Normal sample Flagged anomaly Regression trend line
Drag to pan · scroll / pinch to zoom

Reactor Sensor Correlation & Anomaly Stream (2D)

Two virtual reactor sensors — temperature, held near its setpoint by a proportional thermostat feedback loop, and pressure, physically coupled to it — stream new readings several times a second. Each sample joins a 2D scatter plot where the X and Y axes are the two sensor values, so a tight diagonal band of points is a visual readout of correlation strength; older samples fade out as they age past the correlation window. A sliding-window Pearson coefficient, a rate-of-change readout, and a regression-residual z-score anomaly detector all update live alongside a rolling strip chart of both raw streams. Trigger a simulated malfunction and watch the correlated relationship break in real time, exactly as an operator's dashboard would flag a fault.