Comparative Analysis of Machine Learning Models for Day-Ahead Photovoltaic Power Production Forecasting

Abstract

A main challenge for integrating the intermittent photovoltaic (PV) power generation remains the accuracy of day-ahead forecasts and the establishment of robust performing methods. The purpose of this work is to address these technological challenges by evaluating the day-ahead PV production forecasting performance of different machine learning models under different supervised learning regimes and minimal input features. Specifically, the day-ahead forecasting capability of Bayesian neural network (BNN), support vector regression (SVR), and regression tree (RT) models was investigated by employing the same dataset for training and performance verification, thus enabling a valid comparison. The training regime analysis demonstrated that the performance of the investigated models was strongly dependent on the timeframe of the train set, training data sequence, and application of irradiance condition filters. Furthermore, accurate results were obtained utilizing only the measured power output and other calculated parameters for training. Consequently, useful information is provided for establishing a robust day-ahead forecasting methodology that utilizes calculated input parameters and an optimal supervised learning approach. Finally, the obtained results demonstrated that the optimally constructed BNN outperformed all other machine learning models achieving forecasting accuracies lower than 5%.

A hybrid methodology for distribution level photovoltaic power production forecasting verified at the distribution system of Cyprus

“…Finally, optimal prediction performance for each reference PV system predictive model was achieved by including the input features of the on‐site forecasted , and solar angles of and α, against the output variable of the maximum power [56].…”

“…Previous work demonstrated that ML models constructed with input feature combinations from computed NWP models and solar position algorithms (forecasted irradiance, temperature, and the Sun's position angles) outperformed any other input parameter selection [56]. This is attributed to the correction of underlying biases of NWP forecasted data.…”

“…The P (Da, 𝛽, M) is a normalisation factor that ensures that the total summation of the probability is unity and represents the conditional probability between the dataset and constructed model. Previous work demonstrated that ML models constructed with input feature combinations from computed NWP models and solar position algorithms (forecasted irradiance, temperature, and the Sun's position angles) outperformed any other input parameter selection [56]. This is attributed to the correction of underlying biases of NWP forecasted data.…”

“…Amongst the multitude of data-driven prediction models a Bayesian regularised neural network (BRNN) was employed in this study due to its proven convergence and performance accuracy in forecasting PV production compared to other statistical and ML models [56]. In addition, the BRNN technique was selected amongst a list of candidate ML models, due to the following reasons:…”

A Review of Machine Learning-Based Photovoltaic Output Power Forecasting: Nordic Context

“…Similar work that implements conventional ML algorithms for PV output power forecasting has also been pursued by others. SVM [62], BNN (Binarized Neural Network), SVR, and RT [63], and ANN [64], [65] are few examples. An important limitation of implementing the above methods arises when they are used for a PV plant located in a region where large and sudden weather variations over a short period are frequent.…”