Supervisory Control and Data Acquisition Analysis for Wind Turbine Maintenance Management

Abstract

Wind energy is growing to become a competitive energy source. An efficient wind turbine maintenance management is required for ensuring the reliability of the energy production and the costs reduction. Supervisory control and data acquisition system provide information about the condition of the wind turbine by signals of the different subsystems and alarm activations in case of failure or malfunction. Due to the volume and variety of the data, operators require advanced analytics to control the performance of the wind turbines and the identification and prediction of failures. The novelty proposed in this work is based on statistical analysis for analyzing supervisory control and data acquisition data to optimize the use of the data in neural networks. The first phase is the alarm analysis, quantifying the critical alarms regarding on the number and time of activation. A filtering algorithm is developed for considering only interest periods with enough range to make the study. The second phase is based on the initial data treatment, classifying alarms and signals identifying the interest time periods. Neural network is defined and trained for evaluating the signal trends, with the aim of detecting the alarm activations cause. This information will be used in the maintenance management plan for programming maintenance tasks.