CN106777752B - A kind of bullet train tracking operation curve optimal setting method - Google Patents
A kind of bullet train tracking operation curve optimal setting method Download PDFInfo
- Publication number
- CN106777752B CN106777752B CN201611252207.3A CN201611252207A CN106777752B CN 106777752 B CN106777752 B CN 106777752B CN 201611252207 A CN201611252207 A CN 201611252207A CN 106777752 B CN106777752 B CN 106777752B
- Authority
- CN
- China
- Prior art keywords
- bullet train
- formula
- speed
- train
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/15—Vehicle, aircraft or watercraft design
Abstract
The invention discloses a kind of bullet trains to track operation curve optimal setting method, the characteristics of for bullet train tracking interval under movable block " movement, distance to go ", route and bullet train operation data of the method based on collection in worksite, establish bullet train echo state network speed prediction model, the tracking moving model based on movable block, line characteristics model, and the tracking operation curve multiple target setting model using innovative evaluation index.Efficient multi-objective particle swarm algorithm is used again, using algorithmic statement condition as one of constraint of setting model, bullet train is carried out based on the above real time data and tracks operation curve optimal setting.Finally using section efficiency of operation and stability as the evaluation index of setting method, one group of optimal operation curve is filtered out, so that bullet train operational process is safe and efficient, while improving high-speed railway section efficiency of operation and stability under movable block.
Description
Technical field
The optimal setting method of " speed-mileage " curve that the present invention relates to bullet trains under movable block belongs to high speed
Train tracks running optimizatin control and automatic Pilot technical field.
Background technique
With the rapid development of society, transportation demand amount is continuously increased, and high-speed railway transportation more highlights it and passes through in its people
It helps developing critical positions.According to national " ' 13 ' programme of railway ", " 13 " are still what railway construction developed
Golden period, China will accelerate the development of High Speed Railway for Passenger Trans port, put forth effort to improve national speed rail net.In high speed, high density operation
High speed rail system in, the tracking running optimizatin control of bullet train based on movable block is to ensure that bullet train safety, high
One of the core technology of operation is imitated, and the related innovation of country, the comprehensive of manufacturing capacity embody.Current China railways train control system
Fixed blocking, quasi-moving block mode are generallyd use, movable block is used widely in urban rail system.Compared to other two
Kind block system, the not default block section of movable block, dynamic calculate the safe distance between adjacent train, have stronger operation
Adjustment capability and higher tracking efficiency, are the development trends of high-speed rail train control system.With the increasing of high-speed railway rate of traffic flow,
The mutual influence of train also becomes larger, so that bullet train service condition is more complicated, changeable.Therefore, it explores based on movement
The optimal tracking operation curve setting method of efficient, reliable bullet train of occlusion, to the bullet train realized under movable block
Safe and efficient operation has important research and application value.
Existing train tracks operation curve setting method, in terms of being concentrated mainly on subway (urban rail), for high-speed railway
Tracking operation curve setting research be in the starting stage.It is high compared to running environment closing, stable subway (urban rail) system
Fast railway operation condition is more complicated, changeable, proposes higher technology to bullet train tracking operation curve setting method and wants
It asks.Due to the strong nonlinearity dynamic characteristic of bullet train operational process, it is difficult to accurate prediction of speed mechanism model is established, and
Accurate bullet train speed prediction model is the basis of operation curve setting.Line characteristics are determining for train operation curvilinear motion
Determine one of factor, accurate line characteristics model is the basis of bullet train tracking operation curve setting research.And it existing sets
Determine method and often has ignored influence of some important line features to setting result, such as phase-separating section power-off.In addition, train is tracked
Operation curve sets the comprehensive evaluation index of result, determines the validity and feasibility of the optimal tracking operation curve of gained.Cause
This, establishes the key of efficient, practical comprehensive evaluation index and bullet train tracking operation curve setting.
Summary of the invention
The object of the present invention is to which the tracking running optimizatin for being bullet train under movable block provides a practicable
Simulation study platform establishes bullet train speed prediction model, tracing model, line characteristics model and operation based on field data
Curve multiple target setting model obtains optimal bullet train tracking operation curve using improved multi-objective optimization algorithm, real
Existing safe and efficient operation of the bullet train under movable block, while improving section efficiency of operation and stability.
A kind of bullet train tracking operation curve optimal setting method, comprising the following steps:
Step 1: the bullet train speed prediction model based on echo state network is established;Establish bullet train tracking mould
Type;Establish line characteristics model;
Step 2: the high speed based on echo state network established with the actual operating data of acquisition, training step one
Train speed prediction model parameters;
Step 3: based on trained bullet train speed prediction model and random generation N group control sequence collection { cli},
Obtain corresponding operation curve collection { vsi};
Step 4: acquiring speed limit, speed and position data in real time, and solving bullet train tracing model can obtain: D, Lm、Ln;
Step 5: judge D-LnWhether < ε is true;Practical tracking interval of the D between forward and backward vehicle, LnFor forward and backward workshop
Every distance threshold, ε is a sufficiently small positive real number;If establishment skips to step 7, step 6 is otherwise skipped to;
Step 6: the speed limit of rear car is adjusted are as follows: V (l)=v1(l), V (l) is the ATP speed limit of bullet train operation;
Step 7: being based on line characteristics model, establishes bullet train tracking operation curve multiple target setting model, is transported
Row curve Pareto disaggregation;
Step 8: it according to the efficiency of operation index of setting method assessment models, is solved from Pareto and screening is concentrated to obtain one group
Optimal solution vsx;
Step 9: vs is assessed using operation stability indicatorx;
Step 10: judge vsxWhether meet stability requirement, if it is satisfied, then terminating, otherwise skips to step 3 and continue to hold
Row.
The bullet train tracks operation curve optimal setting method, in step 1, the bullet train prediction of speed
Model is described as follows:
Y (t+1)=v (t+1), U (t+1)=[v (t);v(t-1);v(t-3);cl(t+1)] (3)
In formula, U, X, Y be an externally input respectively, intrinsic nerve member state, network output matrix/vector, Win、W、WoutPoint
Weight, intrinsic nerve member connection weight, output weight matrix/vector Wei not inputted;V (t+1) is bullet train current time
The speed of service, cl (t+1) ∈ { -1,0,1 } are the train handle level at current time, and " -1,0,1 " respectively indicates bullet train system
Dynamic, coasting, traction working condition.
The bullet train tracks operation curve optimal setting method, described to establish bullet train tracking in step 1
The method of model are as follows:
Provide minimum safety interval distance LmWith the spacing distance threshold for judging whether rear car operating status is influenced by front truck
Value Ln, it is as follows:
Ln=Lm+L1, D=d1-d2 (5)
In formula, laAnd lbIt is two vehicle minimums parking spacing and train length of wagon respectively;c1And c2It is front truck and rear car respectively
Deceleration, L1And L2It is the emergency stopping distance of front truck and the service braking distance of rear car;v1And v2It is the speed of front truck and rear car
Degree;ew1、ev1And ew2、ev2It is the measurement error of the positioning of front truck and rear car, the speed of service respectively;Reality of the D between forward and backward vehicle
Border tracking interval, d1And d2It is the position of forward and backward vehicle;D≤LnWhen, the operating status of rear car is influenced by front truck operating status.
The bullet train tracks operation curve optimal setting method, described to establish line characteristics model in step 1
Are as follows:
wa=wr(pr,lr)+wc(rc,lc)+wt(v,lt) (6)
In formula, (pr,lr),(rc,lc) and (v, lt) respectively indicate (gradient, ramp length), (curvature, length of curve) and
(train speed, length of tunnel);Under the collective effect of air drag and self gravity, fortune of the bullet train in split-phase section
Row velocity characteristic can be described as:
In formula, gsin (pr) it is gravity acceleration g in the component being parallel on the direction of ramp.
The bullet train tracks operation curve optimal setting method, in the step 7, establishes bullet train tracking
The method of operation curve multiple target setting model are as follows:
Provide safe and comfortable Measure Indexes and model constraint are as follows:
A1 safety
Using D as constraint of velocity, then safety evaluation index may be expressed as:
In formula, D is bullet train actual interval distance, and v and V (l) are the bullet train speed of service and ATP speed limit respectively.
In bullet train tracking operation course, when v will be more than V (l), vehicle-mounted ATP system can be tight to train automatic implementation
Anxious braking, so V (l)-v > 0 is set up always;ε is a sufficiently small positive real number;fsIt is smaller, then train tracking operation course
The risk that safety accident occurs is smaller;
A2 energy conservation
Entire tracing process is divided into limited sufficiently small section, traction energy consumption of the bullet train in each section can
It indicates are as follows:
Ei=F (vi)ΔSiM (11)
In formula, Δ Sj=vj·Δt,vj∈ v, Δ t are the sampling period, and M is train weight, F (vi) it is bullet train traction
Power is speed viFunction;
Total traction energy consumption in entire section are as follows:
A3 is comfortable
It is as follows by acquiring Comfort Evaluation index using weighting factor method to acceleration change amount and rate of acceleration change
It is shown:
fc=ω1|a|+ω2|r| (14)
In formula, a is acceleration, and a > 0 indicates to accelerate, and a < 0 indicates to slow down, and r is rate of acceleration change, ω1And ω2It is to add
Weight coefficient;Due to the powerful inertia of bullet train itself, influence of the r to comfort level is smaller with respect to a, and weighting coefficient is set as
ω1=0.8 and ω2=0.2;
The constraint of A4 Algorithm Convergence
For the multi-objective particle of use as a kind of efficient Stochastic Optimization Algorithms, convergence conditions are its bases
One of this constraint;
The algorithm search mechanism may be expressed as:
In formula, xjIt (n) is the position of j-th of particle, pbj(n) and gb (n) respectively indicates the particle optimum position and population
Optimum position,WithIt is acceleration factor, r1 and r2 are equally distributed random numbers in [0,1], and ω is changeable weight, nmax
It is maximum number of iterations;
Enable x*For particle optimal location, then pbj(n) it can transform to gb (n):
In formula, dpj(n) and dg(n) pb is respectively indicatedj(n) and gb (n) arrives x*Euclidean distance;
Since the position of Algorithm Convergence and particle is closely related, and it is almost unrelated with the search speed of particle, by formula
(16) formula (15) are substituted into and eliminates vj(n+1) it can obtain:
Ask expectation that can obtain formula (17):
In formula, r1 and r2 are to be uniformly distributed, then have E (r1)=E (r2)=1/2, X=E (x*) it is Pareto disaggregation center
The expectation of position,It is all constant with X;
Based on formula (18), optimum results global convergence be may be expressed as:
Based on formula (18) and (19), the adequate condition of algorithmic statement can be obtained are as follows:
The convergent adequate condition of this setting method can be obtained from above are as follows:
It establishes high speed using the convergence conditions in A4 as model constraint condition based on evaluation index in the above A1-A3 and arranges
Vehicle is tracked shown in operation curve setting model such as formula (22)-(23):
min f(vsi)={ fs(vsi),fe(vsi),fc(vsi),fcg(vsi)} (22)
In formula, f (vsi) it is composite evaluation function, vsiIt is bullet train tracking operation " speed-mileage " curve, ξ is given
Allowance on schedule;In formula (23) on schedule: being 1. to constrain, be 2. speed limit and security constraint, be 3. algorithmic statement constraint.It solves
The rate curve setting model obtains the Pareto disaggregation of optimized operation curve.
The bullet train tracks operation curve optimal setting method, described to assess according to setting method in step 8
The appraisal procedure of model are as follows:
The screening index of one group of optimal solution is obtained using efficiency of operation and stability as from Pareto solution concentration;
(1) efficiency of operation:
By period TsThe interior bullet train theory quantity N by certain block section between stationstIt is defined as efficiency of operation, as follows:
Nt=Ts/ti (13)
In formula, Nt∈R+, tiRuning time between station needed for the bullet train for passing through the section for the i-th column.We set herein
It is fixed, according to the bullet train for the optimized operation curve motion that the present invention is set:The calculating formula of k
ForΔSj=vj·Δt,vj∈v,L0For distance between sites, Δ t is the sampling period, and ρ is given parking
Precision;By formula it is found that under the premise of meeting high-speed railway multiple target service requirement, tiSmaller then efficiency of operation is higher;
(2) stability is runed:
By subsequent Train Schedule tiTime t late to front truckdRecovery capability be defined as operation stability, following institute
Show:
In formula, T0The service time between station given in operation timetable, λ is the coefficient of stability.λ > 0 indicates to stablize, and
In a certain range, the λ the big, and it is better to run stability;λ < 0 then indicates unstable.
The beneficial effect of the present invention compared with the prior art is:
(1) due to the strong nonlinearity dynamic characteristic in bullet train operational process, it is difficult to establish high speed train dynamics
Mechanism model is predicted to the accurate speed of service.Therefore, the present invention, which is established, establishes bullet train speed based on echo state network
It spends prediction model and replaces original prediction of speed mechanism model, the technical foundation as operation curve setting;With improved tracking
Moving model, line characteristics model, operation curve multiple target setting model are technological core, in which: 1. consider GSM-R channel radio
Letter module can realize that front and back vehicle location information exchanges, and for simplicity, cancel the front and back vehicle radio communication in original tracing model
Route;2. the phase-separating section feature in line characteristics model is indicated with speed variables;3. to improve efficiency of algorithm, by tracking interval
Rule of judgment of the distance as the switching of rear car speed limit, while algorithmic statement condition and on schedule index are changed to operation curve multiple target
The constraint of setting model;4. the efficiency of operation of block section between stations and stability assessment refer under the movable block defined according to the present invention
Mark is concentrated from Pareto solution and filters out one group of optimal solution.
(2) this invention simplifies the computation complexities of Multi-target evaluation index, and algorithmic statement condition is more as solving
One of object module constraint, cancels " sensitivity " and " energy conservation " preference, is changed to more preferably embody the section fortune of tracking operational effect
Efficiency and stability indicator are sought, the efficiency of optimized operation curve setting method and the practicability of setting result are improved;
(3) setting of optimized operation curve is more efficient, and the final result that sets can meet bullet train multiple target well
Service requirement, while improving high-speed railway efficiency of operation and stability under movable block.
(4) present invention is suitable for bullet train optimal control and automatic Pilot.
Detailed description of the invention
Fig. 1 is the bullet train tracking moving model based on movable block;
Fig. 2 is bullet train line characteristics model;
Fig. 3 is optimal tracking operation curve set algorithm flow chart;
Fig. 4 is the bullet train prediction of speed effect based on echo state network;
Fig. 5 is the practical tracking operation curve of bullet train;
Fig. 6 is the tracking operation curve not being set;
Tracking operation curve after Fig. 7 setting;
Fig. 8 runs stability test effect;
The optimal tracking operation curve setting strategy of Fig. 9;
Specific embodiment
Below in conjunction with specific embodiment, the present invention is described in detail.
Using advantage of the echo state network in terms of time series forecasting, bullet train speed prediction model is established;Needle
To the tracking interval between bullet train under movable block with the variation of the operating statuses such as front and back train traction, braking, and it is in
The characteristics of revealing " movement, distance to go ", bullet train tracing model is established, accurate description movable block tracks bullet train
The influence of operation curve;Fully consider the characteristics of high-speed railway route feature and split-phase are powered to bullet train tracking operation " speed
Accurate line characteristics model is established in the limitation of degree-mileage " curve;It is proposed new multiple target Measure Indexes: safe, energy saving, easypro
It is suitable, and using Algorithm Convergence condition as one of model constraint, solving model obtains the Pareto disaggregation of setting curve;It will move
The evaluation criterion of the lower block section between stations efficiency of operation and stability curve setting method of occlusion screens and obtains one group and optimal chase after
Track operation curve so that bullet train operational process meets safety, energy conservation and comfortable commuter service requirement, while improving movement
High-speed railway efficiency of operation and reliability under block system.
The technical solution specifically describes are as follows:
1, the bullet train speed prediction model based on echo state network
The prediction of the bullet train speed of service is a kind of typical time series forecasting problem, and complicated nonlinear kinetics
Feature makes the forecasting problem be difficult to solve by establishing mechanism model.Echo state network as recurrent neural network one
Kind innovation form has higher training effectiveness while retaining the memory function of Recursive Networks, has been successfully applied to many
The prediction of industrial data, such as the interval prediction of telephony access amount prediction and blast-furnace gas amount.Therefore, the present invention is arranged for high speed
The difficult point of vehicle speed forecasting problem is established in conjunction with advantage of the echo state network in terms of time series forecasting and is based on echo shape
The bullet train speed prediction model of state network.Model is described as follows:
Y (t+1)=v (t+1), U (t+1)=[v (t);v(t-1);v(t-3);cl(t+1)] (3)
In formula, U, X, Y be an externally input respectively, intrinsic nerve member state, network output matrix/vector, Win、W、WoutPoint
Weight, intrinsic nerve member connection weight, output weight matrix/vector Wei not inputted.V (t+1) is bullet train current time
The speed of service, cl (t+1) ∈ { -1,0,1 } are the train handle level at current time, and " -1,0,1 " respectively indicates bullet train system
Dynamic, coasting, traction working condition.
2, the bullet train tracing model based on movable block:
Bullet train tracking interval distance based on movable block, with the operating status of front and back train, braking ability
Change and dynamic change.In bullet train tracing process, minimum safety interval distance Lm, minimum interval time of departure TmAnd judgement
The spacing distance threshold value L whether rear car operating status is influenced by front truckn, it is three key variables of tracing computation.Due to opening
The interference such as the magnetic field in running environment, communication for putting, bullet train tests the speed, position error is inevitable, the foundation of tracing model
Need influence of these errors of reasonable consideration to train tracking operational safety.Fig. 1 describes the bullet train based on movable block
Tracing model.
Minimum safety interval distance may be expressed as:
In formula, laAnd lbIt is two vehicle minimum spacings and train length of wagon respectively;c1And c2The respectively system of front truck and rear car
Dynamic deceleration, L1And L2The respectively service braking distance of the emergency stopping distance of front truck and rear car;v1And v2It is front truck and rear car
Speed;ew1、ev1And ew2、ev2It is positioning, the data noise of front truck and rear car respectively.
Judgment threshold can be obtained are as follows:
Ln=Lm+L1 (5)
Practical tracking interval between forward and backward vehicle are as follows:
D=d1-d2 (6)
In formula, d1And d2It is the position of forward and backward vehicle.In tracking operation course of the rear car to front truck, as D≤LnWhen, then after
The operating status of vehicle is influenced by front truck operating status.
Minimum tracking time of departure TmIt may be expressed as:
In formula, TxAnd TyIt is the reaction time and front truck dwell time of train control respectively.
3, bullet train working line characteristic model:
Bullet train tracking operating status is closely related with line characteristics, establishes accurate line characteristics model, is tracking
Operation curve setting method is effectively basic.Line characteristics mainly include line condition (vertical section, curve and tunnel) and traction
Mistake split-phase in power supply, in detail as shown in Figure 2.It may be expressed as: in conjunction with Fig. 2 mathematical description that can obtain line characteristics model
wa=wr(pr,lr)+wc(rc,lc)+wt(v,lt) (8)
In formula, (pr,lr),(rc,lc) and (v, lt) respectively indicate (gradient, ramp length), (curvature, length of curve) and
(train speed, length of tunnel).Speed of service feature of the bullet train in split-phase section can be described as:
In formula, gsin (pr) it is gravity acceleration g in the component being parallel on the direction of ramp.
4, bullet train tracks operation curve multiple target setting model:
In moving block system, bullet train is operated in environment complicated and changeable and length dynamic change is closed
It fills in section.Meanwhile the commuter service that bullet train provides needs to meet safety, comfortably requires, energy conservation is also optimal speed
The important goal of curve setting, convergence rate be evaluate Stochastic Optimization Algorithms efficiency important indicator, efficiency of algorithm be then this most
The guarantee of excellent operation curve setting method feasibility.Therefore, based on innovative Measure Indexes, it is bent to establish bullet train tracking operation
Line setting model is as follows:
A1 safety
The bullet train safety accident occurred in recent years is analyzed it is found that knocking into the back with overspeed is major accident reason.?
This, bullet train tracking operational safety is defined as guaranteeing the safe separation distance and not overspeed between two vehicles.Consider real
During the operation of border, table is run at the time of bullet train needs to follow stringent, therefore the size of actual interval distance D is mainly by transporting
Scanning frequency degree v is determined.Using D as constraint of velocity, then safety evaluation index may be expressed as:
In formula, D is bullet train actual interval distance, and v and V (l) are the bullet train speed of service and ATP speed limit respectively.
In bullet train tracking operation course, when v will be more than V (l), vehicle-mounted ATP system can be tight to train automatic implementation
Anxious braking, so V (l)-v > 0 is set up always.ε is a sufficiently small positive real number.fsIt is smaller, then train tracking operation course
The risk that safety accident occurs is smaller.
A2 energy conservation
Energy-saving run is one of the main target of bullet train curve of pursuit setting.Bullet train in tracking operation course
It is a nonlinear dynamic system, operation energy consumption is train traction feature, line condition and train handling sequence collective effect
As a result, it is difficult to directly be calculated.Due to bullet train have powerful inertia, train running speed with control force change and
Variation needs a reaction time (Δ t) as in formula (11).Therefore, entire tracing process is divided into limited sufficiently small
Section, traction energy consumption of the bullet train in each section may be expressed as:
Ei=F (vi)ΔSiM (11)
In formula, Δ Sj=vj·Δt,vj∈ v, Δ t are the sampling period, and M is train weight, F (vi) it is bullet train traction
Power is speed viFunction.
Total traction energy consumption in entire section are as follows:
A3 is comfortable
Ride comfort is the important evaluation index of high-speed rail quality of passenger service.Bullet train acceleration change is too big or changes
Frequency is too high, can all seriously affect riding comfort.This method is added by using to acceleration change amount and rate of acceleration change
Weight coefficient acquires Comfort Evaluation index, as follows:
fc=ω1|a|+ω2|r| (14)
In formula, a is acceleration (a > 0 indicates to accelerate, and a < 0 indicates to slow down), and r is rate of acceleration change, ω1And ω2It is to add
Weight coefficient.Due to the powerful inertia of bullet train itself, influence of the r to comfort level is smaller with respect to a, and weighting coefficient is set as
ω1=0.8 and ω2=0.2.
The constraint of A4 Algorithm Convergence
Algorithm Convergence is the important restrictions for guaranteeing optimized operation rate curve setting result validity.What the present invention used
For multi-objective particle as a kind of efficient Stochastic Optimization Algorithms, convergence conditions are that it one of is constrained substantially.
The algorithm search mechanism may be expressed as:
In formula, xjIt (n) is the position of j-th of particle, pbj(n) and gb (n) respectively indicates the particle optimum position and population
Optimum position,WithIt is acceleration factor, r1 and r2 are equally distributed random numbers in [0,1], and ω is changeable weight, nmax
It is maximum number of iterations.
Enable x*For particle optimal location, then pbj(n) it can transform to gb (n):
In formula, dpj(n) and dg(n) pb is respectively indicatedj(n) and gb (n) arrives x*Euclidean distance.
Since the position of Algorithm Convergence and particle is closely related, and it is almost unrelated with the search speed of particle, by formula
(16) formula (15) are substituted into and eliminates vj(n+1) it can obtain:
Ask expectation that can obtain formula (17):
In formula, r1 and r2 are to be uniformly distributed, then have E (r1)=E (r2)=1/2, X=E (x*) it is Pareto disaggregation center
The expectation of position,It is all constant with X.
Based on formula (18), optimum results global convergence be may be expressed as:
Based on formula (18) and (19), the adequate condition of algorithmic statement can be obtained are as follows:
The convergent adequate condition of this setting method can be obtained from above are as follows:
It establishes high speed using the convergence conditions in A4 as model constraint condition based on evaluation index in the above A1-A3 and arranges
Vehicle is tracked shown in operation curve setting model such as formula (22)-(23):
min f(vsi)={ fs(vsi),fe(vsi),fc(vsi),fcg(vsi)} (22)
In formula, f (vsi) it is composite evaluation function, vsiIt is bullet train tracking operation " speed-mileage " curve, ξ is given
Allowance on schedule.In formula (23) on schedule: being 1. to constrain, be 2. speed limit and security constraint, be 3. algorithmic statement constraint.It solves
The rate curve setting model obtains the Pareto disaggregation of optimized operation curve.
5, the optimal tracking operation curve setting method assessment models of bullet train
Bullet train high density under movable block tracks operation, the fortune between the train tracked in same block section between stations
Row state is closely related.Therefore, the setting of bullet train operation curve by directly affect entire block section between stations efficiency of operation and
Stability.Here, obtaining the screening index of one group of optimal solution using efficiency of operation and stability as from Pareto solution concentration.
B1 efficiency of operation:
Under moving block system, bullet train tracking operation does not set fixed block section, so the evaluation of efficiency of operation
Using runing time required between standing as scale.The present invention is by period TsThe interior bullet train theory quantity by certain block section between stations
NtIt is defined as efficiency of operation, as follows:
Nt=Ts/ti (24)
In formula, Nt∈R+, tiRuning time between standing needed for the bullet train for passing through the section for the i-th column.We set herein
It is fixed, according to the bullet train for the optimized operation curve motion that the present invention is set:The calculating formula of k
ForΔSjIdentical as formula (11), Δ t is sampling period, L0For distance between sites, ρ is given parking essence
Degree.Under the premise of meeting high-speed railway multiple target service requirement, tiSmaller then efficiency of operation is higher.
B2 runs stability:
Under movable block in multiple row bullet train tracking operation course, since emergency case causes certain column bullet train to start
It is unavoidable to there is late situation.Therefore, the subsequent bullet train runing time t run between stationiTime t late to front truckd
Recovery capability will directly affect the operation stability in entire section.It is as follows that the present invention will run definition of stability:
In formula, T0The service time between station given in operation timetable, λ is the coefficient of stability.λ > 0 indicates to stablize, and
In a certain range, the λ the big, and it is better to run stability;λ < 0 then indicates unstable.
Implementation of the invention is research object with CHR380AL type bullet train (14 dynamic 2 drag, and 1,16 be trailer).Acquisition should
Model bullet train is in JinanActual operating data in the East of Xuzhou, in conjunction with data such as actual motion condition, routes,
The optimal tracking operation curve carried out under movable block sets simulating, verifying.
The actual operating data of acquisition is used for the trained bullet train speed prediction model based on echo state network, is surveyed
Try model accuracy, test effect such as Fig. 4.The test experiments are repeated N=100 times, 100 prediction of speed error mean squares can be obtained
The average and standard deviation of root error (RMSE) is respectively Em=0.0051 and std=± 5.74e-5。
Based on the model constraint in the above speed prediction model and formula (9), (23), solving speed curve setting model can be obtained
Operation curve disaggregation after setting, and compared with the operation curve before setting, verify the validity of this setting method.Before setting
Operation curve includes: the actual operation curve of collection in worksite, as shown in Figure 5;The operation curve generated at random based on model above
(the multiple target setting model is not used), as shown in Figure 6.The corresponding fitness value of Fig. 5 and Fig. 6 is respectively i=a and i in table 1
=b.It is compared by Fig. 5 and Fig. 6 and their fitness function value it is found that the operation curve not being set still has very big optimization
Space.Pareto disaggregation obtained by setting model is solved as shown in i=c in table 1.I=a in contrast table 1, b, c is it is found that after setting
The corresponding energy consumption of bullet train operation curve and comfort level have larger improvement, safety allowance is also improved.
The fitness function value of 1 Pareto disaggregation of table
In table, i and j are respectively Pareto disaggregation and the serial number for tracking operation curve.I=a, b, c, d are respectively Fig. 5, and 6,
The corresponding fitness function value of operation curve in 7,8.
2 bullet train of table runs Comfort Evaluation standard
With efficiency of operation evaluation index defined in formula (24)-(25), the corresponding efficiency of operation of disaggregation i=c can be obtained such as
(setting T shown in j=3-7 in table 3s=7200s).It can show that optimal solution is j=6 by table 3, according to the height of the setting curve motion
Fast train tracking effect is as shown in Figure 7.
To test effect of this setting method in terms of guaranteeing that stability is runed in section, it is assumed that emergency event together occurs and leads
Front truck is caused to occur the reduction of speed of 303km/h-293km/h in the 653.5-656.5km of section, while the reduction of speed leads to the late t of front truckd
=36s, service time T between the station given in time-table0=4611s.Based on the late situation, using setting side of the invention
Method sets the optimal tracking operation curve of rear car, sets result as shown in the i=d in table 1.It should with operation stability indicator assessment
Disaggregation, as a result as shown in the j=8-12 on 3 right side of table.
3 Pareto disaggregation efficiency of operation of table and stability assessment result
In table, j is identical as table 1." Jinan-Xu Zhoudong " running schedule is inquired, the parking of intermediate station " Tai'an " station is obtained
Time is 60s, and sets the bullet train run in the section herein and fix in the station down time.
By j=8-12 in table 3 and formula (26) it is found that in the case where front truck is late, solve obtained by multiple target setting model
Operation curve all meet operation stability requirement.Wherein the stability of j=10 is best, and bullet train is according to the curve tracing
The effect of operation is as shown in Figure 8.It may be seen that the operation curve of rear car fluctuates.After the fluctuation is mainly slowed down by front truck
Caused by the variation of two vehicle spacing distances, i.e., as D < LnWhen, rear car is slowed down;As D > LnWhen, rear car correspondingly accelerates to meet other
Target.
Claims (6)
1. a kind of bullet train tracks operation curve optimal setting method, which comprises the following steps:
Step 1: the bullet train speed prediction model based on echo state network is established;Establish bullet train tracing model;It builds
Vertical line characteristics model;
Step 2: the bullet train based on echo state network established with the actual operating data of acquisition, training step one
Speed prediction model parameter;
Step 3: based on trained bullet train speed prediction model and random generation N group control sequence collection { cli, it obtains pair
Operation curve collection { the vs answeredi};
Step 4: acquiring speed limit, speed and position data in real time, and solving bullet train tracing model can obtain: D, Lm、Ln;
Step 5: judge D-LnWhether < ε is true;Practical tracking interval of the D between forward and backward vehicle, LnFor forward and backward workshop gauge
From threshold value, ε is a sufficiently small positive real number;If establishment skips to step 7, step 6 is otherwise skipped to;
Step 6: the speed limit of rear car is adjusted are as follows: V (l)=v1(l), V (l) is the ATP speed limit of bullet train operation;
Step 7: being based on line characteristics model, establishes bullet train tracking operation curve multiple target setting model, it is bent to obtain operation
Line Pareto disaggregation;
Step 8: according to setting method assessment models efficiency of operation index, from Pareto solve concentrate screening obtain one group it is optimal
Solve vsx;
Step 9: vs is assessed using operation stability indicatorx;
Step 10: judge vsxWhether meet stability requirement, if it is satisfied, then terminating, otherwise skips to step 3 and continue to execute.
2. bullet train according to claim 1 tracks operation curve optimal setting method, which is characterized in that step 1
In, the bullet train speed prediction model is described as follows:
Y (t+1)=v (t+1), U (t+1)=[v (t);v(t-1);v(t-3);cl(t+1)] (3)
In formula, U, X, Y be an externally input respectively, intrinsic nerve member state, network output matrix/vector, Win、W、WoutRespectively
Input weight, intrinsic nerve member connection weight, output weight matrix/vector;V (t+1) is the operation at bullet train current time
Speed, cl (t+1) ∈ { -1,0,1 } be current time train handle level, " -1,0,1 " respectively indicate high-speed train braking,
Coasting, traction working condition.
3. bullet train according to claim 1 tracks operation curve optimal setting method, which is characterized in that step 1
In, the method for establishing bullet train tracing model are as follows:
Provide minimum safety interval distance LmWith the spacing distance threshold value L for judging whether rear car operating status is influenced by front truckn,
It is as follows:
Ln=Lm+L1, D=d1-d2 (5)
In formula, laAnd lbIt is two vehicle minimums parking spacing and train length of wagon respectively;c1And c2It is subtracting for front truck and rear car respectively
Speed, L1And L2It is the emergency stopping distance of front truck and the service braking distance of rear car;v1And v2It is the speed of front truck and rear car;
ew1、ev1And ew2、ev2It is the measurement error of the positioning of front truck and rear car, the speed of service respectively;D actually chasing after between forward and backward vehicle
Track interval, d1And d2It is the position of forward and backward vehicle;D≤LnWhen, the operating status of rear car is influenced by front truck operating status.
4. bullet train according to claim 1 tracks operation curve optimal setting method, which is characterized in that step 1
In, it is described to establish line characteristics model are as follows:
wa=wr(pr,lr)+wc(rc,lc)+wt(v,lt) (6)
In formula, (pr,lr),(rc,lc) and (v, lt) respectively indicate (gradient, ramp length), (curvature, length of curve) and (train
Speed, length of tunnel);Under the collective effect of air drag and self gravity, operation speed of the bullet train in split-phase section
Degree feature can be described as:
In formula, gsin (pr) it is gravity acceleration g in the component being parallel on the direction of ramp.
5. bullet train according to claim 1 tracks operation curve optimal setting method, which is characterized in that the step
In seven, the method that bullet train tracks operation curve multiple target setting model is established are as follows:
Provide safety, energy conservation and comfortable Measure Indexes and model constraint are as follows:
A1 safety
Using D as constraint of velocity, then safety evaluation index may be expressed as:
In formula, D is bullet train actual interval distance, and v and V (l) are the bullet train speed of service and ATP speed limit respectively;In height
In fast train tracking operation course, when v will be more than V (l), vehicle-mounted ATP system can promptly make train automatic implementation
It is dynamic, so V (l)-v > 0 is set up always;ε is a sufficiently small positive real number;fsSmaller, then train tracking operation course occurs
The risk of safety accident is smaller;
A2 energy conservation
Entire tracing process is divided into limited sufficiently small section, traction energy consumption of the bullet train in each section can indicate
Are as follows:
Ei=F (vi)ΔSiM (11)
In formula, Δ Sj=vj·Δt,vj∈ v, Δ t are the sampling period, and M is train weight, F (vi) it is bullet train tractive force, be
Speed viFunction;
Total traction energy consumption in entire section are as follows:
A3 is comfortable
By acquiring Comfort Evaluation index, following institute using weighting factor method to acceleration change amount and rate of acceleration change
Show:
fc=ω1|a|+ω2|r| (14)
In formula, a is acceleration, and a > 0 indicates to accelerate, and a < 0 indicates to slow down, and r is rate of acceleration change, ω1And ω2It is weighting system
Number;Due to the powerful inertia of bullet train itself, influence of the r to comfort level is smaller with respect to a, and weighting coefficient is set as ω1=
0.8 and ω2=0.2;
The constraint of A4 Algorithm Convergence
The multi-objective particle of use as a kind of efficient Stochastic Optimization Algorithms, convergence conditions be its substantially about
One of beam;
The algorithm search mechanism may be expressed as:
In formula, xjIt (n) is the position of j-th of particle, pbj(n) and gb (n) respectively indicates the particle optimum position and population is best
Position,WithIt is acceleration factor, r1 and r2 are equally distributed random numbers in [0,1], and ω is changeable weight, nmaxIt is most
Big the number of iterations;
Enable x*For particle optimal location, then pbj(n) it can transform to gb (n):
In formula, dpj(n) and dg(n) pb is respectively indicatedj(n) and gb (n) arrives x*Euclidean distance;
Since the position of Algorithm Convergence and particle is closely related, and it is almost unrelated with the search speed of particle, by formula (16) generation
Enter formula (15) and eliminates vj(n+1) it can obtain:
Ask expectation that can obtain formula (17):
In formula, r1 and r2 are to be uniformly distributed, then have E (r1)=E (r2)=1/2, X=E (x*) it is Pareto disaggregation center
It is expected thatIt is all constant with X;
Based on formula (18), optimum results global convergence be may be expressed as:
Based on formula (18) and (19), the adequate condition of algorithmic statement can be obtained are as follows:
The convergent adequate condition of this setting method can be obtained from above are as follows:
It establishes bullet train using the convergence conditions in A4 as model constraint condition based on evaluation index in the above A1-A3 and chases after
Shown in track operation curve setting model such as formula (22)-(23):
min f(vsi)={ fs(vsi),fe(vsi),fc(vsi),fcg(vsi)} (22)
In formula, f (vsi) it is composite evaluation function, vsiIt is bullet train tracking operation " speed-mileage " curve, ξ is given standard
Point allowance;In formula (23) on schedule: being 1. to constrain, be 2. speed limit and security constraint, be 3. algorithmic statement constraint;Solve the speed
It writes music line setting model, obtains the Pareto disaggregation of optimized operation curve.
6. bullet train according to claim 1 tracks operation curve optimal setting method, which is characterized in that step 8
In, the appraisal procedure according to setting method assessment models are as follows:
The screening index of one group of optimal solution is obtained using efficiency of operation and stability as from Pareto solution concentration;
(1) efficiency of operation:
By period TsThe interior bullet train theory quantity N by certain block section between stationstIt is defined as efficiency of operation, as follows:
In formula, Nt∈R+, tiRuning time between station needed for the bullet train for passing through the section for the i-th column;It sets herein, according to
The bullet train of the optimized operation curve motion of setting:The calculating formula of k isΔSj=vj·Δt,vj∈v,L0For distance between sites, Δ t is the sampling period, and ρ is given parking essence
Degree;By formula it is found that under the premise of meeting high-speed railway multiple target service requirement, tiSmaller then efficiency of operation is higher;
(2) stability is runed:
By subsequent Train Schedule tiTime t late to front truckdRecovery capability be defined as operation stability, it is as follows:
In formula, T0The service time between station given in operation timetable, λ is the coefficient of stability;λ > 0 indicates to stablize, and certain
In range, the λ the big, and it is better to run stability;λ < 0 then indicates unstable.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611252207.3A CN106777752B (en) | 2016-12-30 | 2016-12-30 | A kind of bullet train tracking operation curve optimal setting method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611252207.3A CN106777752B (en) | 2016-12-30 | 2016-12-30 | A kind of bullet train tracking operation curve optimal setting method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106777752A CN106777752A (en) | 2017-05-31 |
CN106777752B true CN106777752B (en) | 2019-04-02 |
Family
ID=58928113
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611252207.3A Active CN106777752B (en) | 2016-12-30 | 2016-12-30 | A kind of bullet train tracking operation curve optimal setting method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106777752B (en) |
Families Citing this family (24)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109918687B (en) * | 2017-12-08 | 2021-01-15 | 通号城市轨道交通技术有限公司 | Train dynamics simulation method and simulation platform based on machine learning |
CN108263449B (en) * | 2017-12-08 | 2020-04-28 | 天津津航计算技术研究所 | Urban rail train tracking method based on speed tracking |
CN108394429B (en) * | 2018-02-06 | 2020-04-07 | 重庆交通大学 | Method for generating automatic driving curve for urban rail train group |
CN108725518B (en) * | 2018-04-18 | 2020-08-18 | 中国神华能源股份有限公司 | Heavy haul train tracking interval calculation method and device and heavy haul railway system |
CN108764571B (en) * | 2018-05-28 | 2020-04-14 | 西南交通大学 | Multi-objective optimization method for heavy-duty train operation |
CN109033521B (en) * | 2018-06-25 | 2021-04-20 | 中南大学 | Newly-built railway slope-limiting optimization decision method |
CN109345022B (en) * | 2018-10-17 | 2021-08-31 | 重庆交通大学 | Method for optimizing railway fixed block partition |
CN109635246B (en) * | 2018-12-06 | 2021-07-13 | 西南交通大学 | Multi-attribute data modeling method based on deep learning |
CN110155126B (en) * | 2019-06-03 | 2020-09-15 | 北京交通大学 | High-speed train dispatching and control integrated optimization method under temporary speed limit |
CN112977546B (en) * | 2019-12-13 | 2022-11-11 | 上海申通地铁集团有限公司 | Track traffic train tracking interval shrinking and changing adjustment control method |
CN111257827B (en) * | 2020-01-16 | 2023-07-14 | 玉林师范学院 | High-precision non-line-of-sight tracking and positioning method |
CN111460633A (en) * | 2020-03-19 | 2020-07-28 | 南京理工大学 | Train energy-saving operation method based on multi-target particle swarm algorithm |
CN112329340B (en) * | 2020-10-31 | 2022-11-18 | 同济大学 | Auxiliary parking area arrangement optimization method for normally-conductive high-speed magnetic levitation unidirectional running line |
CN112419204B (en) * | 2021-01-25 | 2021-04-13 | 江苏德劭信息科技有限公司 | Driving intelligent control method based on improved echo state network |
CN112883474B (en) * | 2021-03-02 | 2021-09-24 | 中南大学 | Layout method, system, terminal and readable storage medium for intelligent reconstruction of existing track line |
CN112977556B (en) * | 2021-04-27 | 2021-08-31 | 北京全路通信信号研究设计院集团有限公司 | Train braking curve segmentation method and system and train control method |
CN113111442A (en) * | 2021-04-27 | 2021-07-13 | 上海大学 | Subway real-time simulation method and system and computer readable storage medium |
CN113147841A (en) * | 2021-05-13 | 2021-07-23 | 中车长春轨道客车股份有限公司 | Rail vehicle capacity management and energy-saving auxiliary driving method and related device |
CN113361150B (en) * | 2021-08-11 | 2021-12-24 | 华东交通大学 | Multi-objective optimization method and system for urban train operation |
CN114212124A (en) * | 2021-11-25 | 2022-03-22 | 通号城市轨道交通技术有限公司 | Train main line tracking method and device, electronic equipment and storage medium |
CN113997915B (en) * | 2021-11-26 | 2022-06-24 | 北京大象科技有限公司 | Big data-based automatic train operation ATO (automatic train operation) accurate parking control method |
CN114735047B (en) * | 2022-06-09 | 2022-08-23 | 西南交通大学 | Train control method and system for high-speed railway train control system |
CN114919632A (en) * | 2022-06-14 | 2022-08-19 | 通号城市轨道交通技术有限公司 | Traction calculation simulation method and device |
CN115146389B (en) * | 2022-07-19 | 2024-03-12 | 江西理工大学 | Permanent magnet magnetic levitation train dynamics feature modeling method |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102955884A (en) * | 2012-11-23 | 2013-03-06 | 同济大学 | Safety distance calibration method in full-speed areas during following operation of high-speed train |
CN104424507A (en) * | 2013-08-28 | 2015-03-18 | 杨凤琴 | Prediction method and prediction device of echo state network |
CN105978732A (en) * | 2016-06-27 | 2016-09-28 | 华北理工大学 | Method and system for optimizing parameters of minimum complexity echo state network based on particle swarm |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
PT104246B (en) * | 2008-10-31 | 2010-12-28 | Manutencao De Equipamento Ferroviario S A Emef Empresa De | SYSTEM AND METHOD FOR TELEMANUTENTION AND PERICIAL TROUBLE DIAGNOSIS |
CN103336908B (en) * | 2013-07-23 | 2016-01-20 | 哈尔滨工业大学 | The method for predicting residual useful life of dull echo state network |
CN104346517A (en) * | 2013-08-02 | 2015-02-11 | 杨凤琴 | Echo state network based prediction method and prediction device |
CN104765916B (en) * | 2015-03-31 | 2018-01-19 | 成都天佑创软科技有限公司 | A kind of Dynamics Performance of High Speed Trains parameter optimization method |
CN104809292B (en) * | 2015-04-28 | 2017-11-28 | 西安理工大学 | A kind of on-line identification method of bullet train non-linear dynamic model parameter |
-
2016
- 2016-12-30 CN CN201611252207.3A patent/CN106777752B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102955884A (en) * | 2012-11-23 | 2013-03-06 | 同济大学 | Safety distance calibration method in full-speed areas during following operation of high-speed train |
CN104424507A (en) * | 2013-08-28 | 2015-03-18 | 杨凤琴 | Prediction method and prediction device of echo state network |
CN105978732A (en) * | 2016-06-27 | 2016-09-28 | 华北理工大学 | Method and system for optimizing parameters of minimum complexity echo state network based on particle swarm |
Non-Patent Citations (4)
Title |
---|
Online optimization for the multi-objective operation of electric multiple units based speed limits curve;Hui Yang et al.;《Proceeding of the 11th World Congress on Intelligent Control and Automation》;20141231;全文 |
动车组追踪运行多目标实时优化策略;杨辉 等;《控制工程》;20150331;第22卷(第2期);全文 |
基于任意闭塞系统的高速列车追踪运行优化控制;刘鸿恩;《中国优秀硕士学位论文全文数据库 信息科技辑》;20151115(第11期);全文 |
基于自适应LSSVM模型的动车组运行速度控制;杨辉 等;《铁道学报》;20150930;第37卷(第9期);全文 |
Also Published As
Publication number | Publication date |
---|---|
CN106777752A (en) | 2017-05-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106777752B (en) | A kind of bullet train tracking operation curve optimal setting method | |
CN109835375B (en) | High-speed railway train automatic driving system based on artificial intelligence technology | |
CN104809554B (en) | A kind of DSS and method towards road traffic accident emergency disposal | |
De Martinis et al. | Data-driven perspectives for energy efficient operations in railway systems: Current practices and future opportunities | |
CN112614344B (en) | Hybrid traffic system efficiency evaluation method for automatic driving automobile participation | |
CN104881527B (en) | Urban railway transit train ATO speed command optimization methods | |
CN106710215B (en) | Bottleneck upstream lane grade traffic status prediction system and implementation method | |
CN108053645A (en) | A kind of signalized intersections cycle flow estimation method based on track data | |
CN109508751B (en) | Deep neural network model modeling method for high-speed railway train late time prediction | |
CN112307564A (en) | Method and device for optimizing ATO target running speed curve of train | |
CN110299011A (en) | A kind of traffic flow forecasting method of the highway arbitrary cross-section based on charge data | |
CN112686453B (en) | Intelligent prediction method and system for locomotive energy consumption | |
CN107346460A (en) | Following operating mode Forecasting Methodology based on the lower front truck operation information of intelligent network contact system | |
CN103116608A (en) | Method of reproducing traffic flow on express way | |
CN111063204B (en) | Expressway vehicle speed prediction model training method based on toll station flow | |
Valera et al. | Driving cycle and road grade on-board predictions for the optimal energy management in EV-PHEVs | |
CN108960496A (en) | A kind of deep learning traffic flow forecasting method based on improvement learning rate | |
CN102376075A (en) | Rail transit line transporting capacity evaluating method based on CBTC (Communication Based Train Control) | |
Zhu et al. | An automated vehicle fuel economy benefits evaluation framework using real-world travel and traffic data | |
Qu et al. | Robust optimization of train timetable and energy efficiency in urban rail transit: A two-stage approach | |
CN114187766B (en) | Road service level evaluation method based on saturation rate | |
Hamid et al. | Investigation into train positioning systems for saving energy with optimised train trajectories | |
Liu et al. | Optimizing energy savings for a fleet of commercial autonomous trucks | |
Jin et al. | Energy-optimal speed control for connected electric buses considering passenger load | |
Yin et al. | Balise arrangement optimization for train station parking via expert knowledge and genetic algorithm |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |