State Estimation Using DCGNN-CLO Approach for IoT-Enabled Wind Energy Conversion System

Wind farms, which comprise multiple wind turbines, play an important role in renewable energy generation. However, network connectivity issues and data transmission delays in the Internet of Things (IoT) infrastructure can affect monitoring and decision-making, reducing the accuracy and responsiveness of state estimation. This paper proposes a hybrid methodology for state estimation in an IoT-enabled wind energy conversion system (WECS), with the primary objective of enhancing system efficiency and reliability. The proposed method combines Density Clustering and Graph Neural Network (DCGNN) and Clouded Leopard Optimization (CLO) and is named the DCGNN-CLO approach. The DCGNN model effectively predicts the operational state of the wind turbine by applying advanced graph-based learning. To improve the prediction accuracy and model robustness, the weight parameters of DCGNN are fine-tuned using the CLO method. The performance of the proposed method is evaluated using MATLAB and compared with several existing techniques such as Space Vector Pulse Width Modulation (SVPWM) Algorithm, Fractional Order Darwinian Particle Swarm Optimization (FODPSO) and Convolutional Neural Network (CNN). From the findings, the proposed method achieved a low THD of 3.26%with a shorter response time of 0.23[Formula: see text]s, minimizing harmonic distortions while ensuring rapid system adaptation. Additionally, with a low RMSE of 0.112 and a correlation coefficient of 0.997, the proposed method outperforms existing approaches by achieving higher prediction accuracy and stronger correlation with actual turbine states.