2 February 2021 in Journal of Mathematics
This paper presents the application of a multiple number of statistical methods and machine learning techniques to model the relationship between rice yield and climate variables of a major region in Sri Lanka, which contributes significantly to the countrys paddy harvest. Rainfall, temperature (minimum and maximum), evaporation, average wind speed (morning and evening), and sunshine hours are the climatic factors considered for modeling. Rice harvest and yield data over the last three decades and monthly climatic data were used to develop the prediction model by applying artificial neural networks (ANNs), support vector machine regression (SVMR), multiple linear regression (MLR), Gaussian process regression (GPR), power regression (PR), and robust regression (RR). The performance of each model was assessed in terms of the mean squared error (MSE), correlation coefficient (R), mean absolute percentage error (MAPE), root mean squared error ratio (RSR), BIAS value, and the Nash number, and it was found that the GPR-based model is the most accurate among them. Climate data collected until early 2019 (Maha season of year 2018) were used to develop the model, and an independent validation was performed by applying data of the Yala season of year 2019. The developed model can be used to forecast the future rice yield with very high accuracy.
31 July 2021 in Mathematical Problems in Engineering
This paper presents the development of models for the prediction of power generation at the Samanalawewa hydropower plant, which is one of the major power stations in Sri Lanka. Four regression-based machine learning and statistical techniques were applied to develop the prediction models. Rainfall data at six locations in the catchment area of the Samanalawewa reservoir from 1993 to 2019 were used as the main input variables. The minimum and maximum temperature and evaporation at the reservoir site were also incorporated. The collinearities between the variables were investigated in terms of Pearsons and Spearmans correlation coefficients. It was found that rainfall at one location is less impactful on power generation, while that at other locations are highly correlated with each other. Prediction models based on monthly and quarterly data were developed, and their performance was evaluated in terms of the correlation coefficient (R), mean absolute percentage error (MAPE), ratio of the root mean square error (RMSE) to the standard deviation of measured data (RSR), BIAS, and the Nash number. Of the Gaussian process regression (GPR), support vector regression (SVR), multiple linear regression (MLR), and power regression (PR), the machine learning techniques (GPR and SVR) produced the comparably accurate prediction models. Being the most accurate prediction model, the GPR produced the best correlation coefficient closer to 1 with a very less error. This model could be used in predicting the hydropower generation at the Samanalawewa power station using the rainfall forecast.
1 December 2022 in Gazi University Journal of Science
This paper presents the development of wind energy prediction models for the Nala Danavi wind farm in Sri Lanka by using machine learning and statistical techniques. Wind speed and ambient temperature were used as the input variables in modeling while the daily wind energy production was the output variable. Correlation between the wind energy and each weather index was investigated using the Pearson’s and Spearman’s correlation coefficients and it was found that daily wind energy output is positively correlated with both daily averaged input variables. Statistical prediction models of Multiple Linear Regression (MLR) and Power Regression (PR) and the machine learning techniques of Support Vector Regression (SVR), Gaussian Process Regression (GPR), Feed Forward Backpropagation Neural Network (FFBPNN), Cascade-Forward Backpropagation Neural Network (CFBPNN) and Recurrent Neural Network (RNN) were developed. The accuracy of the prediction models was measured in terms of the coefficient of determination, Bias, Percent Root mean square error (RMSE)Bias, and Nash-Sutcliffe Efficiency (NSE). Results of the performance evaluation indicated that all the models are highly accurate while the FFBPNN-based model demonstrates outstanding performance with very low error. Such prediction models are highly important for a country like Sri Lanka whose power generation mainly depends on imported coal followed by hydropower and expanding the on-shore and off-shore wind farms gradually in many potential locations scattered over the country.