Probabilistic Calculation of Tidal Currents for Wind Powered Systems Using PSO Improved LHS

Abstract

This paper introduces the Particle Swarm Optimization (PSO) algorithm to enhance the Latin Hypercube Sampling (LHS) process. The key objective is to mitigate the issues of lengthy computation times and low computational accuracy typically encountered when applying Monte Carlo Simulation (MCS) to LHS for probabilistic trend calculations. The PSO method optimizes sample distribution, enhances global search capabilities, and significantly boosts computational efficiency. To validate its effectiveness, the proposed method was applied to IEEE34 and IEEE-118 node systems containing wind power. The performance was then compared with Latin Hypercubic Important Sampling (LHIS), which integrates significant sampling with the Monte Carlo method. The comparison results indicate that the PSO-enhanced method significantly improves the uniformity and representativeness of the sampling. This enhancement leads to a reduction in data errors and an improvement in both computational accuracy and convergence speed.