions

SemiOps can evaluate particle readings and pressure gradients before they become action items:

CheckInputOutput
Particle monitoringSampling point, particle size, particle count, and ISO class.Normal, warning, or critical status with per-size summary and alert details.
Pressure cascadeZone, target pressure differential, actual pressure differential, and tolerance.Normal, warning, reversed, or sensor fault status per zone.
Environmental trendRecent parameter values, timestamps, horizon, and optional thresholds.Predicted value, trend direction, confidence, and threshold timing.
Environmental correlationTime-aligned values for temperature, humidity, pressure, particle counts, or other parameters.Pairwise correlations, strong correlations, and generated insights.
Soft sensorsTemperature, humidity, particles, air changes, cleanroom age, filter pressure drop, and operating hours.Estimated AMC, dew point, and filter loading with contributing factors.

Trend and correlation outputs are review aids. They depend on time alignment, source freshness, and sensor quality.

Review SMT line performance

Use the SMT views when production, process, or quality teams need to connect equipment context with line performance:

  1. Open SemiOps > SMT to review line status and OEE.
  2. Open SMT defects to review defect records by line and period.
  3. Use defect classification to normalize defect types and inspect typical cause categories.
  4. Use DPMO inputs when defect counts, unit counts, and opportunities per unit are known.
  5. Use bottleneck simulation to compare station cycle times, uptime assumptions, and line throughput.

The SMT simulation uses discrete-event simulation when the runtime dependency is available and falls back to analytical calculation. Treat it as a planning and review tool for line balance, not as a high-fidelity physics model.

Review utilities and energy signals

The energy and utility views support operational review of facility power and cooling signals:

WorkflowWhat SemiOps calculates
PUEFacility power divided by IT or process load, with benchmark category and trend review.
Load forecastShort-horizon forecast from recent load history, typical fab load profile, and recent trend.
Chiller loadingCandidate chiller loading plans based on demand, rated capacity, COP curves, ambient temperature, and electricity cost.

Use these outputs to support energy review, shift planning, and facilities discussion. Confirm setpoints, safety limits, and operating constraints with the site facilities team before field changes.

Review maintenance risk

SemiOps maintenance views connect equipment status, alerts, work orders, schedules, and AI-assisted risk checks:

ViewWhat to review
Filter maintenanceCurrent pressure drop, pressure-drop trend, remaining life estimate, and replacement recommendation.
Equipment healthComposite health score using equipment age, recent alerts, maintenance state, and sensor deviation.
VibrationFrequency-domain velocity measurements against VC curve requirements and violation details.
SchedulePreventive maintenance schedule for cleanroom-linked or semiconductor facility equipment.

If source data is incomplete, some demo or fallback flows may provide typical values for review screens. Production workflows should connect actual equipment, sensor, alert, and work-order sources through governed data pipelines.

Use the decision center

The decision center brings recent alerts and operational records into a structured review workflow:

  1. Open SemiOps > Decision Center.
  2. Review recent alerts from the last operating period.
  3. Open decision history to compare alerts, work orders, and past actions.
  4. Run a closed-loop recommendation to create diagnosis, recommendation, and work-order draft.
  5. Review action priority, what-if comparison, and capacity planning before preparing a plan.
  6. Prepare execution or escalation only after the responsible owner confirms the action.

The current backend returns recommendation drafts, plan summaries, and escalation suggestions. Writing decision records and preparing execution paths require semiops.edit.

API surface

SemiOps application APIs are grouped under /api/v1/semiops:

GroupExample endpoints
Cleanroom/cleanrooms, /cleanrooms/{id}, /cleanrooms/{id}/particles, /cleanrooms/{id}/iso-assessments, /pressure-gradients
SMT/smt/lines, /smt/lines/{id}/oee, /smt/lines/{id}/production, /smt/defects
Utilities/utilities, /energy/pue
Maintenance/maintenance/filters, /maintenance/health, /maintenance/vibration, /maintenance/schedule
Compliance/compliance/iso14644, /compliance/assessments, /compliance/assess
Decision/decision/recent, /decision/history, /decision/closed-loop, /decision/execute-plan, /decision/escalate
Prediction/prediction/action-priority, /prediction/what-if, /prediction/capacity

AI-assisted SemiOps tools are exposed under /ai/semiops:

ToolPurpose
/iso-assessAssess particle data against an ISO cleanroom target class.
/pressure-checkEvaluate pressure cascade readings against target values and tolerance.
/particle-monitorEvaluate particle readings against ISO limits and generate alert summaries.
/env-predictPredict a short-term environmental parameter trend.
/env-correlateCompute correlations across time-aligned environmental parameters.
/soft-sensorEstimate AMC, dew point, and filter loading from measurable cleanroom inputs.
/filter-predictEstimate filter life from pressure-drop history.
/health-scoreCalculate equipment health from age, alerts, maintenance, and sensor deviation.
/vibration-assessAssess vibration measurements against VC curve requirements.
/pue and /pue/trendCalculate PUE and summarize PUE trend.
/load-forecastForecast fab electrical load from recent history and profile assumptions.
/cop-optimizeCompare candidate chiller loading plans.
/defect-classifyClassify SMT defect records and build a Pareto-style priority view.
/defect-dpmoCalculate DPMO and approximate sigma level from production quality inputs.
/bottleneck-simSimulate or estimate SMT line throughput and bottlenecks.

Validation checklist

Before using SemiOps output in an operational meeting, confirm:

  • cleanroom, equipment, line, and utility identifiers are mapped to the correct tenant objects;
  • timestamps are aligned across sensors, production records, alerts, and work orders;
  • engineering units are consistent, especially pressure, particle concentration, vibration velocity, cooling demand, and power;
  • source values are current enough for the decision being reviewed;
  • alert and work-order owners are visible;
  • AI-assisted outputs have been reviewed by the responsible facility, manufacturing, quality, or maintenance owner.