Study On Parameter Identification Algorithms Of PV Array And Equivalent Model Of PV Power Station

As the traditional fossil energy drying up,photovoltaic(PV)power generation becomes hot today.But at the present stage,the PV power generation industry in the world is still in the primary stage of development,and there are still some problems to be solved in the development.After the PV system is connected to the grid on a large scale,the analysis and study on the running characteristics of PV system has become one of the core related topics.And the establishment of an accurate and simple model of PV power station is the most important part of the theoretical research.Single stage PV power generation system mainly consists of PV array,voltage stabilizing capacitor,DC-AC inverter,filter,transformer and control system.Where,the control system includes the control of the inverter and the maximum power point tracking control.On the basis of the model,the U_L-I_L output curves under the change of temperature and irradiance are analyzed.PV array is an important part of the PV system.Its mechanism model has a clear concept and high precision and the simplified engineering model has less parameters,simple expression and strong practicability.These two models are more widely used in practical engineering and simulation research.But in the actual applicat ion process,PV array can't be accurately represented by the parameters formed of PV cell data multiplied by series and parallel numbers provided by the manufacturers.And the parameters in PV array are very sensitive to changes in weather and climate.Therefore,the parameters of the model should be identified in multi scene.Based on the measured data of a PV power station,the hybrid artificial fish swarm and frog leaping algorithm is adopted to identify the parameters in the mechanism model and the simplified engineering model of PV array.The identification results of hybrid algorithm and the results of individual identification of artificial fish swarm algorithm(AFSA)and shuffled frog leaping algorithm(SFLA)are compared and analyzed which proves that the hybrid algorithm has the superiority and effectiveness of the two algorithms.On the basis of effective algorithm,multi scene parameter identification is carried out in four seasons,spring,summer,autumn and winter,and four weather types,sunny,cloudy,cloudy and rainy.The output curves of PV array model can be well fitted to the measured curves of the actual PV power station for any day,and further verifies the accuracy and effectiveness of the algorithm.The equivalent model of the whole PV power station is established on the basis of PV array model which can accurately express the output power of PV array in the actual power station.Selecte irradiance S,temperature T and output active power PL as the grouping index and use the subtractive fuzzy clustering algorithm and fuzzy C mean clustering algorithm to cluster PV power unit of the whole power station.According to the established unit classification,the equivalent values of the related parameters of the PV power generation unit are calculated.The results show that the output of the equivalent model is basically consistent with the output of the actual photovoltaic power station.Then the accuracy and effectiveness of the equivalent method of the power station are verified by evaluating the clustering results through appropriate evaluation indexes.