| With the continuous development of technology and science,the popularity of high-speed railways has become a necessary trend for social development.While our country is at a critical stage of sustainable development,research on high-speed railways is an indispensable part of the comprehensive transportation system.In the planning process of constructing new traction substations,there is a higher demand for research on the influence of traction characteristics on the power quality.Therefore,the method for studying the characteristics of the traction load probability distribution of high-speed railways also needs constant exploration and improvement.First of all,this article uses a large number of high-speed railway traction substation measured data,using the stochastic process theory and statistical analysis method based on ? distribution,the traction transformer wiring form as a single-phase wiring form,statistical calculation of high-speed traction load negative sequence current 95% The measured values of large values of probabilities were used to derive the incomplete ? function.Then the calculated values of 95% probability of the negative sequence current of the traction load were obtained by looking up the table,and the measurement and calculation were unified by comparison and error calculation.This paper examines the distribution of two-arm currents during the regenerative conditions by statistical analysis of the duration of the traction process and the study of when the negative sequence voltage unbalance takes the maximum value and the 95% probability maximum,and observes the regenerative braking.The effect of negative sequence voltage imbalance.Through research,people found that high-speed railway,the duration of regenerative braking accounts for a short duration of the whole day,and it has little effect on the unbalanced degree of the negative sequence of the feeder voltage.Due to the large number of traction load samples and the different wiring conditions,the study found that the moments of the traction load samples can be combined with the traction of their digital characteristic parameters,and the fuzzy C-means is used a for clustering.Data mining finds that the sample data gathered together have similar trends in the probability distribution characteristics.And the use of cluster evaluation indicators to determine the impact of different moments and different clustering methods on the results of cluster analysis,found that the use of traction load samples of each moment combination of traction and its digital characteristics of the cluster analysis of clustering results are relative the best.When studying the capacity design of traction transformers,the common typical load curves given by the design standards often do not consider the different characteristics of traction transformers.Therefore,in this paper,for traction loads that have already converged into one class,the probability distribution characteristics of samples in clusters are similar,considering designing a typical traction transformer typical load curve for a cluster.In past studies,the typical load curve is often periodic,and by studying the feeder current waveforms of the clustered samples,the unloaded load section of the traction load feeder waveform gathered together is relatively concentrated and does not exhibit periodicity throughout the day.Considering designing a typical load curve of a segmented traction transformer,the idle period corresponds to the no-load period of the feeder current,and the step density is obtained by using the probability density.Using the latest railway design standards as a reference,the winding hot spot temperature is calculated,and the insulation aging rate and life loss are calculated by the hot spot temperature.Compared with the typical load curves given by the design standards,the load factors of the typical load curves based on different clusters are selected to be more reasonable and the life loss is lower,which proves the feasibility of the method. |
PDF Full Text Request