Abstract:Under the background of widespread utilization of wind energy, the improvement of power generation efficiency and the reduction of operation and maintenance costs have increased the demand for anomaly detection and localization of wind turbines through fault propagation analysis, while the large-scale data collected by SCADA systems provide sufficient data support and promising assistance for this purpose. It is an effective and cost-saving solution to this demand to construct a high-precision wind turbine digital twin model under normal operating conditions based on multivariate monitoring data, and to realize root cause localization based on prediction deviation and backtracking analysis. This paper proposes a root cause localization framework for wind turbines based on explicit-implicit knowledge fusion and multivariate time-series graph neural networks, fully utilizing the inter-variable dependencies mining capability of graph structure learning and the temporal prediction advantage of time-series graph neural networks. Five types of explicit knowledge based on explicit rules are constructed to describe the explicit relationships between monitoring variables. The fused knowledge (i.e., graph adjacency matrix) is obtained by fusing the explicit knowledge with the implicit knowledge hidden in the data through a quasi-hard-attention mechanism based on the idea of graph structure learning. A high-precision digital twin model for wind turbines based on fused knowledge is constructed using the multivariate time-series graph neural network. The anomaly degree of monitoring variables’ state is defined based on the prediction deviation, with which the propagation of fault's state over time can be accurately described, and the root cause localization through backtracking analysis can be achieved. Case studies are conducted using field fault data from a wind farm, a sugar factory and a thermal power plant. The results show that the proposed framework can provide a solution for studying the fault propagation process and root cause localization, also demonstrating certain robustness and generalizability.

Fusion Structure-based Fault State Prediction of Wind Turbine Spindle

Fatigue Load Prediction of Large Wind Turbine by Big Data and Deep Learning

Wind Turbine Fault Detection with Multimodule Feature Extraction Network and Adaptive Strategy

Fuzzy Reliability Evaluation and Machine Learning-based Fault Prediction of Wind Turbines

Typical fault prediction method for wind turbines based on an improved stacked autoencoder network

Research on Wind Turbine Fault Diagnosis Technology Based on Big Data

Deep Learning Approach for Sensor Data Prediction and Sensor Fault Diagnosis in Wind Turbine Blade

Early fault prediction for wind turbines based on deep learning

Fault Detection of Wind Turbine Generator Bearing Using Attention-Based Neural Networks and Voting-Based Strategy

A model fusion optimization strategy for lithium mill equipment state prediction

Wind turbine failure prediction and health assessment based on adaptive maximum mean discrepancy

Root Cause Localization for Wind Turbines Using Physics Guided Multivariate Graphical Modeling and Fault Propagation Analysis

Fault prediction model in wind turbines using deep learning structure with enhanced optimisation algorithm

SPLSTM: A Prediction Model for Operational Faults in Rotating Machinery

Gearbox Fault Prediction Of Wind Turbines Based On A Stacking Model And Change-Point Detection

Wind Turbine Rotor Fault Prediction Based on Convolutional Neural Network

Prediction of Icing Fault of Wind Turbine Blades Based on Deep Learning

An early fault warning and diagnosis model based on a two-step analysis strategy for wind turbine bearings

Misalignment Fault Diagnosis for Wind Turbines Based on Information Fusion

Machine Learning for Wind Turbine Fault Prediction through the Combination of Datasets from Same Type Turbines

Early Fault Detection of Wind Turbines Based on Operational Condition Clustering and Optimized Deep Belief Network Modeling