A condition monitoring system for wind turbine generator temperature by applying multiple linear regression model

The development and implementation of condition monitoring system become very important for wind industry with the increasing number of failures in wind turbine generators due to over temperature especially in offshore wind turbines where higher maintenance costs than onshore wind farms have to be paid due to their farthest locations. Monitoring the wind generators temperatures is significant and plays a remarkable role in an effective condition monitoring system. Moreover, they can be easily measured and recorded automatically by the Supervisory Control and Data Acquisition (SCADA) which gives more clarification about their behavior trend. An unexpected increase in component temperature may indicate overload, poor lubrication, or possibly ineffective passive or active cooling. Many techniques are used to reliably predict generator's temperatures to avoid occurrence of failures in wind turbine generators. Multiple Linear Regression Model (MLRM) is a model that can be used to construct the normal operating model for the wind turbine generator temperature and then at each time step the model is used to predict the generator temperature by measuring the correlation between the observed values and the predicted values of criterion variables. Then standard errors of the estimate can be found. The standard error of the estimate indicates how close the actual observations fall to the predicted values on the regression line. In this paper, a new condition-monitoring method based on applying Multiple Linear Regression Model for a wind turbine generator is proposed. The technique is used to construct the normal behavior model of an electrical generator temperatures based on the historical generator temperatures data. Case study built on a data collected from actual measurements demonstrates the adequacy of the proposed model.