Wind power forecasting of an offshore wind turbine based on high-frequency SCADA data and deep learning neural network

Abstract

Accurate wind power forecasting is essential for efficient operation and maintenance (O&M) of wind power conversion systems. Offshore wind power predictions are even more challenging due to the multifaceted systems and the harsh environment in which they are operating. In some scenarios, data from Supervisory Control and Data Acquisition (SCADA) systems are used for modern wind turbine power forecasting. In this study, a deep learning neural network was constructed to predict wind power based on a very high-frequency SCADA database with a sampling rate of 1-second. Input features were engineered based on the physical process of offshore wind turbines, while their linear and non-linear correlations were further investigated through Pearson product-moment correlation coefficients and the deep learning algorithm, respectively. Initially, eleven features were used in the predictive model, which are four wind speeds at different heights, three measured pitch angles of each blade, average blade pitch angle, nacelle orientation, yaw error, and ambient temperature. A comparison between different features shown that nacelle orientation, yaw error, and ambient temperature can be reduced in the deep learning model. The simulation results showed that the proposed approach can reduce the computational cost and time in wind power forecasting while retaining high accuracy.

A Critical Review of Wind Power Forecasting Methods—Past, Present and Future

“…ANNs, as one of the most commonly used methods for wind power prediction, can identify the non-linear relationships between input features and output data [22]. One of the reasons for the tendency to use neural networks is to avoid the complexity of the mechanical structure in wind turbines [23]. Typically, an ANN model consists of an input layer, one or more hidden layers [11], and an output layer, where the historical data/features are fed for training and testing (see Figure 1).…”

“…Comparison results showed that IF is a more effective way of providing accurate forecasting, especially when the investigated data do not follow the normal distribution. In another paper [23], the authors critically evaluated eleven features from a 7 MW wind turbine in Scotland, including four wind speeds at different heights, average blade pitch angle, three measured pitch angles for three blades, ambient temperature, yaw error and nacelle orientation. The results revealed that the blade pitch angle had the greatest effect on the performance of the prediction model, even more than wind speed and wind shear.…”

“…Lin and Liu [23], in an effort for reducing the computational time, removed minor influencer features in the proposed model, including air temperature, nacelle orientation and yaw error. Even though this reduction resulted in a small saving of processing time (0.77 min) for a single wind turbine, the saved simulation time can be considered while taking into account a typical wind farm comprising of more than 100 turbines.…”

“…Explained variation estimates the proportion to which a forecasting model scores for the dispersion of a specified dataset. For the best prediction, the value of EVS is 1.0, while lower scores represent worse prediction [23]. The EVS is defined as follows:…”

“…Wind direction is another factor with the effect on power generation. Considering the current design of a wind turbine, turbines are allowed to face into the wind during the time of operation [23]. Singh et al [8] showed that wind speed and wind direction were the top two influence factors on wind power prediction through their MLP prediction model.…”