A Self-Data-Driven Method for Lifetime Prediction of PV Arrays Considering the Uncertainty and Volatility

This article proposes a self-data-driven method for remaining useful life prediction of PV arrays based on self-condition monitoring data considering the uncertainty and volatility. First, a health indicator reconstruction method is presented to eliminate the uncertainty and volatility of condition monitoring data. Second, a nonlinear Gamma stochastic process model is established to describe the probability distribution of the degradation trend. Then, the model parameter solution is transformed into an optimization problem, and a hybrid particle swarm and gray wolf optimization algorithm is developed to estimate the model parameters avoiding trapping in local optimization and divergence. Finally, two case studies are demonstrated to verify the effectiveness of the proposed method based on the Desert Knowledge Australia Solar Center and NREL datasets, and the performance is further evaluated in comparisons with the empirical models, statistical models, and long short-term memory network. Experimental results demonstrate that the proposed method has excellent RUL prediction accuracy.