StackGuard

StackGuard is a web + mobile AI app for electrolyzer operators that predicts stack degradation and flags likely failure modes early (membrane pinholes, catalyst poisoning, drying/flooding, hot spots). It ingests time-series data from plant historians (OPC UA/PI exports), power profiles, water quality logs, and maintenance events, then produces a simple “health score” per stack and a ranked list of actionable checks. The goal is not fancy dashboards; it’s fewer forced outages and better timing for stack swaps, water treatment interventions, and operating setpoint changes. It also tracks warranty/guarantee conditions (ramp rates, current density limits, inlet conductivity) and alerts when operations drift toward voiding coverage. Realistically, accuracy will be imperfect at first—so the MVP focuses on conservative anomaly detection, transparent explanations, and operator feedback loops to improve models over time.