Optimization of Wind Power Configuration in Distribution Network Based on Scenario Clustering and Power Flow Entropy

When the wind power is connected to the distribution network, the access points and the access capacity of each point in the distribution network will also affect the stability of the distribution network. Therefore, based on the fuzzy C clustering of wind speed, this paper optimizes the location and capacity of wind power in the distribution network by using the proposed entropy index of power flow which reflects the power flow equilibrium degree of the system. Firstly, typical wind speed and output scenarios are selected through scenario clustering, and then the flow entropy index reflecting system stability is adopted to optimize the location and capacity of wind power. The improved IEEE33 was used to verify the method. Introduction With the development of energy technology, it is imperative to connect renewable energy such as wind power to distribution network. However, due to the inherent uncertainty of wind power, it is bound to affect the stability of distribution network. Therefore, how to plan the access location of wind power in distribution network and the access capacity of wind power in each access location is very important [1,2]. Since the randomness of wind power output is strong, the randomness of wind power generation needs to be analyzed before planning and optimization. A probability distribution can be used to describe the randomness of variables comprehensively. Since the randomness of different meteorological variables is not the same, they must correspond to different theoretical distribution models. The results show that the wind speed probability distribution curve can be fitted by a statistical model. Common fitting models include Rayleigh model, log-normal distribution model, Gamma distribution model and two-parameter Weibull distribution model [3]. However, in the process of calculating the probability distribution function, parameter estimation is difficult. At the same time, the accuracy of the probability function is difficult to guarantee because of too many uncertain factors. Scenario analysis can exactly solve the above problems. The scenario set describes the probability of the possible occurrence of uncertain events in the future. Since the probability measure of random events can be observed, the scenario set can comprehensively reflect the occurrence of full probability scenarios [4,5]. Too many scenarios will result in too much computation. Therefore, the scene needs to be clustered. Fuzzy C Mean (FCM) is a classification based Fuzzy clustering method. By describing the membership of samples to different categories, this algorithm can objectively cluster, which can overcome the defects of traditional clustering algorithm of either/or, and occupies an important position in Fuzzy clustering [6]. Since the access of wind power will inevitably affect the power flow distribution of the system, the power flow distribution of the system also affects the stability of the system. Therefore, this paper proposes to use the power flow entropy index to quantify the power flow distribution of the system, and to analyze the optimal node and capacity of system access to wind power [7,8]. To sum up, in the process of selecting the location of the wind power distribution network and determining the node capacity, this paper firstly adopts the method of fuzzy C clustering to analyze the uncertain wind speed scenario. After obtaining the typical scenario, based on the theory of entropy, the flow entropy of the system which can quantitatively describe the imbalance of power flow of system is taken as the optimization index to optimize the access location of wind power and the access capacity of each location. Clustering of Scenarios Scenario Clustering based on Fuzzy C In practical engineering application, it is not difficult to find the development law of random events through many data statistical analysis. The continuous probability function is discretized into several samples by means of sampling discretization. Each sample is called a "scenario". The content of each scenario consists of two parts: the value of the random variable w and the probability p of the value. This method is called the scenario method for solving optimization problems involving random variables, so as to obtain the probability of random events in each possible situation. The greater the probability of the scenario, the greater the impact on the target event, thus forming the scenario analysis method. In order to obtain enough scene sets, many samples are needed, which leads to an exponential increase in the size of the final scene set, resulting in "dimension disaster", which is not conducive to analysis and calculation. Scene reduction is to use a more concise set of scenes to represent the original set of scenes, so scene reduction is also known as scene simplification. Therefore, this paper adopts fuzzy c-means algorithm to cluster renewable energy scenarios. The objective function of the fuzzy c-means is: