Metaheuristic-supported smart integration of solar-wind hybrid systems in mesh distribution networks using grey wolf with SVM and LSTM forecasting

As the world moves more quickly toward sustainable energy, it is becoming more important to include renewable sources like wind and solar in power distribution networks. This work presents a novel framework that integrates advanced forecasting with intelligent optimisation to improve the efficiency of mesh distribution systems, which offer enhanced reliability over conventional radial networks. A standard support vector machine learning model and a long short-term memory deep learning model are used to make accurate predictions about renewable resources. Comparative results show that LSTM is more accurate than SVM, which makes fewer mistakes in predicting the future and better reflects changes in solar irradiance and wind speed over time. These better predictions are used with a Grey Wolf Optimisation algorithm to find the best placement and size for wind turbines and bifacial photovoltaic panels. The suggested method makes the power system much more efficient, cutting active power loss by 41.83% and reactive power loss by 36.47% in the 33-bus mesh distribution network. It shows a 32.73% drop in active power loss and a 38.06% drop in reactive power loss in the 118-bus system. This shows that it can greatly improve the efficiency of the whole system.