64 / 2025-03-30 10:49:10

A Distributed Condition Monitoring and Fault Diagnosis Method of Wind Turbine Using SCADA Data

wind turbine, condition monitoring, fault diagnosis, distributed modeling, cointegrated analysis

A distributed monitoring and diagnosis framework is proposed for wind turbines, which consists of three modules: data decomposition, condition monitoring and fault diagnosis. Considering the difference of temporal characteristics in SCADA

data, this paper decomposes high-dimensional operation data into stationary and nonstationary blocks based on the stationary test. Different monitoring and diagnostic strategies are applied to the stationary and nonstationary blocks. This paper first proposes a distributed monitoring scheme combining principal component analysis (PCA) and sparse cointegration analysis (SCA), which extracts faulty features from stationary and nonstationary parts respectively, and applies multivariate statistical process monitoring (MSPM) to detect faulty samples. Then, a fault isolation method based on distributed reconstruction is proposed to filter the process variables affected by the fault. The proposed method

is applied in the operation and maintenance of a real wind farm, and the performance advantages are verified by the comparison with popular benchmarks.