Skip to main content
The dashboard displays real-time sensor telemetry and derived health metrics to provide operators with instant visibility into asset condition.

​
Primary Sensor Metrics

Four key metrics are displayed in the Status Cards row:

Voltage (V)

AC voltage measured in volts. Baseline target: 230.0 V

Current (A)

Electrical current measured in amperes. Baseline target: 12.0 A

Power Factor

Ratio of real power to apparent power (0.0 to 1.0). Baseline target: 0.92

Vibration (g)

Vibration intensity measured in g-force. Baseline target: 0.15 g

​
Health Score Ring

The Health Summary card displays a circular progress ring showing the asset’s overall health on a scale of 0-100.

​
Color Coding

Health scores are color-coded based on risk level:
Health ScoreRisk LevelColorMeaning
75-100LOWGreenAsset is healthy, normal operation
50-74MODERATEYellow/OrangeMinor degradation detected
25-49HIGHOrangeSignificant anomalies, schedule maintenance
0-24CRITICALRedImmediate maintenance required
Health scores are derived from the Cumulative Degradation Index (DI), which increases monotonically as the system experiences anomalies. The formula is: health_score = 100 × (1 - DI)

​
Maintenance Window

The dashboard displays an estimated maintenance window in days based on the current degradation rate: 
RUL (hours) = (1.0 - DI) / max(damage_rate, 1e-9)
maintenance_window_days = RUL / 24
Example Display:
  • ~60 days — Asset is healthy, no urgency
  • ~19 days — MODERATE risk, plan preventive maintenance
  • ~4 days — HIGH risk, schedule maintenance soon
  • < 1 day — CRITICAL risk, immediate action required

​
Baseline Target Values

After calibration, the system learns baseline target values from healthy sensor data. These targets are displayed on Status Cards as small gray text below the live reading:

Example: Voltage Status Card

230.2 V
Target: 230.0 V

​
How Baselines Are Calculated

During the CALIBRATING state (see System Controls), the backend collects 100+ sensor readings and computes: 
baseline_targets = {
    "voltage_v": round(mean(voltage_samples), 1),
    "current_a": round(mean(current_samples), 1),
    "power_factor": round(mean(power_factor_samples), 2),
    "vibration_g": round(mean(vibration_samples), 3)
}
These targets are returned in the /api/health response and displayed on Status Cards for quick visual comparison.

​
Streaming Chart Metrics

The Streaming Sensor Telemetry chart plots three signals over a 60-second sliding window:
  • Color: Blue (#3b82f6)
  • Y-Axis Range: 220-240 V
  • Use Case: Detect voltage spikes or brownouts

​
Anomaly Shading

When the risk level is not LOW, red dashed vertical lines and shaded regions appear on the chart to highlight anomalous time periods.
Anomaly markers use majority-rules aggregation (≥15/100 points) and 2-second debounce to suppress noise and prevent flickering.

​
Health Score Calculation Details

The health score is derived from two machine learning models:

​
Dual-Model Architecture

ModelFeaturesSampling RateF1 ScoreAUC-ROC
Legacy (v2)6 features1 Hz78.1%1.000
Batch (v3)16 features100 Hz99.6%1.000
The Batch model (v3) is the primary detector. It extracts 16 statistical features (mean, std, peak-to-peak, RMS) from 100-sample windows to detect subtle faults like Jitter (high variance with normal mean).

​
Dead-Zone for Healthy Noise

To prevent healthy fluctuations from accumulating damage, the system uses a dead-zone threshold: 
HEALTHY_FLOOR = 0.65
if batch_score < HEALTHY_FLOOR:
    effective_severity = 0.0  # No damage accumulated
else:
    effective_severity = (batch_score - HEALTHY_FLOOR) / (1.0 - HEALTHY_FLOOR)
Only scores above 0.65 contribute to the Cumulative Degradation Index (DI).

​
Status Card Indicators

Each Status Card changes color based on the sensor value and fault state:

Green Border

Value is within normal range and no fault detected

Yellow Border

Value deviates from baseline but not critical

Red Border

Fault detected or value is critically out of range

​
Status Logic Example (Vibration)

function vibrationStatus(value) {
    if (value == null) return 'gray';
    if (value < 0.18) return 'green';  // Healthy
    if (value < 0.25) return 'yellow'; // Warning
    return 'red';                      // Critical
}
If is_faulty flag is true, the status is overridden to red regardless of value.

​
Next Steps

System Controls

Learn how to calibrate the system and inject faults

Operator Logs

Log maintenance events to create ground-truth data

Build docs developers (and LLMs) love

Get started for freeTalk to us