Predictive Equipment Maintenance

osop_version: "1.0"
id: "equipment-failure-prediction"
name: "Predictive Equipment Maintenance"
description: "AI analyzes sensor data to predict equipment failures before they occur."

nodes:
  - id: "sensor_stream"
    type: "event"
    name: "Sensor Data Stream"
    description: "Vibration, temperature, and pressure readings from machinery."

  - id: "ai_analysis"
    type: "agent"
    subtype: "llm"
    name: "Vibration & Thermal Analysis"
    description: "ML model detects anomalous patterns indicating wear or imminent failure."
    security:
      risk_level: "medium"

  - id: "failure_probability"
    type: "system"
    name: "Failure Probability Score"
    description: "Classify risk: low (<20%), medium (20-60%), high (>60%)."

  - id: "maintenance_alert"
    type: "api"
    name: "Maintenance Team Alert"
    description: "Push alert to CMMS and team mobile devices."

  - id: "schedule_repair"
    type: "human"
    subtype: "review"
    name: "Schedule Repair"
    description: "Maintenance lead reviews, schedules downtime window, assigns crew."
    security:
      approval_gate: true

edges:
  - from: "sensor_stream"
    to: "ai_analysis"
    mode: "sequential"
  - from: "ai_analysis"
    to: "failure_probability"
    mode: "sequential"
  - from: "failure_probability"
    to: "maintenance_alert"
    mode: "conditional"
    when: "failure_risk > 0.2"
  - from: "maintenance_alert"
    to: "schedule_repair"
    mode: "sequential"
  - from: "ai_analysis"
    to: "maintenance_alert"
    mode: "timeout"
    timeout_sec: 60
    label: "Escalate if analysis stalls"