A designed predictive modelling strategy based on data decomposition and machine learning to forecast solar radiation

• A hybrid solar radiation forecasting model has been designed. • Robust Local Mean Decomposition (RLMD) splits the solar radiation data into the subset of PFs, AMPs and FMs signals. • The low frequency based PFs play a vital role as pertinent features. • Random Forest algorithm (RF) used the PFs to design RLMD-RF to forecast solar radiation. • The proposed RLMD-RF model provides significant energy management implications. Consistent with the United Nations Sustainable Development Goal 7, the design and optimization of clean energy resources ( e.g ., solar energy) is a highly motivating task for all researchers globally who continue to build research synergies that can tackle the future likelihood of energy crises due to socioeconomic and strategic environmental policy. In this paper, a weekly solar radiation ( S R ) forecasting model is designed using a robust local mean decomposition (RLMD) technique unified with a random forest (RF) algorithm to generate a fully optimised hybridized RLD-RF model that has a promising capability to forecast the S R values. In the first stage of model design, the RLMD, a frequency resolution method, is applied to decompose the original S R time series into amplitude modulation subseries (AMs), frequency modulation subseries (FMs), and the low-frequency product functions (PFs) to reveal the internal structure of the model construction data to incrementally optimize the RLD-RF model where only PFs were used. Subsequently, the statistically significant lagged subseries at a week ahead forecasting horizon ( t – 1) of the low-frequency PFs with residual components are extracted individually, via partial autocorrelation function ( PACF ), to capture the historical behaviour of frequency-resolved S R component in order to build a robust modelling framework. Consequently, the random forest (RF) algorithm is employed to forecast each of the subseries using PACF-based lagged inputs to construct a fully optimised hybrid RLMD-RF predictive model. RLMD-RF is benchmarked against a baseline RF, M5tree, and multiple linear regression (MLR), Artificial neural network (ANN) and Gaussian process regression (GPR) algorithms, including their hybridized counterparts ( i.e. , RLMD-M5tree, RLMD-MLR, RLMD-ANN, and RLMD-GPR) using statistical score metrics in the independent testing phase. The results generated at test sites in Queensland State, Australia that have high solar energy potential confirm that the RLMD-RF method can produce quality predictions of weekly solar radiation against the benchmarking comparison models. For instance, RLMD-RF for Barcaldine are higher in terms ( E WI = 0.938 , E NS = 0.878) against RLMD-MLR ( E WI = 0.845 , E NS = 0.705), the RLMD-M5tree ( E WI = 0.836 , E NS = 0.684), RLMD-ANN ( E WI = 0.836 , E NS = 0.715), RLMD-GPR ( E WI = 0.839 , E NS = 0.716), the RF ( E WI = 0.720 , E NS = 0.564), the M5tree ( E WI = 0.692 , E NS = 0.522), the MLR ( E WI = 0.683 , E NS = 0.508), the ANN ( E WI = 0.708 , E NS = 0.519) and the GPR ( E WI = 0.708 , E NS = 0.520). Similarly, the RLMD-RF also outperformed in Rockhampton, Clermont, and Lockyer Valley stations as compared to other models. This research establishes the practical usefulness of hybridised RLMD-RF modelling framework for accurate S R forecasting and advocates its possible consideration in renewable and sustainable energy production and monitoring systems that can aid in decision-making by energy utilities and stakeholders ( e.g ., climate and energy policy experts).

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.