Abstract
Abstract An enhanced convolutional neural network (CNN) has been proposed to address the issue of low accuracy in wind farm power prediction using traditional methods, which leverages an improved version of the Gray Wolf Optimizer (IGWO). The standard Gray Wolf Optimizer (GWO) exhibits a linear change in the search range, which can limit its dynamic convergence capabilities. By incorporating nonlinear convergence factors and dynamic convergence strategies, the improved algorithm modulates the global search pace, effectively preventing it from getting stuck in local optima. The refined IGWO is subsequently applied to fine-tune the optimal CNN architecture, mitigating the uncertainty typically associated with the structural configuration of CNNs. When predicting the actual power measurements from a wind plant in Inner Mongolia, the results indicate that the IGWO-based CNN method outperforms both the conventional GWO-CNN approach and the standalone CNN regarding prediction accuracy.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have