Electrical Load Forecasting by using CANFIS Controller with Gaussian Function

The significant significance of load forecasting models is necessary for electrical utilities to manage power intelligently. Power generation is backed by a multitude of resources due to the digitalization of the sectors. As a result, different models have differing forecasting accuracy. When projecting future electrical consumption, power utilities with multiple generation modalities have difficulties and a sizable degree of error. A utility must use forecasting technologies to estimate future electricity demand with the least amount of inaccuracy in order to control the power flow with minimal power outages. In paper this we highlighted and more focus on Momentum learning rule, Bell membership function and Gaussian membership functions are used to enable the Co-Adaptive Neuro-Fuzzy Inference System (CANFIS) architecture’s adaptation to a variety of data patterns and complexities. The membership functions hold significant importance in facilitating the model’s ability to precisely represent the innate uncertainties and nonlinearities found in real-world datasets. This research thoroughly examines the effects of various CANFIS framework configurations on important performance parameters, including learning efficiency, prediction accuracy, and generalisation capacity.