Abstract
To maximize energy extraction, the nacelle of a wind turbine follows the wind direction. Accurate prediction of wind direction is vital for yaw control. A tandem hybrid approach to improve the prediction accuracy of the wind direction data is developed. The proposed approach in this paper includes the bilinear transformation, effective data decomposition techniques, long-short-term-memory recurrent neural networks (LSTM-RNNs), and error decomposition correction methods. In the proposed approach, the angular wind direction data is firstly transformed into time-series to accommodate the full range of yaw motion. Then, the continuous transformed series are decomposed into a group of subseries using a novel decomposition technique. Next, for each subseries, the wind directions are predicted using LSTM-RNNs. In the final step, it decomposed the errors for each predicted subseries to correct the predicted wind direction and then perform inverse bilinear transformation to obtain the final wind direction forecasting. The robustness and effectiveness of the proposed approach are verified using data collected from a wind farm located in Huitengxile, Inner Mongolia, China. Computational results indicate that the proposed hybrid approach outperforms the other single approaches tested to predict the nacelle direction over short-time horizons. The proposed approach can be useful for practical wind farm operations.
Introduction
Wind energy generation is expanding with about 12% of world’s electricity to be supplied by 2020 (). Compared with the traditional form of power generation, wind energy has the advantages of zero pollution and low operation cost. Hence, it has become one of the fastest growing renewable energy power supplies globally ().
Although it has obvious advantages over others, wind energy still faces technical challenges due to the characteristics of chaos, randomness, and intermittence which make the wind data complex. The wind direction is one of the most complex aspect of the wind data due to its high dynamics in both spatial and temporal domains. To follow the wind direction, the nacelle of a wind turbine orientes the controlling of yaw and maximizes the energy output. For most efficient energy extraction, the nacelle orientation of a wind turbine needs to agree with wind direction which calls for accurate and prediction of the wind direction ().
According to literature review, statistical approaches based on meteorological and geographic information are widely applied to forecast wind direction (; ; ). applied a neural Kriging method to spacially estimate the distribubiton of wind directions. developed autoregressive moving average (ARMA) model to forecast the short-term wind directions. used a mixture of Von Mises models to fit the wind direction series.
Therefore, machine learning adoptions for wind direction forecasting have evolved from the classic approach to deep learning, which is then improved in this study (; ). In the wind direction forecasting sector, selected least-square support vector machines (LS-SVM) to predict the wind directions. developed artificial neural networks to forecast the short-term wind directions. developed an adaptive neuro-fuzzy inference system to predict the wind directions. improved the wind direction forecasting using the echo state network (ESN) which is a deep-learning algorithm. integrated the ESN network with IFPA optimization algorithm and developed a two-step deep-learning wind direction framework.
Considering the complexity and high dynamics of the wind direction series, additional measures are essential to study in the pattern inside. Even though deep learning algorithms have achieved promising results in the field of time-series prediction, it is still challenging for a single deep-learning approach to adapt all wind direction patterns. To further improve the prediction performance, hybrid prediction models are considered to be the mainstream since last year. The signal decomposition is one of the most popular components within the hybrid models published. It contains wavelet decomposition (), empirical mode decomposition (EMD) (), complete ensemble empirical model decomposition (CEEMD) (), complete EEMD with adaptive noise (CEEMDAN) (), and the improved CEEMDAN (ICEEMDAN) (). In particular, the ICEEMDAN has demonstrated its superior performance in decomposing a complex signal into a finite number of intrinsic mode functions with transient frequencies. The decomposed subseries contains the detailed characteristics of the signal and can essentially reflect the spatial and temporal patterns of the wind direction series ().
Based on the above considerations, in this research, we propose a new hybrid approach combining ICEEMDAN and error correction methods for short-term wind direction forecasting. First, the angular wind direction data has been transformed via bilinear transformation. Then, the transformed wind direction series are decomposed into a series of relatively simple subseries by the ICEEMDAN modules. Next, the LSTM-RNN is established as the prediction module to predict each sub-series. After that, the prediction errors are obtained and decomposed by ICEEMDAN modules. The statistical ARIMA model is used to predict the error subsequence and compute the prediction error. In the last step, the final prediction of the wind direction is made by summing all predicted subseries together with current predicted error and then transformed into angular data by inverse bilinear transformation.
The major contribution of this research can be summarized as follows: First, the wind direction forecasting system based on ICEEMDAN decomposition, LSTM-RNN and error correction has been proposed; Second, the comparative analysis is performed against other benchmarking deep-learning algorithm; Third, the experiments were performed in different seasons to explore seasonal patterns of wind directions.
The remainder of the manuscript is configured as follows. In Section “dataset description and transformation”, it summarizes the data collection process and patterns inside the wind direction dataset. In Section “methodologies”, it introduces the ICEEMDAN decomposition, LSTM-RNN, error correction, benchmarking deep-learning algorithms, and error correction procedures. The experimental results are provided in Section “experimental results” and the Conclusion is made in Section “conclusions” respectively.
Dataset Description and Transformation
Data Analysis
In this study, the data has been collected during the year of 2020 from a wind farm namely Huitengxile wind power plant in Inner Mongolia, Northern China. It is one of the largest wind farms in Asian and it’s located in the suburbs between Chaha’er youyi zhongqi and Ulanqab city. The whole wind farm has multiple wind turbines that are distributed in an open flat grassland which provides rich wind resources. The prevailing wind directions are northwest and southeast which are very stable in recent years. The location and the annual wind rose diagrams has been illustrated in Figure 1 below.
FIGURE 1
According to Figure 1B, the two prevailing wind directions, around 180° and 315° are visible. The geographic center coordinate is 112°40′E and 41°05′N. It’s annual average wind speed at 10 m height is 7.2 m/s and its annual average wind speed at 40 m height is 8.8 m/s. In the wind farm, the annual average air density is 1.07 kg/m and it contains an effective wind speed of 5–25 m/s with strong stability and high quality.
The data used in this research has been collected by the supervisory control and data acquisition (SCADA) system. Usually, data on more than 100 parameters at 10 s intervals is collected and stored in a SCADA system. The SCADA collected data of individual wind turbines is streamed to a central computer for condition monitoring, performance evaluation, and other forms of analysis.
In this research, the SCADA data collected at 20 wind turbines over the period of 2020 has been analyzed. According to Figure 1B, there are two annual prevailing wind directions and it can be partitioned into four seasons independently as illustrated with the wind roses in Figure 2. In the fall and spring, two prevailing wind directions around 150° and 315°are observed. In the winter and summer, one prevailing wind direction is noted. Since the wind direction data is captured as a discrete angular variable, it needs to be transformed for modeling. A bilinear transformation of the angular wind direction is applied in the next section.
FIGURE 2
Bilinear Transformation of Angular Data
The value of wind direction ranges from 0° to 360°. It is likely that the wind direction may change from the interval, i.e (0°, 10°) to (350°, 360°). Practice shows that bilinear transformation is a better way for transforming discrete wind direction data to continuous data than the sine and cosine transformation (). Geometrically, the two intervals are close to each other and therefore this change would lead to a large prediction error (). To avoid such error, transformation of the discrete angular variable into a standardized continuous variable is essential. One option is to use a sine and cosine transformation which is not the best approach due to two variables needed for prediction which enlarges the prediction errors. A better option is to apply a bilinear transformation ().
The bilinear transformation maps the analog plane (s-plane) into the digital plane (z-plane) () (see Figure 3). The transformation function, the ratio of two polynomials (), is expressed in Eq. 1.where: s is the original value of angular variable in s-plane; T is the time interval of the transformation. The bilinear transformation expressed the angular variable between 0° and 360° as continuous and normalized. The inverse bilinear transformation is expressed in Eq. 2.where: s is the inversed value of angular variable in s-plane; and H(s)z is the transformed angular variable.
FIGURE 3
Since the wind direction data is noisy, a bilinear transformation function acting as a low-pass filter in the continuous-time domain reduces the noise (). A prediction model developed with the transformed data is more accurate than the model based on the discrete time-series angular data.
Methodologies
The use of deep learning algorithms in regression, multi-class classification, collaborative filtering, and graphic learning is growing (). The concept of deep learning originates from research in neural networks and it avoids the local optima dilemma. However, any single deep learning algorithms can offer limited extraction of patterns inside the dataset. Hybrid frameworks containing multiple deep learning algorithm is becoming the new mainstream in academia.
Iceemdan
In this research, the improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN) is served as the major module in the hybrid forecasting framework. It is considered as an improvement on empirical mode decomposition (EMD) which decomposes the wind directions in the temporal domain ().
The time-series of wind direction can be expressed as the sum of multiple IMFs and the residual after the ICEEMDAN decomposition which can be expressed in Eq. 3 as follows:
The amplitude energy , , …, of the IMFs is calculated as Eq. 4:where N denotes the total number of sampling points of the jth IMF. Assuming that the energy carried by r(t) can be ignored, the total energy of the transformed direction series can be expressed as Eq. 5 as follows:
To remain the data in the same magnitude, the amplitude of the IMFs is normalized to facilitate the subsequent calculations and the impact of singular data has been reduced. Hence, the energy entropy of the ICEEMDAN framework can be expressed as Eq. 6 below:
Compared with other decomposition methods, the ICEEMDAN can not only reduce the noise in the original time-series data but also reduce the residual spurious pattern problems based by signal overlap. Thus, the decomposed subseries gains more orthogonality among each other and it can provide more accurate reconstruction of the original series.
Short-Term Wind Direction Forecasting Using ICEEMDAN
To integrate the wind direction series with the ICEEMDAN modules, the implementations are introduced as follows ():
Step 1: Compute the local means of realizations using the EMD algorithm described in Eq. 7:where ; compute the standard deviation; and is the reciprocal of the desired signal-to-noise ratio between the first added noise and the analyzed signal.
Step 2: Compute residual term R1 in the first component using Eq. 8:
Step 3: Compute the first mode at the first stage (k = 1) using Eq. 9:
Step 4: Estimate the second residue as the average of local means of the realizations and then define the second mode using Eq. 10 as follows:
Step 5: For the other terms (k = 3, … ,K) of residuals, they can be computed by Eq. 11:
Step 6: Compute the other terms (k = 3, … ,K) of the mode by Eq. 12:
Step 7: Implement step 4 for the next iteration.
For the transformed wind direction series, the IMF components are obtained via the above steps which can be illustrated by the diagram presented in Figure 4.
FIGURE 4
Long-Short-Term Memory Recurrent Neural Network
A major drawback of the classical deep neural networks is that they do not have memory of the past periods. The time series information such as the past clusters of seasonal patterns and seasonal trend may not be reflected (). Introduced by and , the long-short-term memory recurrent neural network (LSTM-RNN) matches the needs and it is used in this paper to predict wind direction.
The long-short-term memory recurrent neural network (LSTM-RNN) contains units called memory blocks composed of memory cells with self-connections storing temporal states. Each memory block includes an input and output gate. The input gate controls the flow of input data into the cell. The output gate controls the output data flow into the rest of the network (). In addition, the LSTM-RNN has peephole connections () from its internal cells to the gates in the same cell to learn precise timing of the output. The architecture of LSTM-RNN is illustrated in Figure 5.
FIGURE 5
With a long-short-term memory recurrent neural network (LSTM-RNN) architecture, the mapping from an input to an output layer is iteratively computed from Eqs 13–18 ().where: W are the weight matrices (i.e., Wix is the weight matrix from the input to the input layer; Wic, Wfc, Woc are diagonal weight matrices of the peephole connections ()); bi,bf,bo, and bc are the bias vectors; m is the cell output activation vector; sig () is the sigmoid function; i, f, o, and c are the input gate, forget gate, output gate, and cell activation vectors, respectively, with all having the same size as the cell output activation vector m; is the element-wise product of the vectors; and g () and h () are the cell input and cell output activation functions, respectively.
Benchmarking Machine Learning Algorithms
Comparative analysis is performed in this research against the other benchmarking popular deep learning algorithms. All algorithms tested here are using the same ICEEMDAN framework as described in Section “Short-term Wind Direction Forecasting using ICEEMDAN”. The benchmarking deep learning algorithms compared includes deep neural network (DNN) (; ; ), deep belief network (DBN) (; ), kernel-based extreme learning machine (KELM) (; ), and gated recurrent unit network (GRU) (; ).
The DNN is a fully connected feedforward network that consists of a cascade of multiple layers and hidden units. It’s structure with multiple processing layers enables it to handle highly nonlinear patterns inside the dataset. The deep temporal representations in the temporal domain can be effectively extracted by DNN.
Similar to DNN, the DBN consists of multiple layers of restricted Boltzmann machines (RBMs). It also contains a supervised regression layer stacked on the top of all RMBs for classification or regression tasks. Inside each RBM, it contains an input layer and a hidden layer with hidden-to-all-visible connections.
The ELM is a single hidden layer feedforward network. Instead of conventional back-propagation, it uses Penn-Moore pseudo-inverse to compute the wights between the hidden layer and output layer. The KELM is the improvement of vanilla ELM which uses the kernel matrix to replace the randomly initialized weights between the input layer and output layer. The most popular applied kernel functions include RBF, linear function, and polynomial function.
The GRU is another type of recurrent neural network other than LSTM-RNN proposed by . In a typical GRU unit, it has one less gate than the LSTM unit and consists of two gates: the reset gate and the update gate. Hence, the GRU is also popular in modeling time-series dataset.
Measurement Matrices
To assess prediction accuracy of the proposed deep learning model, four metrics are computed: the MAE [Mean absolute error (Eq. 19)], the MAPE [Mean absolute percentage error (Eq. 20)], the MSE [Mean square error (Eq. 21)], and the RMSE [Root mean square error (Eq. 22)].where: oj is the jth predicted wind direction; tj is the jth measured wind direction; and N denotes total number of samples.
Error Correction
To improve the forecasting accuracy, the error correction is implemented in this research. First, after the forecasting outcome produced by each LSTM-RNN, the error series E(t) of the training dataset can be computed by comparing the original transformed wind direction series. The step can be expressed in Eq. 23 as follows:where is the final forecasted transformed series in the first module; and is the actual transformed wind direction series after bilinear transformation.
The forecasted errors of wind direction series E(t) are oscillatory in the time-series domain (). The relationship between oscillatory and decaying property of the wind direction errors can be represented by an ARIMA model which predicts the errors. In detail, the ARIMA can be constructed by computing autocorrelation expressed with the autocorrelation factor (ACF) (See. Eq. 24) and the partial autocorrelation factor (PACF)) (See. Eq. 25). Here, Cov() denotes the covariance; Var() denotes the variance; and Corr () denotes the Pearson’s correlation coefficient.
For each IMF, an ARIMA model is developed and then all outcomes of each ARIMA are integrated to obtain the final error series. Last, as illustrated in Figure 4, the final prediction is achieved by Eq. 26 as follows:where denotes the forecasting results from the LSTMs; denotes the errors forecasted by ARIMA models; and is the final forecasting outcomes.
Experimental Results
Training Strategies
In this section, computational experience with models predicting wind direction is presented. Wind data from four seasons, spring, summer, fall, and winter are used. Prediction of wind direction is conducted using dataset at 10, 20, and 30 s resolution. The prediction horizons are 2, 5, 10 min, and 1 h. The prediction model is expressed in Eqs 27–29.where: f(D) represents the whole framework illustrated in Figure 4; Dt-i is the ith lagged vector containing 1 hour of the historical wind direction data; and denotes the ith lagged transformed wind direction series.
The wind speed and wind direction are 10 s data. One hour of data (360 data points) is used as the input vector. The six-time lagged vectors containing 1 hour of the historical wind direction data are selected as the input vectors. The wind direction of the 2, 5, 10 min, and 1 h horizon is predicted. The input vector is normalized beforehand and the predicted values are inverse-normalized.
Short-Term Predictions
Based on the training strategy stated in Section “Training Strategies”, experiments with the five selected algorithms have been performed. In all experiments, the wind direction has been predicted for the next 2, 5, 10 min, and 1 h. Experiments have been conducted in each of the four seasons of 2020.
The prediction accuracy results in Figure 6 demonstrate that the long-short-term memory recurrent neural network (LSTM-RNN) performs better over short-term horizons than the other algorithms. Since the LSTM-RNN contains long/short term memory, it produces smaller prediction errors than the DNN, DBN, KELM, and GRU. For the short-term horizons (i.e., 2 and 5 min), prediction accuracy of all five algorithms is similar. However, the LSTM-RNN provides more promising results for longer-term predictions (i.e., 10 min and 1 h) of wind direction.
FIGURE 6
The prediction accuracy in four seasons varies. In the fall and spring season, the prediction errors are larger than the errors in the summer and winter season. This is due to a larger variability of the wind direction over short-term horizons. Hence, training specific prediction models in different seasons is necessary.
Table 1 provide the MAPE for different resolution data (i.e., 10, 20, 30 s) and different prediction horizons (i.e., 2, 5, 10 min, and 1 h) before the error correction. Obviously, the MAPE errors are smaller for the 10 s data than for 20 and 30 s. With the increase of the prediction horizon, the MAPEs increase. The LSTM-RNN algorithm has the smallest MAPE at all resolutions and all prediction horizons. Therefore, it is an effective algorithm for wind direction prediction at short-term horizons.
TABLE 1
| Season | Resolution (s) | Algorithm | Prediction horizon | Season | Resolution (s) | Algorithm | Prediction horizon | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2 min | 5 min | 10 min | 1 h | 2 min | 5 min | 10 min | 1 h | ||||||
| Spring | 10 | DNN | 0.082 | 0.106 | 0.133 | 0.211 | Summer | 10 | DNN | 0.047 | 0.057 | 0.092 | 0.181 |
| DBN | 0.072 | 0.09 | 0.121 | 0.207 | DBN | 0.056 | 0.064 | 0.097 | 0.182 | ||||
| KELM | 0.081 | 0.098 | 0.137 | 0.224 | KELM | 0.062 | 0.074 | 0.106 | 0.186 | ||||
| GRU | 0.077 | 0.095 | 0.125 | 0.211 | GRU | 0.059 | 0.07 | 0.095 | 0.181 | ||||
| LSTM-RNN | 0.072 | 0.091 | 0.108 | 0.195 | LSTM-RNN | 0.047 | 0.056 | 0.084 | 0.171 | ||||
| 20 | DNN | 0.088 | 0.112 | 0.139 | 0.269 | 20 | DNN | 0.049 | 0.061 | 0.099 | 0.221 | ||
| DBN | 0.077 | 0.093 | 0.127 | 0.252 | DBN | 0.056 | 0.065 | 0.105 | 0.227 | ||||
| KELM | 0.085 | 0.102 | 0.141 | 0.273 | KELM | 0.065 | 0.077 | 0.113 | 0.228 | ||||
| GRU | 0.082 | 0.101 | 0.131 | 0.264 | GRU | 0.064 | 0.073 | 0.104 | 0.227 | ||||
| LSTM-RNN | 0.076 | 0.094 | 0.117 | 0.231 | LSTM-RNN | 0.049 | 0.058 | 0.096 | 0.214 | ||||
| 30 | DNN | 0.096 | 0.114 | 0.15 | 0.33 | 30 | DNN | 0.052 | 0.066 | 0.112 | 0.278 | ||
| DBN | 0.082 | 0.099 | 0.139 | 0.316 | DBN | 0.058 | 0.072 | 0.116 | 0.277 | ||||
| KELM | 0.089 | 0.109 | 0.153 | 0.328 | KELM | 0.068 | 0.083 | 0.125 | 0.29 | ||||
| GRU | 0.087 | 0.105 | 0.142 | 0.311 | GRU | 0.063 | 0.078 | 0.116 | 0.286 | ||||
| LSTM-RNN | 0.081 | 0.096 | 0.128 | 0.305 | LSTM-RNN | 0.051 | 0.063 | 0.108 | 0.261 | ||||
| Fall | 10 | DNN | 0.094 | 0.112 | 0.154 | 0.239 | Winter | 10 | DNN | 0.049 | 0.069 | 0.125 | 0.212 |
| DBN | 0.079 | 0.095 | 0.145 | 0.226 | DBN | 0.06 | 0.078 | 0.134 | 0.213 | ||||
| KELM | 0.14 | 0.163 | 0.232 | 0.302 | KELM | 0.077 | 0.091 | 0.125 | 0.213 | ||||
| GRU | 0.109 | 0.128 | 0.153 | 0.232 | GRU | 0.055 | 0.073 | 0.12 | 0.209 | ||||
| LSTM-RNN | 0.075 | 0.092 | 0.133 | 0.211 | LSTM-RNN | 0.035 | 0.053 | 0.097 | 0.183 | ||||
| 20 | DNN | 0.098 | 0.116 | 0.156 | 0.285 | 20 | DNN | 0.053 | 0.072 | 0.127 | 0.247 | ||
| DBN | 0.082 | 0.101 | 0.147 | 0.279 | DBN | 0.064 | 0.084 | 0.136 | 0.264 | ||||
| KELM | 0.145 | 0.167 | 0.224 | 0.365 | KELM | 0.08 | 0.094 | 0.131 | 0.256 | ||||
| GRU | 0.113 | 0.131 | 0.158 | 0.285 | GRU | 0.059 | 0.077 | 0.124 | 0.243 | ||||
| LSTM-RNN | 0.079 | 0.095 | 0.137 | 0.266 | LSTM-RNN | 0.041 | 0.057 | 0.102 | 0.234 | ||||
| 30 | DNN | 0.102 | 0.124 | 0.168 | 0.364 | 30 | DNN | 0.058 | 0.084 | 0.14 | 0.314 | ||
| DBN | 0.086 | 0.11 | 0.16 | 0.353 | DBN | 0.069 | 0.095 | 0.149 | 0.33 | ||||
| KELM | 0.149 | 0.179 | 0.237 | 0.459 | KELM | 0.083 | 0.101 | 0.142 | 0.313 | ||||
| GRU | 0.117 | 0.138 | 0.169 | 0.357 | GRU | 0.064 | 0.087 | 0.136 | 0.321 | ||||
| LSTM-RNN | 0.083 | 0.104 | 0.149 | 0.332 | LSTM-RNN | 0.046 | 0.066 | 0.114 | 0.276 | ||||
Mape of the five algorithms for 2020 before error correction.
Bold only shows the optimal solution results.
Error Correction
To correct the errors made by the ICEEMDAN modules, the ARIMA model has been developed to forecast the errors. In the second part of Figure 4, To illustrate this step, the forecasted errors using ARIMA versus the actual errors produced by LSTM-RNNs are visualized in Figure 7. It is obvious that the aggregated results from ARIMAs can represent the temporal trend of the forecasted errors produced from the first component of the proposed framework.
FIGURE 7
Table 2 provides the MAPE for different resolution data (i.e., 10, 20, 30 s) and different prediction horizons (i.e., 2, 5, 10 min, and 1 h) after the error correction. There exists significant performance for all algorithms tested with respect to the MAPE computed before and after error correction. It validates the effectiveness of implementing error correction in improving the forecasting accuracy of time-series dataset. Meanwhile, the LSTM produces the smallest errors which also demonstrates its superior performance in forecasting wind directions.
TABLE 2
| Season | Resolution (s) | Algorithm | Prediction horizon | Season | Resolution (s) | Algorithm | Prediction horizon | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2 min | 5 min | 10 min | 1 h | 2 min | 5 min | 10 min | 1 h | ||||||
| Spring | 10 | DNN | 0.082 | 0.106 | 0.133 | 0.211 | Summer | 10 | DNN | 0.047 | 0.057 | 0.092 | 0.181 |
| DBN | 0.072 | 0.094 | 0.121 | 0.207 | DBN | 0.056 | 0.064 | 0.097 | 0.182 | ||||
| KELM | 0.081 | 0.098 | 0.137 | 0.224 | KELM | 0.062 | 0.074 | 0.106 | 0.186 | ||||
| GRU | 0.077 | 0.095 | 0.125 | 0.211 | GRU | 0.059 | 0.077 | 0.095 | 0.181 | ||||
| LSTM-RNN | 0.072 | 0.091 | 0.108 | 0.195 | LSTM-RNN | 0.047 | 0.056 | 0.084 | 0.171 | ||||
| 20 | DNN | 0.088 | 0.117 | 0.139 | 0.269 | 20 | DNN | 0.049 | 0.061 | 0.099 | 0.221 | ||
| DBN | 0.077 | 0.093 | 0.127 | 0.252 | DBN | 0.056 | 0.065 | 0.105 | 0.227 | ||||
| KELM | 0.085 | 0.102 | 0.141 | 0.273 | KELM | 0.065 | 0.077 | 0.113 | 0.228 | ||||
| GRU | 0.082 | 0.146 | 0.131 | 0.264 | GRU | 0.064 | 0.073 | 0.104 | 0.227 | ||||
| LSTM-RNN | 0.076 | 0.094 | 0.117 | 0.231 | LSTM-RNN | 0.049 | 0.058 | 0.096 | 0.214 | ||||
| 30 | DNN | 0.096 | 0.114 | 0.154 | 0.332 | 30 | DNN | 0.052 | 0.066 | 0.112 | 0.278 | ||
| DBN | 0.082 | 0.099 | 0.139 | 0.316 | DBN | 0.058 | 0.072 | 0.116 | 0.277 | ||||
| KELM | 0.089 | 0.109 | 0.153 | 0.328 | KELM | 0.068 | 0.083 | 0.125 | 0.294 | ||||
| GRU | 0.087 | 0.105 | 0.142 | 0.311 | GRU | 0.063 | 0.078 | 0.116 | 0.286 | ||||
| LSTM-RNN | 0.081 | 0.096 | 0.128 | 0.305 | LSTM-RNN | 0.051 | 0.063 | 0.108 | 0.261 | ||||
| Fall | 10 | DNN | 0.071 | 0.096 | 0.117 | 0.204 | Winter | 10 | DNN | 0.036 | 0.056 | 0.092 | 0.181 |
| DBN | 0.06 | 0.083 | 0.113 | 0.214 | DBN | 0.044 | 0.065 | 0.098 | 0.182 | ||||
| KELM | 0.107 | 0.142 | 0.176 | 0.254 | KELM | 0.056 | 0.076 | 0.092 | 0.179 | ||||
| GRU | 0.083 | 0.115 | 0.116 | 0.202 | GRU | 0.041 | 0.061 | 0.088 | 0.179 | ||||
| LSTM-RNN | 0.057 | 0.079 | 0.101 | 0.185 | LSTM-RNN | 0.026 | 0.044 | 0.071 | 0.161 | ||||
| 20 | DNN | 0.074 | 0.121 | 0.124 | 0.255 | 20 | DNN | 0.039 | 0.06 | 0.099 | 0.221 | ||
| DBN | 0.062 | 0.086 | 0.117 | 0.243 | DBN | 0.047 | 0.071 | 0.106 | 0.234 | ||||
| KELM | 0.117 | 0.144 | 0.175 | 0.319 | KELM | 0.059 | 0.079 | 0.101 | 0.215 | ||||
| GRU | 0.086 | 0.114 | 0.125 | 0.239 | GRU | 0.044 | 0.064 | 0.097 | 0.212 | ||||
| LSTM-RNN | 0.061 | 0.083 | 0.109 | 0.228 | LSTM-RNN | 0.03 | 0.048 | 0.081 | 0.205 | ||||
| 30 | DNN | 0.077 | 0.107 | 0.135 | 0.305 | 30 | DNN | 0.043 | 0.074 | 0.111 | 0.275 | ||
| DBN | 0.065 | 0.095 | 0.129 | 0.291 | DBN | 0.051 | 0.079 | 0.118 | 0.277 | ||||
| KELM | 0.113 | 0.155 | 0.187 | 0.387 | KELM | 0.061 | 0.084 | 0.113 | 0.285 | ||||
| GRU | 0.089 | 0.119 | 0.136 | 0.314 | GRU | 0.047 | 0.072 | 0.108 | 0.274 | ||||
| LSTM-RNN | 0.063 | 0.089 | 0.122 | 0.282 | LSTM-RNN | 0.034 | 0.055 | 0.092 | 0.242 | ||||
Mape of the five algorithms for 2020 after error correction.
Bold only shows the optimal solution results.
Error Analysis
The experiments reported in Section “Short-term Predictions” have been conducted using the transformed wind direction data from four seasons. An inverse bilinear transformation, expressed in Eq. 2, is applied to transform the predicted transformed wind direction into the original angular range [0°, 360°]. The actual angular values versus the forecasted angular values by the proposed framework using ICEEMDAN and LSTM-RNN are presented in Figure 8. It can be seen that the majority of the forecasted values fall within a relatively small range with respect to the actual values. It demonstrates the proposed framework can sufficiently provide accurate forecasting performances.
FIGURE 8
In this section, performance of the ICEEMDAN framework integrated with the long-short-term memory recurrent neural network (LSTM-RNN) for prediction of wind direction at four seasons is discussed. The prediction error of the inverse transformed wind direction at 2, 5, 10 min, and 1 h horizons are presented in Table 3. The mean absolute error (MAE) and mean absolute percentage error (MAPE) of wind direction are smaller in the summer and winter.
TABLE 3
| Season | Fall | Winter | Spring | Summer |
|---|---|---|---|---|
| MAE (2 min) | 1.88° | 1.67° | 1.92° | 1.65° |
| MAPE (2 min) | 0.86% | 0.79% | 0.95% | 0.74% |
| MAE (5 min) | 5.18° | 5.01° | 5.57° | 4.93° |
| MAPE (5 min) | 2.37% | 2.28% | 2.75% | 2.21% |
| MAE (10 min) | 6.53° | 6.25° | 6.97° | 6.10° |
| MAPE (10 min) | 3.18% | 3.01% | 3.36% | 2.96% |
| MAE (1 h) | 10.64° | 9.05° | 10.58° | 9.34° |
| MAPE (1 h) | 5.62% | 4.39% | 5.29% | 4.71% |
Angular prediction error at four seasons.
The wind direction error shows less variability over short horizons. The changes of a nacelle position are usually made within 5 min and the prediction error should be under 3% (). A control chart is applied to monitor the prediction error and facilitate changing the nacelle position. A control chart with lower and upper limits enables monitoring the yaw error. Any prediction error that exceeds the bound (i.e., 3%) may trigger a change of the nacelle position. The final forecasting errors in the angular perspectives are illustrated in Figure 9.
FIGURE 9
Validation
The long-short-term memory recurrent neural network (LSTM-RNN) has been demonstrated to perform better than other algorithms. To validate the effectiveness and robustness of the LSTM-RNN, the data collected from another wind farm located in Shandong Province in the year 2020 has been used. The experiments are conducted following the similar training strategies as described in Section “Data Analysis”. The computational results are presented in Table 4.
TABLE 4
| Season | Fall | Winter | Spring | Summer |
|---|---|---|---|---|
| MAE (2 min) | 1.61° | 1.63° | 1.62° | 1.59° |
| MAPE (2 min) | 0.79% | 0.78% | 0.80% | 0.76% |
| MAE (5 min) | 4.94° | 4.41° | 5.19° | 4.51° |
| MAPE (5 min) | 2.08% | 1.93% | 2.58% | 1.97% |
| MAE (10 min) | 5.92° | 5.86° | 6.18° | 5.94° |
| MAPE (10 min) | 2.79% | 2.78% | 2.81% | 2.71% |
| MAE (1 h) | 9.81° | 9.02° | 9.75° | 9.32° |
| MAPE (1 h) | 4.77% | 4.37% | 4.92% | 4.74% |
Angular prediction error at four seasons.
The prediction error (see Table 4) in winter and summer seasons of 2020 from the wind farm in Shandong Province produced by the LSTM-RNN is similar to the one based on the 2020 data (see Table 3) in the wind farm in Inner Mongolia. More accurate performance has been observed in the fall and spring seasons with two prevailing wind directions. The favorable prediction error validates the effectiveness and robustness of the LSTM-RNN to predict the nacelle orientation.
Conclusion
A hybrid short-term forecasting framework to orient nacelle based on the predicted wind direction was presented. Industrial data collected from a wind farm in Inner Mongolia, China was utilized to train and validate the prediction models. A bilinear transformation was applied to transform the wind direction data from an angular variable into a continuous time-series. The forecasting framework was developed using ICEEMDAN integrated with LSTM-RNN. Also, the error corrections are implemented to improve the forecasting accuracy. The wind direction was predicted at short-term horizons, i.e., 2, 5, 10 min, and 1 h. Five algorithms, the deep neural network, deep belief network, kernel-based extreme learning machine, gated recurrent unit network, and long-short-term memory recurrent neural network were applied to predict wind direction at short-term horizons. The results of performance analysis of the five algorithms at four seasons were reported.
It was demonstrated that the long-short-term memory recurrent neural network outperformed the other four algorithms tested to predict wind direction. The results presented are of paramount importance in yaw control and can improve the efficiency of energy extraction process.
Statements
Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
Author contributions
HL conceptualized the study, contributed to the study methodology, and wrote the original draft. JD contributed to the study methodology, data curation and investigation. PF contributed to data analysis and investigation. HL contributed to software and formal analysis. CP and DA contributed to investigation and writing-original draft. QC contributed to editing. All authors have read and agreed to the published version of the manuscript.
Funding
This research is supported by the “Miaozi project” of scientific and technological innovation in Sichuan Province, China (Grant No. 2021090), the National Key Research and Development Program of China (2018YFC1505105), the Opening fund of State Key Laboratory of Geohazard Prevention and Geoenvironment Protection (Chengdu University of Technology) (Grant No. SKLGP 2021K014), and the Project from Sichuan Mineral Resources Research Center (SCKCZY2021-YB009) the Project of remote sensing identification and monitoring of geological hazards in Sichuan province, CN (2020) (Grant No. 510201202076888).
Conflict of interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
Author disclaimer
Frontiers Media SA remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
1
AminC. M.SamiF. M.TrzynadlowskiA. M. (2018). Wind Speed and Wind Direction Forecasting Using echo State Network with Nonlinear Functions. Renew. Energ.131 (2), 879–889. 10.1016/j.renene.2018.07.060
2
BilgiliM.SahinB.YasarA. (2007). Application of Artificial Neural Networks for the Wind Speed Prediction of Target Station Using Reference Stations Data. Renew. Energ.32 (14), 2350–2360. 10.1016/j.renene.2006.12.001
3
CastinoF.FestaR.RattoC. F. (1998). Stochastic Modelling of Wind Velocities Time Series. J. Wind Eng. Ind. Aerodynamics74-76, 141–151. 10.1016/s0167-6105(98)00012-9
4
ChoK.Van MerriënboerB.GulcehreC.BahdanauD.BougaresF.SchwenkH.et al (2014). Learning Phrase Representations Using RNN Encoder-Decoder for Statistical Machine Translation. 10.3115/v1/D14-1179
5
ColominasM. A.SchlotthauerG.TorresM. E. (2014). Improved Complete Ensemble EMD: A Suitable Tool for Biomedical Signal Processing.. Biomed. Signal Proce. Contr.146, 19–29.
6
DaviesA. (1974). Bilinear Transformation of Polynomials. IEEE Trans. Circuits Syst.21 (6), 792–794. 10.1109/tcs.1974.1083929
7
DuanJ.ZuoH.BaiY.DuanJ.ChangM.ChenB. (2021). Short-term Wind Speed Forecasting Using Recurrent Neural Networks with Error Correction. Energy217, 119397. 10.1016/j.energy.2020.119397
8
ErdemE.ShiJ. (2011). ARMA Based Approaches for Forecasting the Tuple of Wind Speed and Direction. Appl. Energ.88 (4), 1405–1414. 10.1016/j.apenergy.2010.10.031
9
GersF. A.SchraudolphN. N.SchmidhuberJ. (2003). Learning Precise Timing with Lstm Recurrent Networks. J. Machine Learn. Res.3 (1), 115–143.
10
GroutageF.VolfsonL.SchneiderA. (2003). S-plane to Z-Plane Mapping Using a Simultaneous Equation Algorithm Based on the Bilinear Transformation. IEEE Trans. Automatic Control.32 (7), 635–637. 10.1109/TAC.1987.1104664
11
HochreiterS.SchmidhuberJ. (1997). Long Short-Term Memory. Neural Comput.9 (8), 1735–1780. 10.1162/neco.1997.9.8.1735
12
HuQ.ZhangR.ZhouY. (2016). Transfer Learning for Short-Term Wind Speed Prediction with Deep Neural Networks. Renew. Energ.85 (JAN), 83–95. 10.1016/j.renene.2015.06.034
13
JuryE. (1973). Remarks on "The Mechanics of Bilinear Transformation". IEEE Trans. Audio Electroacoust.21 (4), 380–382. 10.1109/tau.1973.1162485
14
KhosraviA.KouryR. N. N.MachadoL.PabonJ. J. G. (2018). Prediction of Wind Speed and Wind Direction Using Artificial Neural Network, Support Vector Regression and Adaptive Neuro-Fuzzy Inference System. Sustainable Energ. Tech. Assessments25, 146–160. 10.1016/j.seta.2018.01.001
15
KodamaK.BurlsN. J. (2019). An Empirical Adjusted Enso Ocean Energetics Framework Based on Observational Wind Power in the Tropical pacific. Clim. Dyn.53 (5), 3271–3288. 10.1007/s00382-019-04701-8
16
KouZ.YangF.WuJ.LiT. (2020). Application of ICEEMDAN Energy Entropy and AFSA-SVM for Fault Diagnosis of Hoist Sheave Bearing. Entropy22 (12), 1347. 10.3390/e22121347
17
LecunY.BengioY.HintonG. (2015). Deep Learning. Nature521 (7553), 436–444. 10.1038/nature14539
18
LeeT.-H.WhiteH.GrangerC. W. J. (1993). Testing for Neglected Nonlinearity in Time Series Models. J. Econom.56 (3), 269–290. 10.1016/0304-4076(93)90122-l
19
LiH.XuQ.HeY.DengJ. (2018). Prediction of Landslide Displacement with an Ensemble-Based Extreme Learning Machine and Copula Models. Landslides15 (10), 2047–2059. 10.1007/s10346-018-1020-2
20
LiH.XuQ.HeY.FanX.LiS. (2020). Modeling and Predicting Reservoir Landslide Displacement with Deep Belief Network and EWMA Control Charts: a Case Study in Three Gorges Reservoir. Landslides17 (3), 693–707. 10.1007/s10346-019-01312-6
21
LiuD.NiuD.WangH.FanL. (2014). Short-term Wind Speed Forecasting Using Wavelet Transform and Support Vector Machines Optimized by Genetic Algorithm. Renew. Energ.62, 592–597. 10.1016/j.renene.2013.08.011
22
LiuH.ShiJ.ErdemE. (2010). Prediction of Wind Speed Time Series Using Modified Taylor Kriging Method. Energy35 (12), 4870–4879. 10.1016/j.energy.2010.09.001
23
MasseranN.RazaliA. M.IbrahimK.LatifM. T. (2013). Fitting a mixture of von Mises distributions in order to model data on wind direction in Peninsular Malaysia. Energ. Convers. Manag.72, 94–102. 10.1016/j.enconman.2012.11.025
24
McwilliamsB.SprevakD. (1982). The Simulation of Hourly Wind Speed and Direction. Mathematics Comput. Simulation24 (1), 54–59. 10.1016/0378-4754(82)90050-7
25
MohandesM. A.HalawaniT. O.RehmanS.HussainA. A. (2004). Support Vector Machines for Wind Speed Prediction. Renew. Energ.29 (6), 939–947. 10.1016/j.renene.2003.11.009
26
OuyangT.HeY.HuangH. (2018). Monitoring Wind Turbines' Unhealthy Status: a Data-Driven Approach. IEEE Trans. Emerging Top. Comput. Intelligence3 (2), 163–172. 10.1109/TETCI.2018.2872036
27
OuyangT.HeY.LiH.SunZ.BaekS. (2019). Modeling and Forecasting Short-Term Power Load with Copula Model and Deep Belief Network. IEEE Trans. Emerg. Top. Comput. Intell.3 (2), 127–136. 10.1109/tetci.2018.2880511
28
OuyangT.KusiakA.HeY. (2017). Predictive Model of Yaw Error in a Wind Turbine. Energy123, 119–130. 10.1016/j.energy.2017.01.150
29
PanC.YiJ.YinC.YuJ.LiX. (2019). Joint 3D UAV Placement and Resource Allocation in Software-Defined Cellular Networks with Wireless Backhaul. IEEE Access7, 104279–104293. 10.1109/access.2019.2927521
30
PengG.SiC. A.JcB.DiC. (2020). Wind Direction Fluctuation Analysis for Wind Turbines. Renew. Energ.162, 1026–1035. 10.1016/j.renene.2020.07.137
31
RongH.GaoY.GuanL.ZhangQ.ZhangF.LiN. (2019). Gam-based Mooring Alignment for Sins Based on an Improved Ceemd Denoising Method. Sensors19 (16), 3564. 10.3390/s19163564
32
SakH.SeniorA.BeaufaysF. (2014). Long Short-Term Memory Based Recurrent Neural Network Architectures for Large Vocabulary Speech Recognition. arXiv preprint arXiv:1402.1128
33
SanthoshM.VenkaiahC.Vinod KumarD. M. (2018). Ensemble Empirical Mode Decomposition Based Adaptive Wavelet Neural Network Method for Wind Speed Prediction. Energ. Convers. Manag.168, 482–493. 10.1016/j.enconman.2018.04.099
34
SunZ.HeY.GritsenkoA.LendasseA.BaekS. (2020). Embedded Spectral Descriptors: Learning the point-wise Correspondence Metric via Siamese Neural Networks. J. Comput. Des. Eng.7 (1), 18–29. 10.1093/jcde/qwaa003
35
TagliaferriF.ViolaI. M.FlayR. G. J. (2015). Wind Direction Forecasting with Artificial Neural Networks and Support Vector Machines. Ocean Eng.97, 65–73. 10.1016/j.oceaneng.2014.12.026
36
TangZ.ZhangZ. (2019). The Multi-Objective Optimization of Combustion System Operations Based on Deep Data-Driven Models. Energy182, 37–47. 10.1016/j.energy.2019.06.051
37
TangZ.ZhaoG.OuyangT. (2021). Two-phase Deep Learning Model for Short-Term Wind Direction Forecasting. Renew. Energ.173, 1005–1016. 10.1016/j.renene.2021.04.041
38
WasynczukO.ManD. T.SullivanJ. P. (1981). Dynamic Behavior of a Class of Wind Turbine Generators during Random Wind Fluctuations. IEEE Power Eng. Rev.PER-1 (6), 47–48. 10.1109/mper.1981.5511593
39
XuW.YiJ.DasguptaS.CaiJ. F.JacobM.ChoM. (2018). “Sepration-free Super-resolution from Compressed Measurements Is Possible: an Orthonormal Atomic Norm Minimization Approach,” in 2018 IEEE International Symposium on Information Theory (ISIT) (IEEE), 76–80.
40
YangZ.WangJ. (2018). A Hybrid Forecasting Approach Applied in Wind Speed Forecasting Based on a Data Processing Strategy and an Optimized Artificial Intelligence Algorithm. Energy160, 87–100. 10.1016/j.energy.2018.07.005
41
YiJ.XuW. (2020). Necessary and Sufficient Null Space Condition for Nuclear Norm Minimization in Low-Rank Matrix Recovery. IEEE Trans. Inform. Theor.66 (10), 6597–6604. 10.1109/tit.2020.2990948
42
ZhangW.QuZ.ZhangK.MaoW.MaY.FanX. (2017). A Combined Model Based on CEEMDAN and Modified Flower Pollination Algorithm for Wind Speed Forecasting. Energ. Convers. Manag.136, 439–451. 10.1016/j.enconman.2017.01.022
43
ZhouJ.ShiJ.LiG. (2011). Fine Tuning Support Vector Machines for Short-Term Wind Speed Forecasting. Energ. Convers. Manag.52 (4), 1990–1998. 10.1016/j.enconman.2010.11.007
Summary
Keywords
wind direction, bilinear transformation, ICEEMDAN, LSTM-RNN, error correction
Citation
Li H, Deng J, Feng P, Pu C, Arachchige DDK and Cheng Q (2021) Short-Term Nacelle Orientation Forecasting Using Bilinear Transformation and ICEEMDAN Framework. Front. Energy Res. 9:780928. doi: 10.3389/fenrg.2021.780928
Received
22 September 2021
Accepted
12 October 2021
Published
26 October 2021
Volume
9 - 2021
Edited by
Tinghui Ouyang, National Institute of Advanced Industrial Science and Technology (AIST), Japan
Reviewed by
Jirong Yi, The University of Iowa, United States
Arian Dhini, University Indonesia, Indonesia
Updates
Copyright
© 2021 Li, Deng, Feng, Pu, Arachchige and Cheng.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Jiahao Deng, jdeng5@depaul.edu
This article was submitted to Wind Energy, a section of the journal Frontiers in Energy Research
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.