Improving Performance of Recurrent Neural Networks Using Simulated Annealing for Vertical Wind Speed Estimation

An accurate vertical wind speed(WS)data estimation is required to determine the potential for wind farm installation.In general,the vertical extrapolation of WS at different heights must consider different parameters fromdifferent locations,such as wind shear coefficient,roughness length,and atmospheric conditions.The novelty presented in this article is the introduction of two steps optimization for the Recurrent Neural Networks(RNN)model to estimate WS at different heights using measurements from lower heights.The first optimization of the RNN is performed to minimize a differentiable cost function,namely,mean squared error(MSE),using the Broyden-Fletcher-Goldfarb-Shanno algorithm.Secondly,the RNN is optimized to reduce a non-differentiable cost function using simulated annealing(RNN-SA),namely mean absolute error(MAE).Estimation ofWS vertically at 50 m height is done by training RNN-SA with the actualWS data a 10–40 m heights.The estimatedWS at height of 50 m and the measured WS at 10–40 heights are further used to train RNN-SA to obtain WS at 60 m height.This procedure is repeated continuously until theWS is estimated at a height of 180 m.The RNN-SA performance is compared with the standard RNN,Multilayer Perceptron(MLP),Support Vector Machine(SVM),and state of the art methods like convolutional neural networks(CNN)and long short-term memory(LSTM)networks to extrapolate theWS vertically.The estimated values are also compared with realWS dataset acquired using LiDAR and tested using four error metrics namely,mean squared error(MSE),mean absolute percentage error(MAPE),mean bias error(MBE),and coefficient of determination(R2).The numerical experimental results show that the MSE values between the estimated and actualWS at 180mheight for the RNN-SA,RNN,MLP,and SVM methods are found to be 2.09,2.12,2.37,and 2.63,respectively.

Land-sea breeze circulation structure on the west coast of the Yellow Sea,China

Land-sea breeze(LSB)is an atmospheric mesoscale circulation that occurs in the vicinity of the coast and is caused by uneven heating resulting from the difference in specific heat capacity between the sea and land surfaces.The circulation structure of LSB was quantitatively investigated with a Doppler wind lidar Windcube100s on the west coast of the Yellow Sea for the first time.The time of observation was 31 August to 28 September 2018.It was found that the height of LSB development was 700 m to 1300 m.The duration of conversion of LSB was between 6 h and 8 h.The biggest average horizontal sea-breeze wind speed at 425 m was 5.6 m s^(-1),and at 375 m it was 4.5 m s^(-1).During the conversion process from sea breeze to land breeze,the maximum wind shear exponent was 2.84 at 1300 m altitude.During the conversion process from land breeze to sea breeze,the maximum wind shear exponent was 1.28 at 700 m altitude.The differences in wind shear exponents between sea-breeze and landbreeze systems were between 0.2 and 3.6 at the same altitude.The maximum value of the wind shear exponent can reflect the height of LSB development.