An end-to-end, real-time solution for condition monitoring of wind turbine generators

摘要: Data-driven wind generator condition monitoring systems largely rely on multi-stage processing involving feature selection and extraction followed by supervised learning. These stages require expert analysis, are potentially error-prone do not generalize well between applications. In this paper, we introduce a collection of end-to-end Convolutional Neural Networks for advanced turbine generators. End-to-end models have the benefit utilizing raw, unstructured signals to make predictions about parameters interest. This makes it easier scale an existing new predictive tasks (e.g., failure types) since extracting steps required. automated achieve low Mean Squared Errors in predicting operational state (40.85 Speed 0.0018 Load) high accuracy diagnosing rotor demagnetization failures (99.67%) only raw current signals. We show how create, deploy run proposed real-time setting using laptop connected test rig via data acquisition card. Based sampling rate 5 kHz, stored efficient time series database monitored dynamic visualization framework. further discuss options understanding decision process behind made models.