Wind Farm Wind Energy Resource Assessment Based-on The Optimization Algorithm

It's an important measure to develop wind power in the case of many issues arisen, such as shortage of resources, environmental pollution. The wind resource evaluation is the basic work to explore wind power. The accurate and effective assessment can influence the practicality and economy of the wind power development directly. Wind speed probability distribution is seen as the quite important indicator to wind resources. The two-parameter Weibull model is seen as the best model to fit the actual wind speed. Since the traditional methods have shortcomings, this paper tries to introduce the particle swarm algorithm to solve such conventional problems.The particle swarm optimization (PSO) is a new intelligent optimization algorithm, whose prime idea comes from information sharing and evolution between individuals from the group. It is a probability-based heuristic search algorithm, with lots of advantages such as few parameters, simple structure and so on, having been widely used in many fields.The main researching results are as following:(1)This paper makes a detailed study of the relevant content of the traditional PSO, including the analysis of the algorithm mechanism, the contrast of two elementary particle swarm optimization, the impact of various parameters on the algorithm, the processes and characteristics of the algorithm, the comparison with the other optimization algorithm, to get a comprehensive understanding of the PSO.(2)A new improved algorithm is proposed, which the Simulated Annealing is introduced to prevent the traditional PSO falling into local optimum, since it can accept the inferior solutions, at the same time, the average speed is introduced to adjust both the inertia weight and learning factors to raise the convergence speed during the late stage of the algorithm. Through the simulation experiment of the test functions, the superiority of the improved algorithm can be shown by increasing the improvement points one by one.(3)The two-parameter Weibull model is widely recognized as the best model to fit the actual wind speed, the method of determining its parameters are:the least square method, the average wind speed and standard deviation method, the average wind speed and maximum speed method. However, the post-processing work of the first two is quite tedious, the latter one needs the data in a long time, which is difficult to get at the present. So the improved algorithm is introduced to model and optimize the Weibull distribution with the practical data. The fitting result is satisfactory by comparison with that of the other methods.