CN102496064B - Method for acquiring unevenness of track - Google Patents

Abstract

The invention relates to a method for acquiring unevenness of a track, which comprises the following steps that: vehicle vibration response data and orbit dynamic unevenness data detected by a track inspection vehicle are used for estimating the static unevenness of the track by adopting genetic algorithm to be combined with vehicle track coupling model and selecting the least square criterion instruction genetic algorithm, so sum of the error square between the theoretic output value of the calculation model and vehicle system displacement, speed, acceleration and track dynamic unevenness measured by the track inspection vehicle is minimized, and a best track static unevenness value can be obtained. Considering different vehicle track coupling models of different running trains, track vibration response which is possibly experienced by the train when the train passes through the track is obtained through simulation calculation, and then the track dynamic unevenness is calculated through the sum of the static unevenness and the steel track dynamic response, so scientific evidence can be provided for estimating the running safety and comfort of the train as well as track maintenance.

Description

A kind of acquisition methods of track irregularity

Technical field

The present invention relates to a kind of acquisition methods of track irregularity.

Background technology

Along with at a high speed, the developing rapidly of heavy haul railway, the dynamic action between rolling stock and track structure is strengthened day by day, the vehicle causing thus and track vibration, fatigue breakdown are more serious.Track irregularity is a kind of aggravation wheel-rail force, vibrative main outside excitation source to rolling stock system, and track dynamic irregularity detects the main regular detection that relies on track checking car and synthetic detection vehicle both at home and abroad at present.Can be divided into static track irregularity and dynamic track irregularity by detection mode.Static track irregularity refers to the track irregularity not having under rail wheeling action, and dynamic track irregularity refers to that the track irregularity under rail wheeling action detects.Dynamic track irregularity is the track condition that really affects train operating safety and stable operation.

Track dynamic irregularity detects the main regular detection that relies on track checking car and synthetic detection vehicle both at home and abroad at present.Because the various trains of track checking car, synthetic detection vehicle and actual operation there are differences in speed, axle weight, suspension parameter, the track irregularity data that therefore detect vehicle Dynamic Acquisition can not reflect track irregularity state when vehicle in use passes through.

Classical dynamics of vehicle is taking Vehicular system as research object, supposes that rail system is non-yielding prop, does not consider the dynamic effects of rail system to Vehicular system, and track irregularity is considered as to the outside excitation of Vehicular system.Along with the raising of train running speed, the aggravation of wheel track dynamic action, the research method that vehicle, two subsystems of track are independently got up can not solve the INTERACTION PROBLEMS of this complex large system.The a lot of scholars of recent domestic have carried out a large amount of theoretical researches about vehicle-orbit coupling dynamics field successively, have obtained many achievements.Zhai Wan penetrating judgment is awarded at the Space Coupling model of having set up passenger vehicle-track, lorry-track, locomotive-track, and wherein rail portion adopts rail-sleeper-railway roadbed three layer model; Feng Qingsong adopts analytical method to analyze under irregularity condition high speed railway track structural vibration and train speed, the impact of track irregularity on Ballast track structural vibration; Lei Xiaoyan teaches taking TGV vehicle as example, by setting up the vertical coupling model of stock rail, has analyzed the Vehicular vibration response of friction speed, different line conditions under the uneven excitation of track.Above-mentioned Most scholars is all to utilize stock rail coupling model to carry out the Response Analysis of stock rail, using track irregularity as known input stimulus.

Summary of the invention

The present invention proposes a kind of track irregularity acquisition methods, the estimation of track irregularity state while passing through to realize various trains.

Vehicular vibration response data and track dynamic irregularity data that first the present invention utilizes track checking car to detect, adopt genetic algorithm to combine with stock rail coupling model and estimate track static track irregularity.Then in conjunction with the stock rail coupling models of different operation trains, by simulation calculation obtains train by time the track dynamic irregularity that may experience, keep in repair scientific basis be provided for the security of assessment train operation and comfortableness and the maintenance of way.

Object of the present invention is achieved through the following technical solutions:

An acquisition methods for track irregularity, the method is as follows:

1) quality of initialization track checking car and track, rigidity, damping parameter matrix;

2) produce track static track irregularity initial population;

3), by static track irregularity initial population substitution stock rail coupling model, utilize numerical method to obtain the vibratory response of Vehicular system;

4) according to genetic algorithm objective function, i.e. error and minimum between the corresponding measured value of model output valve and model output valve, calculates the fitness of each individuality in population;

5) individuality in population is carried out to selection, intersection, the mutation operation of genetic algorithm, obtain population of new generation;

6) judge whether to arrive evolution end condition, obtain track static track irregularity optimal value;

7) utilize the track static track irregularity obtaining, in conjunction with stock rail coupling model, the vibration displacement while obtaining actual vehicle in use through track, obtains track dynamic irregularity.

Further, described step 1) in parameter comprise quality, damping and the stiffness matrix of track checking car, track.

Further, described step 3) in be to utilize criterion of least squares method to set up the genetic algorithm objective function of population.

Further, described step 6) in end condition be: judge whether the evolutionary generation in genetic algorithm equals 300.

The invention has the advantages that:

The present invention utilizes genetic algorithm to combine with stock rail coupling model and has proposed a kind of new obtain manner of track static track irregularity, the real track dynamic irregularity that may experience can obtain runing train operation accurately time, for maintenance of way maintenance and security of operation provide support.

Brief description of the drawings

The track static track irregularity of Fig. 1 based on genetic algorithm estimated flow process;

The track irregularity estimation scheme of Fig. 2 based on genetic algorithm and stock rail coupling model;

Fig. 3 track static track irregularity estimated result;

Fig. 4 track dynamic irregularity estimated result.

Embodiment

Parameter estimation is by the estimation function model to observation sample constructing variable, under a certain estimation criterion constraint condition, adopts the process of certain numerical value computing method estimation function Model Parameter.The function model that parameter estimation is constructed is generally divided into linear model and nonlinear model.Vehicle-orbit coupling model is a nonlinear system.Parameter Estimation of Nonlinear Model criterion and PARAMETERS IN THE LINEAR MODEL estimate similarly have: least-squares estimation criterion, maximum likelihood estimation criterion, Minimum Mean Square Error estimation criterion, Least absolute sum estimation criterion, Bayes' risk least estimated criterion etc.That be wherein most widely used is non-Linear least square estimation (LS).

Select non-linear least square estimation criterion herein, the track static track irregularity in the vertical coupling model of vehicle-track is estimated.Under the guidance of least-squares estimation criterion, the estimated result of static track irregularity can allow the error sum of squares minimum between output valve and the measured value of the vertical coupled system of vehicle-track.

The derivation algorithm of common non-linear least square parameter mainly contains: linear approximation method, process of iteration, search procedure, high-order local derviation direct method etc.The algorithm of leading for the nonlinear model the demand that can not lead all can not solve.Newton iteration is local convergence simultaneously, required parameter estimation local minimum often.Have a kind of algorithm of directly separating in model space random search can effectively overcome these problems, the method does not need differentiate, and generally has globally optimal solution.These methods mainly comprise Monte Carlo (Monte Carlo) method, simulated annealing (Simulated Annealing) and genetic algorithm (Genetic Algrithms).

The principle of the invention is Vehicular vibration response data and the track dynamic irregularity data of first utilizing track checking car to detect, adopt genetic algorithm to combine with stock rail coupling model, select criterion of least squares to instruct genetic algorithm to estimate track static track irregularity, make the theoretical output valve of computation model and the error sum of squares minimum of Vehicular system displacement, speed, acceleration and track dynamic irregularity that track checking car is measured, estimate track static track irregularity.Then in conjunction with the different stock rail coupling models of runing trains, by simulation calculation obtain train by time the track vibration response that may experience, and then calculate track dynamic irregularity by static track irregularity and rail dynamic response sum, for the assessment security of train operation and comfortableness and maintenance of way maintenance provide scientific basis.

Below in conjunction with accompanying drawing, the present invention is obtained to track irregularity and do specific description.

According to vehicle-orbit coupling kinetic model, system dynamics equation can be write as to following form:

$\left\{\begin{array}{c}\left[M\right]\left\{\stackrel{\·\·}{X}\right\}+\left[C\right]\left\{\stackrel{\·}{X}\right\}+\left[K\right]\left\{X\right\}=\left\{Q\right\}\\ \left\{Q\right\}=f\left(\right\{X\left\}m\right\{\stackrel{\·}{X}\},\{\stackrel{\·\·}{X}\},\{Z_0\left\}\right)\end{array}\right.---\left(1\right)$

{ Z in formula ₀it is track static track irregularity value.

Because whole system is large-scale Nonlinear System of Equations, we cannot directly parse track static track irregularity { Z ₀, but this problem can be converted into the Parameter Estimation Problem of model.In the case of mass matrix, damping matrix and the stiffness matrix of Vehicular system and rail system is definite, different track static track irregularities, can obtain the response of different vehicles-rail system by the numerical integration of second-order equation group.Therefore, vehicle-track coupling system simply can be described as:

X′＝L(M，C，K，Z ₀)+ΔE (2)

The measured value of the dynamic response of X ' expression vehicle-rail system in formula; L (M, C, K, Z ₀) expression vehicle-track coupling system model; Δ E is error.So, track static track irregularity { Z ₀estimation problem can be converted into the Parameter Estimation Problem based on vehicle-orbit coupling model.

If track static track irregularity Z ₀scope be-10mm ~ 10mm.The detected value xk_r of displacement vector x (1: 10), velocity v (1: 10), acceleration a (1: 10) and the track dynamic irregularity of known vehicle system.The software flow pattern of genetic algorithm estimation track static track irregularity as shown in Figure 1.1. initialization track checking car and track are known quality, damping and stiffness matrix; 2. initialization wheel under track static track irregularity population; 3. adopt the method for criterion of least squares to evaluate each individuality, make system responses and measurement error quadratic sum minimum; 4. selection, intersection, the variation three of carrying out genetic algorithm operate greatly; 5. judge whether to arrive end condition.

After the track static track irregularity obtaining according to genetic algorithm optimization, using track static track irregularity as input, obtain the rail dynamic displacement under vehicle in use effect in conjunction with vehicle in use orbit coupling model, with the track static track irregularity of estimating be added can obtain operation train by time track irregularity state.

Below in conjunction with Fig. 2 and embodiment, the present invention is further detailed explanation.

The vertical nonlinear model of passenger vehicle-track that vehicle-orbit coupling model adopts in an embodiment of the present invention, is made up of auto model and model trajectory.Vehicle simulation is become to one in auto model and run on the multi-rigid-body system on track structure with speed v, by car body, trailing or leading bogie and wheel to forming, considered car body vertical vibration, nod, the vertical vibration of bogie, nod, take turns right vertical vibration, totally 10 degree of freedom.Model trajectory adopts rail-sleeper-railway roadbed-roadbed three-decker, and wherein rail adopts the point-supported endless Eular of elasticity beam, and sleeper adopts rigid unit, and railway roadbed is separated into railway roadbed module unit in the vertical.Vehicular system is realized by vertical contact of wheel track with the vertical coupling of rail system, is embodied in the vertical acting force between wheel track, and what adopt at embodiment is the most classical effective Hertz nonlinear elasticity contact model.

In embodiment, adopt hot-short parameter and high-speed line parameter as shown in table 1-table 2.

Table 1 hot-short parameter

Table 2 high-speed line parameter

Describe the acquisition process of track irregularity in detail below in conjunction with embodiment.

(1) produce initial population

In the space of separating, the individuality that produces at random some forms initial population, and individual number is exactly the size of population.The size of initial population and diversity directly have influence on the search efficiency of genetic algorithm.Population size n should not be too little also should not be too large.The too little meeting of population size n declines the Optimal performance of genetic algorithm, adopts larger population size can reduce genetic algorithm and is absorbed in the chance of locally optimal solution, but this means that computation complexity increases.The Population Size of track static track irregularity is taken as 20 in the present embodiment, and dimension NVAR is taken as 4, corresponding four wheels under track irregularity.

(2) calculate fitness function

According to the criterion of least squares of parameter estimation, set up the objective function of genetic algorithm.Objective function is namely according to the variance of the error of the theoretical value of the vertical coupling model output of the corresponding vehicle-track of individuality and measured value.Target function value is less, and individual adaptive value is just larger.The computing formula of objective function is:

$\mathrm{fitfun}=\mathrm{\Δ}{E}^2$

$=\underset{i=1}{\overset{10}{\mathrm{\Σ}}}{w}_x^i{(x_i^{\′}-x_i)}^2+w_v^i{(v_i^{\′}-v_i)}^2+w_A^i{(A_i^{\′}-A_i)}^2+w{(\mathrm{xk}\_\mathrm{ei}-\mathrm{xk}\_r)}^2---\left(3\right)$

In formula, x ' _i, v ' _i, A ' _iwith xk_ei be that the vertical coupling model of i individual corresponding vehicle-track calculates Vehicular system displacement, speed, acceleration and the track dynamic irregularity of output, be to utilize Newmark prediction-correction method integration to obtain; x _i, v _i, A _iwith xk_r be known Vehicular system displacement, speed, acceleration and track dynamic irregularity; W is weight matrix.

(3) select operation

Adopt even sequencing selection operator, adopt optimum conversation strategy, preserve in population 10% optimum individual number, these optimum individual numbers do not participate in crossing operation and variation computing, but with it replace the minimum individuality of fitness producing in Ben Dai colony after the operations such as intersection, variation.

(4) interlace operation

Adopt discrete intersection recombination form, discrete recombination can produce a mask table and decide which father is individual to contribute for filial generation.

(5) mutation operation

The genetic algorithms use that track static track irregularity is estimated be the operation of non-uniform mutation, original variable is being increased to a disturbed value.Computing formula is:

variable＝variable+range×delta

Delta determines the precision of mutation operator energy optimum positioning value, and the minimum value of selecting delta is herein 2 ^-19.

The estimated result of the track static track irregularity obtaining according to genetic algorithm optimization as shown in Figure 3.Blue line is the track static track irregularity of estimating, red line is real track static track irregularity, can find out, estimated value and true value overlap substantially, and their related coefficient is 0.97, and standard error is 0.1752mm.

(6) track dynamic irregularity when calculating vehicle in use passes through

Using the track static track irregularity obtaining by genetic algorithm as initial conditions, in the stock rail coupling model of substitution vehicle in use, obtain the dynamic response of stock rail by numerical integration algorithm, can obtain the dynamic displacement of rail under vehicle in use effect, by with initial static track irregularity be added can obtain vehicle in use by time dynamic track irregularity.Fig. 4 is track dynamic irregularity estimated result.

The present invention can set up on the basis of stock rail coupling model and track checking car detection data, the estimation of track irregularity while realizing a kind of different train operation in conjunction with genetic algorithm, simulation result shows, the variation that adopts the track static track irregularity of this algorithm estimation and the dynamic response of vehicle-rail system can well catch up with true value, has realized the estimation of circuit time of day when dissimilar train is passed through.

Claims (3)

1. an acquisition methods for track irregularity, is characterized in that, the method is as follows:

1) quality of initialization track checking car and track, rigidity, damping parameter matrix;

2) produce track static track irregularity initial population;

3), by static track irregularity initial population substitution vehicle-orbit coupling model, utilize numerical method to obtain the vibratory response of Vehicular system;

Described vehicle-orbit coupling model state into:

X′＝L(M,C,K,Z ₀)+ΔE (2)

The measured value of the dynamic response of X ' expression vehicle-rail system in formula; L (M, C, K, Z ₀) expression vehicle-track coupling system model; M, C, K, Z ₀represent mass matrix, damping matrix, stiffness matrix and track static track irregularity value; Δ E is error;

4) according to genetic algorithm objective function, i.e. error and minimum between the corresponding measured value of model output valve and model output valve, calculates the fitness of each individuality in population;

5) individuality in population is carried out to selection, intersection, the mutation operation of genetic algorithm, obtain population of new generation;

6) judge whether to arrive evolution end condition, obtain track static track irregularity optimal value;

7) using the track static track irregularity obtaining by genetic algorithm as initial conditions, in vehicle-orbit coupling model of substitution vehicle in use, obtain the dynamic response of stock rail by numerical integration algorithm, can obtain the dynamic displacement of rail under vehicle in use effect, by with initial static track irregularity be added can obtain vehicle in use by time dynamic track irregularity;

The step that described genetic algorithm obtains track static track irregularity is:

1. initialization track checking car and track are known quality, damping and stiffness matrix;

2. initialization wheel under track static track irregularity population;

3. adopt the method for criterion of least squares to evaluate each individuality, make system responses and measurement error quadratic sum minimum;

4. selection, intersection, the variation three of carrying out genetic algorithm operate greatly;

5. judge whether to arrive end condition.

2. the acquisition methods of a kind of track irregularity according to claim 1, is characterized in that, described step 4) in be to utilize criterion of least squares method to set up the genetic algorithm objective function of population.

3. the acquisition methods of a kind of track irregularity according to claim 1, is characterized in that, described step 6) in end condition be: judge whether the evolutionary generation in genetic algorithm equals 300.