Predictive MaintenancePrognostics and Health Management (PHM) solutions for mission critical systems require a comprehensive methodology for proactively detecting and isolating failures, avoiding false alarms and missed alarms, recommending and unambiguously guiding condition-based maintenance actions, and estimating the remaining useful life (RUL) of critical components and associated subsystems. 
Binatix data-driven and modular approach allows customers to enhance existing CBM platforms by adding our best-in-class data analysis software modules. Remaining Useful Life (RUL) estimation Binatix achieved top score, 2nd in the professional category, in NASA’s PHM’08 CBM benchmark of RUL Estimation. HD-RNN identifies faults, decay patterns, and spatiotemporal dependencies between multiple sensors, and accurately predicts remaining useful life of complex systems. Our data driven approach is especially powerful when model-based and rule-based solutions are infeasible. The HD-RNN classifier detects gradual deviation from nominal conditions given a data set of samples representing normal (non-faulty) inputs. Deviation from nominal conditions is classified to either known faults or unknown faults. Detecting Unknown Faults is essentially equivalent to anomaly detection above. The HD-RNN classifier detects gradual deviation from nominal conditions given a data set of samples representing normal (non-faulty) inputs. Deviation from nominal conditions is classified to either known faults or unknown faults. Detecting Unknown Faults is essentially equivalent to anomaly detection above. |