CN-115345020-B - Traction load modeling method and system based on Laplace

CN115345020BCN 115345020 BCN115345020 BCN 115345020BCN-115345020-B

Abstract

The invention discloses a traction load modeling method and system based on Laplace, which are used for acquiring traction load data, constructing a traction load frequency histogram based on the traction load data, analyzing zero traction load in the traction load frequency histogram to obtain zero load probability, removing the zero traction load in the traction load frequency histogram, modeling the non-zero traction load by using Laplace mixed distribution to obtain a non-zero load model, integrating the zero load probability with the non-zero load model to obtain a probability model, and analyzing and solving the probability model by using a Laplace mixed distribution parameter iterative solution method to obtain an optimal probability model.

Inventors

WANG YUNLING
LIU ZHIGANG
LIU JIAWEI
LI MIN
Su Yunche
GOU JING
LIU YING
CHEN WEI
Chao Huawei
ZHANG QIAO

Assignees

国网四川省电力公司经济技术研究院

Dates

Publication Date: 20260505
Application Date: 20220826

Claims (4)

1. The traction load modeling method based on the Laplace is characterized by comprising the following steps: Acquiring traction load data, wherein the traction load data is traction load power data and traction load probability density which are obtained in one period; constructing a traction load frequency histogram based on the traction load data; Analyzing the zero traction load in the traction load frequency histogram to obtain zero load probability; Removing zero traction load in the traction load frequency histogram, and modeling non-zero traction load by using Laplacian mixed distribution to obtain a non-zero load model; integrating the zero load probability and the non-zero load model to obtain a probability model; analyzing and solving the probability model by using a Laplace mixed distribution parameter iteration solving method to obtain an optimal probability model; The zero load probability is the proportion of zero traction load to all traction loads; the specific obtaining method of the probability model comprises the following steps: integrating the zero load probability and the non-zero load model by a binomial distribution method; the specific expression of the probability model is as follows: ; f (x) represents traction load probability density, P is zero load probability, pi k is a mixing coefficient, L is a laplace operator, L (x|μ k , λ k ) is called a kth component in the mixing model, K is a mixing component number, μ is a position parameter, and λ is a shape parameter; The specific expression of the Laplace operator L is as follows: ; probability of p (x) being x; the specific expression of the mixing coefficient pi k is as follows: ; Analyzing and solving the probability model by adopting a Laplace mixed distribution parameter iteration solving method, wherein the specific substeps comprise the following steps: setting the number K of mixed components, setting an initial value of pi k , μ k , λ k for each component K, and calculating the value of a log likelihood function; b, calculating posterior probability gamma (i, k) according to the current pi k , μ k , λ k ; Based on the parameter gamma (i, k), recalculating a new parameter value pi k , μ k , λ k ; And D, recalculating the value of the log likelihood function based on the new parameter value pi k , μ k , λ k until the obtained parameter pi k , μ k ,λ k converges, and stopping iteration.
2. The traction load modeling method based on laplace as claimed in claim 1, wherein the specific expression of the log likelihood function is: ; Pi is a mixing coefficient vector, μ is a position parameter vector, λ is a shape parameter vector, and N is the traction load data amount.
3. The traction load modeling method based on laplace as claimed in claim 2, wherein the posterior probability concrete expression is: ; ; ; ; Gamma (i, k) denotes the posterior probability, N k denotes the posterior probability of the kth subcomponent, A new location parameter is indicated and a new location parameter is indicated, Representing the new shape parameter(s), Representing the new scale parameter, N represents the number of samples.
4. A traction load modeling system based on laplace, which is used for realizing the traction load modeling method based on laplace as claimed in any one of claims 1 to 3, and comprises a data acquisition module, a histogram construction module, an analysis module, a modeling module, an integration module and a solving module; the data acquisition module is used for acquiring traction load data, wherein the traction load data is traction load power data and traction load probability density which are acquired in one period; The histogram construction module is used for constructing a traction load frequency histogram based on the traction load data; the analysis module is used for analyzing the zero traction load in the traction load frequency histogram to obtain zero load probability; The modeling module is used for removing zero traction load in the traction load frequency histogram, and modeling the non-zero traction load by using Laplace mixed distribution to obtain a non-zero load model; the integration module is used for integrating the zero load probability and the non-zero load model to obtain a probability model; and the solving module is used for analyzing and solving the probability model by adopting a Laplace mixed distribution parameter iteration solving method to obtain an optimal probability model.

Description

Traction load modeling method and system based on Laplace Technical Field The invention relates to the technical field of traction load modeling, in particular to a traction load modeling method and system based on Laplacian. Background Because the electrified railway in China adopts a single-phase power supply mode, electric split phases are arranged in and among traction stations powered by non-single-phase transformers. When the motor train unit passes through the electric split phase, a roof breaker is required to be opened, and the motor train unit slides through the electric split phase, so that the traction load has obvious impact characteristics. In addition, the motor train unit operates according to the train operation diagram, in order to ensure the safe distance, the tracking time is generally more than 5min, and the tracking time of the spring operation busy period is also more than 3 min. This inevitably causes an intermittence in the traction load. How to accurately characterize traction load intermittence and jerkiness is a key issue faced by traction load modeling. In the prior art, a dynamic modeling method based on traction calculation is generally characterized in that the method is based on a traction calculation basic theory and combines actual parameters of a traction power supply system to establish a dynamic load mathematical model of the whole vehicle network system, and has high simulation precision, but large calculation amount and long simulation time. Therefore, when modeling the traction load by adopting the prior art, the required data is usually large, the calculated amount is excessive, the simulation time is too long, and the modeling speed is low and the efficiency is low. In view of this, the present application has been made. Disclosure of Invention The invention aims to solve the technical problems of low modeling speed and low efficiency caused by large calculated amount and long simulation time when modeling traction load in the prior art, and provides a traction load modeling method and system based on Laplacian, which can reduce calculated amount and improve modeling speed and modeling efficiency. The invention is realized by the following technical scheme: the traction load modeling method based on the Laplace comprises the following steps: Acquiring traction load data, wherein the traction load data is traction load power data and traction load probability density which are obtained in one period; constructing a traction load frequency histogram based on the traction load data; Analyzing the zero traction load in the traction load frequency histogram to obtain zero load probability; Removing zero traction load in the traction load frequency histogram, and modeling non-zero traction load by using Laplacian mixed distribution to obtain a non-zero load model; integrating the zero load probability and the non-zero load model to obtain a probability model; and analyzing and solving the probability model by using a Laplace mixed distribution parameter iteration solving method to obtain an optimal probability model. When the traction load is modeled, the method is generally adopted to build a dynamic load mathematical model of the whole vehicle network system based on a traction calculation basic theory and combined with actual parameters of a traction power supply system, but when the method is adopted to model the traction load, the required data is usually large, the calculated amount is excessive, the simulation time is too long, and the modeling speed is slow and the efficiency is low; the invention provides a traction load modeling method based on Laplace, which adopts the Laplace modeling method to separate zero load from non-zero load for model integration, and introduces the intermittence and the impact of the traction load into the modeling method, thereby reducing the calculation time required for modeling, improving the modeling speed and increasing the modeling efficiency. Preferably, the zero load probability is the specific magnitude of zero traction load over all traction loads. Preferably, the specific obtaining method of the probability model is as follows: and integrating the zero load probability with the non-zero load model through a binomial distribution method. Preferably, the specific expression of the probability model is: f (x) represents traction load probability density, P is zero load probability, pi k is a mixture coefficient, L is laplace operator, L (x|μ k,λk) is referred to as kth component in the mixture model, K is a mixture component number, μ is a position parameter, and λ is a shape parameter. Preferably, the laplacian operator L is expressed as follows: p (x) is the probability of x. Preferably, the specific expression of the mixing coefficient pi k is: Preferably, the probability model is analyzed and solved by adopting a Laplace mixed distribution parameter iteration solving method, and the specific substeps comprise the followin