P6: Condition-Based Reliability

Overview

The Condition-Based Reliability package integrates equipment health monitoring with reliability analysis, enabling predictive maintenance and dynamic risk assessment based on real-time condition data.

Key Concepts

Traditional reliability analysis uses fixed failure rates (MTBF). Condition-based reliability adjusts these rates based on actual equipment health indicators:

Key Classes

ConditionBasedReliability

ConditionBasedReliability cbr = new ConditionBasedReliability(
    "C-200",              // Equipment ID
    "Export Compressor"   // Equipment name
);

// Set baseline reliability
cbr.setBaselineMTBF(12000);  // hours (from OREDA or similar)
cbr.setInstallationDate("2020-01-15");
cbr.setOperatingHours(15000);

Adding Condition Indicators

Define monitoring parameters with their healthy and failure thresholds:

// addIndicator(name, healthyValue, failureThreshold, degradationModel)

// Vibration (increasing is bad)
cbr.addIndicator("vibration", 5.0, 15.0, 
    ConditionBasedReliability.DegradationModel.LINEAR);

// Bearing temperature (increasing is bad)  
cbr.addIndicator("bearing_temp", 60.0, 90.0, 
    ConditionBasedReliability.DegradationModel.EXPONENTIAL);

// Oil particulate count (increasing is bad)
cbr.addIndicator("oil_particulates", 0.0, 100.0, 
    ConditionBasedReliability.DegradationModel.LINEAR);

// Efficiency (decreasing is bad)
cbr.addIndicator("efficiency", 85.0, 70.0, 
    ConditionBasedReliability.DegradationModel.LINEAR);

Degradation Models

Model Description Use Case
LINEAR Constant degradation rate Most mechanical wear
EXPONENTIAL Accelerating degradation Bearing failures, fatigue
WEIBULL Bathtub curve General equipment
LOGARITHMIC Rapid initial, then slowing Erosion, corrosion

Setting Weights

Assign importance to each indicator:

cbr.setIndicatorWeight("vibration", 0.35);      // Most important
cbr.setIndicatorWeight("bearing_temp", 0.25);
cbr.setIndicatorWeight("oil_particulates", 0.20);
cbr.setIndicatorWeight("efficiency", 0.20);

Updating Conditions

Real-time Updates

Map<String, Double> currentConditions = new HashMap<>();
currentConditions.put("vibration", 8.5);        // Elevated
currentConditions.put("bearing_temp", 72.0);    // Slightly elevated
currentConditions.put("oil_particulates", 35.0); // Moderate
currentConditions.put("efficiency", 80.0);      // Slightly degraded

cbr.updateConditions(currentConditions);

Historical Data

// Add historical readings for trend analysis
cbr.addHistoricalReading("vibration", 5.0, "2024-01-01");
cbr.addHistoricalReading("vibration", 5.5, "2024-02-01");
cbr.addHistoricalReading("vibration", 6.2, "2024-03-01");
cbr.addHistoricalReading("vibration", 7.0, "2024-04-01");
cbr.addHistoricalReading("vibration", 8.5, "2024-05-01");

Health Assessment

Health Index

double healthIndex = cbr.calculateHealthIndex();
System.out.println("Health Index: " + (healthIndex * 100) + "%");

// Health categories
if (healthIndex > 0.8) {
    System.out.println("Status: Good");
} else if (healthIndex > 0.5) {
    System.out.println("Status: Monitor closely");
} else if (healthIndex > 0.3) {
    System.out.println("Status: Plan maintenance");
} else {
    System.out.println("Status: Critical - immediate action");
}

Individual Indicator Status

for (ConditionBasedReliability.ConditionIndicator indicator : cbr.getIndicators()) {
    double normalized = indicator.getNormalizedValue();  // 0 = healthy, 1 = failed
    
    String status;
    if (normalized < 0.3) status = "🟢 Good";
    else if (normalized < 0.7) status = "🟡 Warning";
    else status = "🔴 Critical";
    
    System.out.println(indicator.getName() + ": " + 
        indicator.getCurrentValue() + " (" + status + ")");
}

Adjusted MTBF

double adjustedMTBF = cbr.calculateAdjustedMTBF();
double baseline = cbr.getBaselineMTBF();

System.out.println("Baseline MTBF: " + baseline + " hours");
System.out.println("Adjusted MTBF: " + adjustedMTBF + " hours");
System.out.println("Reliability reduction: " + 
    ((1 - adjustedMTBF/baseline) * 100) + "%");

Remaining Useful Life

double rul = cbr.estimateRUL();
System.out.println("Estimated RUL: " + rul + " hours");
System.out.println("Days remaining: " + (rul / 24));

// With confidence interval
double[] rulCI = cbr.estimateRULWithConfidence();
System.out.println("RUL (P10-P90): " + rulCI[0] + " - " + rulCI[2] + " hours");

Trend Analysis

ConditionBasedReliability.TrendAnalysis trend = cbr.calculateTrend("vibration");

System.out.println("Trend Direction: " + trend.getDirection());  // INCREASING, STABLE, DECREASING
System.out.println("Rate of Change: " + trend.getRateOfChange() + " per month");
System.out.println("Time to Alarm: " + trend.getTimeToThreshold() + " days");
System.out.println("Confidence: " + (trend.getConfidence() * 100) + "%");

Failure Probability

// Probability of failure within time horizon
double failProb30d = cbr.calculateFailureProbability(30 * 24);  // 30 days
double failProb90d = cbr.calculateFailureProbability(90 * 24);  // 90 days

System.out.println("Failure probability (30 days): " + (failProb30d * 100) + "%");
System.out.println("Failure probability (90 days): " + (failProb90d * 100) + "%");

Recommendations

String action = cbr.getRecommendedAction();
System.out.println("Recommended Action: " + action);

// Possible recommendations:
// - "Continue normal operation"
// - "Increase monitoring frequency"
// - "Schedule maintenance within 30 days"
// - "Plan maintenance within 14 days"
// - "Immediate maintenance required"

Integration with Risk Analysis

With Dynamic Simulation

// Update equipment MTBF based on condition
DynamicRiskSimulator sim = new DynamicRiskSimulator("Condition-Based Risk");
sim.setBaseProductionRate(100.0);

// Use adjusted MTBF instead of baseline
double adjustedMTBF = cbr.calculateAdjustedMTBF();
sim.addEquipment("Compressor", adjustedMTBF, 72, 1.0);

DynamicRiskResult result = sim.runSimulation();

With Real-time Monitor

RealTimeRiskMonitor monitor = new RealTimeRiskMonitor("Platform", "P-001");
monitor.registerEquipment("C-200", "Compressor", cbr.getBaselineMTBF());

// Update health from CBR
double health = cbr.calculateHealthIndex();
monitor.updateEquipmentHealth("C-200", health);

Use Cases

1. Predictive Maintenance Scheduling

// Find all equipment with RUL < 30 days
List<ConditionBasedReliability> equipmentList = getAllEquipmentCBR();
List<MaintenanceTask> schedule = new ArrayList<>();

for (ConditionBasedReliability eq : equipmentList) {
    double rul = eq.estimateRUL();
    if (rul < 30 * 24) {  // Less than 30 days
        MaintenanceTask task = new MaintenanceTask();
        task.equipment = eq.getEquipmentId();
        task.priority = (rul < 7 * 24) ? "HIGH" : "MEDIUM";
        task.recommendedDate = calculateDate(rul * 0.8);  // 80% of RUL
        schedule.add(task);
    }
}

schedule.sort((a, b) -> Double.compare(a.rul, b.rul));

2. Spare Parts Planning

// Estimate parts needed based on RUL
Map<String, Integer> sparesNeeded = new HashMap<>();

for (ConditionBasedReliability eq : equipmentList) {
    double failProb90d = eq.calculateFailureProbability(90 * 24);
    
    if (failProb90d > 0.3) {
        // High probability of needing spares
        String[] parts = getEquipmentParts(eq.getEquipmentId());
        for (String part : parts) {
            sparesNeeded.merge(part, 1, Integer::sum);
        }
    }
}

3. Risk-Based Inspection

// Prioritize inspections based on health degradation
List<InspectionTask> inspections = new ArrayList<>();

for (ConditionBasedReliability eq : equipmentList) {
    double health = eq.calculateHealthIndex();
    double degradationRate = eq.calculateTrend("vibration").getRateOfChange();
    
    double priority = (1 - health) * 0.6 + degradationRate * 0.4;
    
    InspectionTask task = new InspectionTask();
    task.equipment = eq.getEquipmentId();
    task.priority = priority;
    task.inspectionType = determineInspectionType(eq);
    inspections.add(task);
}

inspections.sort((a, b) -> Double.compare(b.priority, a.priority));

Output Format

JSON Export

String json = cbr.toJson();

Example output:

{
  "equipmentId": "C-200",
  "equipmentName": "Export Compressor",
  "baselineMTBF": 12000,
  "operatingHours": 15000,
  "healthAssessment": {
    "healthIndex": 0.68,
    "adjustedMTBF": 8160,
    "estimatedRUL": 2450,
    "recommendedAction": "Schedule maintenance within 30 days"
  },
  "indicators": [
    {
      "name": "vibration",
      "currentValue": 8.5,
      "healthyValue": 5.0,
      "failureThreshold": 15.0,
      "normalizedValue": 0.35,
      "weight": 0.35,
      "trend": {
        "direction": "INCREASING",
        "ratePerMonth": 0.7,
        "timeToThreshold": 45
      }
    }
  ],
  "failureProbability": {
    "30days": 0.12,
    "90days": 0.35,
    "180days": 0.58
  }
}

Best Practices

  1. Indicator Selection: Choose indicators that directly correlate with failure modes

  2. Threshold Setting: Use OEM recommendations and historical failure data

  3. Weight Calibration: Adjust weights based on failure mode analysis (FMEA)

  4. Data Quality: Ensure sensor calibration and data validation

  5. Baseline Updates: Recalibrate after major maintenance

  6. Integration: Connect to existing CMMS/EAM systems

Standards References