Solar and wind predictions using evolution differentiated modular and LSTM long-term modelling based on pattern correlations

Spatiotemporal correlations between meteo-inputs and wind-solar outputs in an optimal regional scale are inevitable in the evolution of robust models that are reliable in mid-term prediction time horizons. Modelling border conditions are vital for early recognition of progress in chaotic atmospheric processes at the central destination. This approach is used in differential and deep learning, the comparison of artificial intelligence (AI) techniques that allow reliable pattern representation of long-term uncertainty and regional irregularities. The proposed day-by-day estimation of RE production potential is based on data first processing in detection of modelling initialisation times from historical databases, considering pattern similarity distance. Optimal data sampling is crucial for AI training in statistically based predictive modelling. Differential learning (DfL) is an unconventional and newly developed biologically inspired strategy that fuses numerical derivative solutions with evolutionary neurocomputing. This hybrid approach is based on the optimal determination of partial differential equations (PDEs) composed at the nodes of gradually expanding binomial networks. This allows modelling highly uncertain weather-related physical systems using unstable RE. The main objective is to improve its self-evolution and the resulting computation in prediction time. Representing patterns in direct relation to their complexity using input-output similarity resampling reduces ambiguity in RE forecasting. Node-by-node feature selection and dynamical PDE representation of DfL are evaluated along with long-short-term memory (LSTM) recurrent processing of Deep Learning (DL), capturing complex spatio-temporal patterns. Parametric C++ executable software with one-month spatial metadata records is available to compare additional modelling strategies.