DriftPulse

DriftPulse is a web app (with optional mobile alerts) for monitoring sensor health in industrial and building systems where “the reading” slowly becomes wrong over weeks or months. It ingests time-series data from common sources (OPC UA, Modbus gateways, MQTT brokers, CSV exports) and runs drift and plausibility checks: cross-sensor consistency (e.g., supply vs return temps), expected physical bounds, step-change detection after maintenance, and calibration interval tracking. The app produces a simple “trust score” per sensor and a prioritized queue of likely-bad instruments, with evidence charts you can show to maintenance or QA. It’s a combination traditional + AI app: traditional rules for hard constraints, plus lightweight anomaly models to flag subtle drift patterns. Reality check: it won’t replace a full CMMS or enterprise historian; it’s a focused layer that sits on top and reduces wasted troubleshooting time.