| Nowadays, power quality problems have become increasingly prominent and effective power quality evaluation is a pressing matter of the moment. The present situation of power quality evaluation is described in this paper. On the basis of the existing power quality evaluation methods, evaluation methods based on cluster analysis are researched.Firstly, power quality evaluation methods of probability and vector algebra, combination weighing method and matter-element method are introduced by analyzing some examples and comparing their advantages and disadvantages, which lays foundation for the improvement of power quality evaluation methods and provides references for selecting the appropriate evaluation methods in practical application.Secondly, a power quality evaluation method based on equivalence relation clustering is proposed. The correlation coefficient is obtained by calculating fuzzy equivalence matrix, which is used to evaluate the power quality of clustering objects. Cluster analysis based on equivalence relation is simple, easy to calculate and combines the knowledge of fuzzy mathematics, which is used to solve the problem that power quality is hard to quantitative analyze.Thirdly, comprehensive applying of genetic algorithm, Mahalanobis distance and fuzzy C means clustering, a power quality evaluation method based on GA-MD-FCM algorithm is proposed. Genetic algorithm is used to overcome the disadvantage that the original FCM algorithm is easy to fall into the local optimal solution, and differences between different attributes can be considered by using Mahalanobis distance instead of Euclidean distance. Examples show that GA-MD-FCM algorithm can effectively evaluate the power quality and has excellent clustering results and fast convergence speed, and the value function is significantly reduced compared with the original FCM algorithm. The method does not need manual intervention, which overcomes the shortcomings of the traditional methods which are influenced by the subjective factors.Finally, a power quality evaluation method based on improved grey clustering and combination weighing method is proposed. Considering the shortcomings of traditional grey clustering, the index whiten weight function is introduced, which can solve the problem of zero value of whiten weight function effectively. By using the combination weighing method instead of the single weighing method, the accuracy of the weights can improve effectively.Aiming at six power quality indexes such as voltage deviation, voltage fluctuation, voltage flicker, harmonic voltage distortion rate, three-phase voltage unbalance degree and frequency deviation, the power quality is evaluated by using the methods mentioned above, and the characteristics and application range of each algorithm are compared and analyzed. The simulation results show the feasibility and effectiveness of the proposed methods. |
