基于群智能算法的四川省太阳辐射模型模拟研究OA

Solar radiation(Rs)serves as the primary energy source for the Earth system,accurate estimation of Rs is crucial for enhancing the precision of regional ET0 simulations and optimizing water resource allocation.To enhance the accuracy and universality of Rs forecasting in Sichuan Province,this study divides the province into three regions(Western Sichuan Plateau,Eastern Sichuan Basin,and Southwestern Sichuan Mountainous Area).Daily meteorological data from seven representative stations spanning 1994-2016 were selected.Least squares method(LSM),Whale Optimization Algorithm(WOA),Quantum Algorithm(QA),and Quantum-Whale Optimization Algorithm(QWOA)were employed to optimize parameters for nine empirical solar radiation(Rs)models,systematically evaluating their simulation accuracy.Results indicate that sunshine duration models(N1～N3)and mixed models(M1～M3)exhibit higher simulation accuracy across different regions of Sichuan Province,while temperature models(T1～T3)demonstrate poorer accuracy.Among these,the El-Sebaii,Angstrom and Ogelman models achieved the highest simulation accuracy in the Western Sichuan Plateau,Eastern Sichuan Basin,and Southwestern Sichuan Mountainous regions,with R² values of 0.844,0.896 and 0.873,respectively.Compared with the least squares method(LSM),optimization by swarm intelligence algorithms significantly improved the simulation accuracy of solar radiation models,with the Quantum-Whale Optimization Algorithm(QWOA)demonstrating the most pronounced enhancement effect.The optimized Rs empirical models across different regions yielded R² values ranging from 0.653 to 0.935,RMSE values from 1.997 to 3.963 MJ/(m2·d),and MAE values from 1.503 to 3.021 MJ/(m2·d).Overall,it is recommended to use the QWOA-optimized sunshine duration model to estimate daily Rs values for Sichuan Province.