Traction Characteristic Control Method and System for Energy Saving of Subway Trains

Abstract

The present invention discloses a traction characteristic control method and system for energy-saving auxiliary driving of subway trains. By adding a driving energy-saving control module to the control system of subway trains, based on the synchronous following asynchronous control of the operation curve and combined with the train central control to jointly control the traction / electromagnetic braking force required by each motor car, according to the train traction characteristics, integrating the finite set model predictive control module, performing rolling optimization control, accurately controlling the exertion of the traction / electromagnetic braking force of each motor car, further optimizing the asynchronous control performance under the coordinated operation of each motor car in the subway train, reducing the energy consumption of the subway train, improving the train operation quality, and providing technical support for the green operation of the subway train.

Description

Technical Field

[0001] The present invention relates to the technical field of energy saving for subway trains, and provides a traction characteristic control method and system for energy-saving auxiliary driving of subway trains. The present invention aims to achieve energy-saving operation control of subway trains. Specifically, it solves problems such as unsatisfactory tracking control of the train operation speed curve and fluctuations in the actual train speed curve. The method uses a controller to precisely control the traction force and electric braking force of the train to ensure precise tracking control between the actual train operation speed curve obtained by the driver's operation and the optimal target speed curve. At the same time, the energy consumption of the train is evaluated through energy consumption analysis, and the power distribution is carried out according to the analysis results. Specifically, the invention monitors the energy utilization of the train traction system and electric braking system in real time and makes corresponding adjustments to the train to achieve the purpose of energy saving. Background Art

[0002] As the largest type of urban rail transit system, subway trains play an important role in the market. However, their energy consumption cannot be ignored. In recent years, with the release of the national dual-carbon policy and the green development initiative, it has become particularly important to optimize the energy saving of subway trains. During the operation of subway trains, traction / electric braking is the most important energy conversion link. Therefore, by precisely controlling the exertion of the traction / electric braking force of the train, unnecessary energy consumption can be effectively reduced to achieve the purpose of energy saving. Summary of the Invention

[0003] The present invention aims to reduce the energy consumption of subway trains. To solve this problem, a traction characteristic control method and system for energy-saving auxiliary driving of subway trains are provided to accurately track the target speed curve of the train. At the same time, through energy consumption analysis and optimization strategies, the power is reasonably distributed to reduce the operation energy consumption of subway trains.

[0004] To solve the above technical problems, the technical method adopted by the present invention is: a traction characteristic control method and system for energy-saving auxiliary driving of subway trains, including the following steps:

[0005] S1. Obtain the actual operation information of the train and the train attribute information;

[0006] S2. Construct an energy-saving model using the actual operation information of the train and the train attribute information; E total = E t + E r ; where E total is the total energy consumption of the subway vehicle, E t is the traction / electric braking energy consumption (traction energy consumption is positive, and electric braking energy consumption is negative), and E r is the resistance energy consumption;

[0007] S3. Define the optimization objective as minimizing the energy consumption within the prediction time window: where represents the traction / electric braking energy consumption of the train within the k-th time step; represents the energy consumed by the train in overcoming various resistances (such as rolling resistance, air resistance, etc.) within the k-th time step; N represents the total number of time steps within the prediction time window, that is, the number of time steps within the prediction interval for minimizing energy consumption. For example, if each time step is 0.1 second and the prediction interval is 0.3 seconds, then N is 3. Set the constraint conditions of the optimization objective;

[0008] S4. Initialize the constraint conditions, traction / electric braking energy consumption, and resistance energy consumption;

[0009] S5. Calculate the value of the optimization objective function corresponding to the current traction / electric braking force sequence, obtain the gradient of the optimization objective function with respect to each traction / electric braking force, and update the traction / electric braking force sequence using the gradient;

[0010] S6. Determine whether the change rate of the optimization objective function value is less than the set threshold or whether the maximum number of iterations is reached. If so, output the optimized traction / electric braking force sequence;

[0011] S7. Obtain the traction / electric braking force at the k-th time step in the optimized traction / electric braking force sequence, compare the traction / electric braking force at the k-th time step with the actual traction / electric braking force, and correct the energy-saving model according to the comparison error;

[0012] S8. Replace the energy-saving model with the corrected energy-saving model and return to step S3;

[0013] S9. End when the train reaches the destination or the stop platform.

[0014] In the traction characteristic control method for energy-saving auxiliary driving of subway trains, the expression of the energy-saving model is:

[0015] E total =E t +E r

[0016] m·a=F t -F r -mgsinθ

[0017]

[0018] F t =K t ·v -1

[0019]

[0020] F r = F r0 + K r1 ·v + K r2 ·v 2

[0021] Among them, E total is the total energy consumption of the subway vehicle, which is the sum of the traction / electric braking energy consumption and the resistance energy consumption; E t is the traction / electric braking energy consumption; E r is the resistance energy consumption; K t is the coefficient related to the characteristics of the traction motor; v is the train speed; F r0 is the basic resistance; K r1 is the linear resistance coefficient; K r2 is the quadratic resistance coefficient; F t is the traction force; F r is the resistance; m is the mass of the subway vehicle; g is the acceleration due to gravity; θ is the track gradient.

[0022] The parameter identification process for the above-established energy-saving model includes:

[0023] The parameters to be identified in the energy-saving model are the initial basic running resistance F r0_Init , the initial resistance coefficient K r_Init , and the initial traction motor-related coefficient K t_Init . The particle swarm algorithm is used to identify these three parameters. It is defined that the three dimensions of each particle correspond to the values of these three parameter variables, that is, the dimension d of the particle is 3. Set the ranges of these three initial parameter values, as well as the initial velocity and position of the particle;

[0024] After setting the ranges of the three initial parameter values and initializing the particle, define its fitness function, that is, the error between the output of the identification model and the output of the actual model, which is used to evaluate the quality of the position of each particle in each dimension:

[0025]

[0026] Among them, E prep is the energy consumption output by the model, and E actual is the energy consumption of the actual output.

[0027] According to the evaluation of the quality of the position of each particle in each dimension, continuously update the velocity and position of each particle until the iteration reaches the predetermined maximum number of iterations, or the change in the fitness function value converges to a certain solution, stop the iteration, and output the optimal values of the particle in the three dimensions, that is, output the three optimal parameter values. The constraint conditions described in S4 include speed constraint and traction / electric braking force constraint.

[0028] In step S5, the specific calculation process of updating the traction / electric braking force is as follows:

[0029] The updated traction / electric braking sequence obtained in the (n + 1)-th iteration process The expression is: is the updated traction / electric braking force sequence obtained in the n-th iteration, is the gradient of the objective function E(F t ) with respect to the traction / electric braking force at the k-th time step in the n-th iteration, α is the learning rate, and this value is updated in each iteration;

[0030] Before starting the iteration process in S5, first determine the target reference traction / electric braking force range. The acquisition process of the target reference traction / electric braking force range includes:

[0031] Determine the target reference traction / electric braking force range. The acquisition process of this range: Based on the current speed of the train and the running state of the vehicle-line-environment, by constructing a state probability matrix and defining a reward function inversely proportional to the energy consumption, obtain the state value and action value functions, and through calculating the current value function, perform policy adjustment to gradually approach the optimal policy. Under this policy, the long-term cumulative reward can be maximized, that is, the energy consumption can be minimized. Apply the optimal policy to guide the train to select operating conditions in different states, and determine the target reference traction / electric braking force range through the selection of operating conditions. Compared with the prior art, the beneficial effects of the present invention are as follows: According to the energy consumption of the subway train and the fluctuations in the tracking of the subway train operation curve, the present invention establishes the kinematics of the subway train and performs system identification in combination with the actual operation data of the subway train, and this model can be modified according to the actual operation feedback, and provides a traction / electric braking characteristic control method and system for energy-saving auxiliary driving of the subway train according to the main reasons for the energy consumption of the subway train operation.

[0032] This system can predict the target traction / electric braking force in advance according to line data, vehicle load, environmental parameters, etc., and can also analyze the energy consumption situation in real time, and perform operation allocation of the power according to the analysis. This system adopts the Markov optimization strategy to provide the optimal rolling optimization direction for the optimization prediction process, that is, to provide the target reference traction / electric braking force range, accelerate the optimization prediction process, ensure the stability of the tracking speed curve, and reduce fluctuations. The present invention can reduce energy consumption while ensuring the safe operation of the train, and realizes the precise control of the traction / electric braking force. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is the network topology structure diagram of the system and related devices in the embodiment of the present invention;

[0034] Figure 2 Structure diagram of the optimized control system according to the embodiment of the present invention;

[0035] Figure 3 Structure diagram of the core prediction model algorithm according to the embodiment of the present invention;

[0036] Figure 4 Block diagram of the auxiliary driving system of the subway vehicle according to the embodiment of the present invention; Specific implementation manners

[0037] The present invention realizes the precise control of the traction / electric braking force of the subway train by adding an energy-saving auxiliary driving system for the subway train on the basis of the subway train control system to ensure the stability of the speed tracking curve, and further achieves the optimization goals of reducing the traction energy consumption and improving the utilization rate of the regenerative braking energy, and finally realizes the purpose of energy saving. Specifically, it includes the following steps:

[0038] (1) Construct a core control module of a traction characteristic control method and system for energy-saving auxiliary driving of subway trains based on finite set model predictive control, which consists of a three-layer model structure. The first is an energy-saving model based on the traction / braking characteristics of train dynamics. This prediction model is established based on the actual operation data of the subway train and finally obtained through parameter identification. This prediction model can be adjusted according to the actual operation of the subway train and is a variable prediction model. The second is a finite set model predictive control module, which mainly consists of a prediction model, rolling optimization, and feedback correction. The prediction model is the energy-saving model of the subway train in the first part. This module takes energy saving as the optimization goal and finally obtains the target traction / electric braking force through rolling optimization. This module can realize the timely correction of the energy-saving model through feedback correction, so that the energy-saving model can better meet the performance requirements and avoid energy waste. The model prediction processing module can adjust the traction characteristics in real time according to the adhesion of the locomotive and the vehicle-line-environment coupling state. In addition, this module is restricted by the expert system to prevent abnormal operation of the subway train. The third is the Markov optimization strategy, which is mainly applied to the rolling optimization process in the second part. This optimization strategy can control the rolling optimization direction. This optimization strategy can make the optimal decision through the Markov decision process and accelerate the optimization prediction process in 2.

[0039] (2)Construct a traction / electric braking characteristic control method and system for energy-saving auxiliary driving of subway trains, which is mainly divided into three parts. The first part is the input information processing module, which collects and summarizes the data and conditions required for model predictive control by communicating with the electric vehicle interface and vehicle network. The second part is the processing and prediction by the finite set model predictive control, Markov optimization strategy, and expert system. Combining the knowledge of the expert system [Cai Zixing, Delkin, Gong Tao. Advanced Expert System: Principles, Design and Applications [M]. Science Press, 2005.], using the knowledge base and inference engine, the constraints on the target traction / electric braking force are realized. The third part is the optimization output and feedback module, which outputs the target traction / electric braking force of the power vehicle output by the core control module to the traction control unit, and finally the traction control unit further realizes the traction / electric braking force control. The tracking situation of the train speed curve is fed back to the finite set model predictive control system to achieve real-time tracking and situation feedback. This can provide more accurate data for the next optimization prediction process, thus obtaining more accurate control results.

[0040] (3)Integrate information such as train operation data, vehicle load, and line data and input it into the finite set model predictive control module, which is convenient for the core control module to master the actual operation data situation. At the same time, it receives the operation situation of the subway train traction condition and electric braking condition in real time, outputs the target traction force / electric braking force and executes it. Input the speed protection curve output by the ATP system into the core control module to facilitate the advance prediction control of the subway train driving condition. At the same time, input and process the fault information in time so that the system can make timely adjustments and processing.

[0041] (4)Construct an output execution and feedback optimization module for the traction / electric braking characteristic control method and system of energy-saving auxiliary driving of subway trains based on finite set model predictive control. This module outputs the target traction / electric braking force output by the energy-saving auxiliary driving system to the train network control system, that is, the TCMS system. The TCMS system inputs the corresponding traction characteristic curve / electric braking characteristic curve to the traction control unit TCU, and then the TCU unit inputs it to the traction converter and drive motor to execute and output the corresponding traction force / electric braking force. Finally, the accurate tracking control of the speed tracking control of the train automatic driving system control layer is realized, avoiding the occurrence of speed fluctuation, so as to achieve the ultimate energy-saving purpose.

[0042] Example 1

[0043] Such as Figure 1 And Figure 4As shown, the structure of the energy-saving auxiliary driving system for subway vehicles 1 describes the relationships between the various parts of the designed energy-saving auxiliary driving system. Embodiment 1 of the present invention includes the relationships between the energy-saving auxiliary driving system 3 and the Automatic Train Operation (ATO) system 4, the Automatic Train Protection (ATP) system 5, the Train Communication Network Control System (TCMS) 6, and the Traction Control Unit (TCU) 7. This system communicates and exchanges data through the train and vehicle bus 2 to ensure that information (such as train attributes 22, line data 23, vehicle load 24, environmental parameters 25, formation data 26, etc.) can be transmitted between the various modules. This information is transmitted to the decision-making layer 12 in the Automatic Train Operation (ATO) system. The decision-making layer 12 uses this information to calculate the operation protection curve, and the energy-saving auxiliary driving system 3 generates specific optimized train control strategies based on this information and outputs them to the Automatic Train Operation (ATO) system. The Automatic Train Operation (ATO) system gives corresponding control instructions according to the optimized control strategy. The Traction Control Unit (TCU) 7 receives the information passed from the Automatic Train Operation (ATO) system to the Train Communication Network Control System (TCMS) 6, and according to the traction characteristic curve / electric braking characteristic curve 17, converts it into specific traction / braking force output instructions. These instructions are executed through the traction converter and the drive motor 16, and are transmitted to the corresponding drive wheels through the drive device 21, thereby achieving precise control of the subway vehicle and reducing energy consumption. This system also includes an energy consumption analysis module 44, which monitors and analyzes the energy consumption of the train in real time, and feeds back the analysis results to the energy-saving auxiliary driving system 3 to continuously optimize the control strategy. And through the control strategy, the operation distribution of the motor 20 can be analyzed, and it can be adjusted in real time according to the current energy consumption situation and train attributes at any time. The energy-saving auxiliary driving system 3 can achieve more efficient and energy-saving operation on the premise of ensuring safety and punctuality.

[0044] Embodiment 2

[0045] As Figure 2 shown, Embodiment 2 of the present invention includes a finite set model predictive control 9, an expert system 10, a Markov optimization strategy 11, an energy consumption calculation 30, etc. The control strategy of the traction characteristic control method and system for energy-saving auxiliary driving of subway trains based on finite set model predictive control constructed is as in Embodiment 2.

[0046] First, the Automatic Train Protection (ATP) system 5 provides the target speed - distance curve 27 to the Automatic Train Operation (ATO) system to ensure the safe operation of the subway vehicle. The Automatic Train Operation (ATO) system generates an instruction u after comprehensive analysis and decision-making according to the designed control strategy and the target speed - distance curve 27 k (u kThe control command output by the train automatic operation system to the traction / braking system is output to the traction / braking system 28, and finally the traction / braking system 28 outputs Z (where Z represents the traction / electric braking force output by the traction / braking system) to the control object 29, the subway train, to ensure the safe operation of the subway vehicle according to the operation plan while reducing energy consumption.

[0047] The specific control strategy of the traction characteristic control method and system for energy-saving auxiliary driving of subway trains based on finite set model predictive control is as follows: First, by collecting historical train operation data 31, including information such as train load, position, gradient, and running speed, the particle swarm algorithm 36 is used for system identification 35 to identify the initial parameter values of the energy-saving model 32 based on the traction / braking characteristics of train dynamics: initial basic running resistance: F ro_Init , initial resistance coefficient: K r_Init , initial traction motor correlation coefficient: K t_Init . Thus, before the subway vehicle runs, the energy-saving auxiliary driving system can obtain an initial energy-saving model 32 based on the traction / braking characteristics of train dynamics. The specific process includes:

[0048] 1. Set the number of particle swarms to 200. Since there are three parameter values to be determined in the above energy-saving model, define the dimension d of each particle as 3, which are: initial basic running resistance F r0_Init , initial resistance coefficient K r_Init , initial traction motor correlation coefficient K t_Init , and define the position of each dimension of the particle corresponding to these three parameter variable values. According to the statistics of train historical operation data, set the ranges of these three initial parameter values and set the initial speed and position of the particle.

[0049] 2. Define the fitness function using the parameter values of each dimension of each particle: that is, the error between the output of the identification model and the output of the actual model, which is used to evaluate the quality of the position of each particle in each dimension:

[0050]

[0051] Among them, E prep is the energy consumption of the model output, and E actual is the energy consumption of the actual output.

[0052] 3. According to the evaluation of the quality of the position of each particle in each dimension, continuously update the speed and position of each particle so that they gradually approach the optimal solution.

[0053] · Update speed:

[0054]

[0055] Among them, vi,d (t) is the velocity of the i-th particle in the d-th dimension at the t-th iteration, ω is the inertia weight, c1 and c2 are learning factors, r1 and r2 are random numbers, is the historical best position of particle i in the d-th dimension, is the global best position in the d-th dimension.

[0056] · Update the position:

[0057] x i,d (t + 1) = x i,d (t) + v i,d (t + 1)

[0058] where x i,d (t + 1) is the position of the i-th particle in the d-th dimension at the (t + 1)-th iteration,

[0059] v i,d (t + 1) is the velocity of the i-th particle in the d-th dimension at the (t + 1)-th iteration.

[0060] 4. Continuously update the velocity and position of each particle in each dimension until the iteration reaches the predetermined maximum number of iterations, or the change in the fitness function value converges to a certain solution, then stop the iteration and output a set of optimal model parameters, that is, three optimal parameter values.

[0061] 5. Output a set of optimal model parameters: the initial basic running resistance F r0_Init , the initial resistance coefficient K r_Init , the initial traction motor related coefficient K t_Init for establishing the initial energy-saving model.

[0062] The above is the entire process of energy-saving model parameter identification. By using the optimal parameters to establish the energy-saving model, the accuracy and prediction ability of the energy-saving model are improved.

[0063] Conduct rolling optimization based on the above energy-saving model. Assume: T represents a rolling optimization period, k is an integer, kT represents the k-th rolling optimization period, and the rolling optimization period is determined by the characteristic period of the train's TCU. Conduct 33 rolling optimization controls based on this energy-saving model. At each control period kT, use the latest updated energy-saving model data to predict the system output within a finite time domain (prediction time domain) in the future. With the goal of minimizing the total energy consumption minE total as the objective, solve the optimization problem within the prediction time domain to determine the traction / electrical braking force in the future period of time. The specific process is as follows: First, define the optimization objective, and the optimization objective is to minimize the energy consumption within the prediction time window: Set the constraint conditions: speed constraint: ensure that the speed is within the allowable range: v min ≤ vk ≤ v max ; Traction / Electronic Braking Force Constraint: Ensure that the traction / electronic braking force is within the allowable range:

[0064] Initialization: Set the initial state of the system, including the initial velocity v0, the initial traction / electronic braking energy consumption E t0 and the initial resistance energy consumption E r0 . Prediction Step: At the current time k, predict the system behavior for the next N steps based on the current state, and predict the target traction / electronic braking force for each step The optimization process includes:

[0065] 1. Set the initial guess value of the traction / electronic braking force

[0066] 2. Perform optimization with minimizing the energy consumption as the objective function to obtain the optimal traction / electronic braking force sequence Calculate the objective function value of the current traction / electronic braking force sequence:

[0067]

[0068] Steps of the optimization process:

[0069] a. Calculate the gradient of the objective function of the traction / electronic braking force sequence with respect to each traction / electronic braking force τ k :

[0070]

[0071] b. After obtaining the gradient, the traction / electronic braking force sequence can be updated:

[0072]

[0073] where α is the learning rate, which determines the step size of each update, and the value of the learning rate can be updated according to each iteration.

[0074] c. Check the constraint conditions: Ensure that the updated satisfies the constraints of speed and force. If not, make corrections and truncate the values beyond the range to the allowable range.

[0075] d. Determine whether the convergence condition is satisfied, that is, the change in the objective function value is less than the set threshold or the maximum number of iterations N max . If it converges, stop the iteration; otherwise, continue the next iteration.

[0076] e. When the optimization process converges, the obtained traction / electronic braking force sequence is the optimal traction / electronic braking force sequence.

[0077] f. Take the control input at the first time step from the optimal traction / electric braking force sequence and input the control input to the control system.

[0078] 3. Apply the control input at the first time step obtained by optimization from the optimal control sequence Rolling step: Update the system state to the next moment k + 1 according to the implemented control input , including traction / electric braking energy consumption and resistance energy consumption Loop: Repeat the above prediction, optimization, implementation, and rolling steps. When the subway vehicle reaches the parking platform, terminate the loop process.

[0079] After the initial energy-saving model is optimized in the first round of iteration, it will be adjusted according to the specific operating conditions after the train starts running, that is, compare the output of the energy-saving model 32 based on the traction / braking characteristics of the train dynamics with the actual output during the train operation, and perform feedback correction. Input the new characteristic parameter values F ro , K r , K t to the energy-saving model 32 based on the traction / braking characteristics of the train dynamics, and correct the parameters to ensure the accuracy and reliability of the prediction model. Such a real-time feedback mechanism ensures that the control system can adapt to the actual operating conditions of the train in a timely manner, thereby effectively optimizing energy utilization.

[0080] In addition, in order to further improve the speed and accuracy of the optimization process, the Markov optimization strategy 11 is introduced to provide the rolling optimization direction decision, that is, it provides the target reference traction / electric braking force range for the rolling optimization control first, provides a reference value for the rolling optimization control, improves the efficiency of the rolling optimization control, and can help the rolling optimization control determine the prediction result faster. The input of the Markov optimization strategy is: the traction / electric braking force in the previous cycle, and the output is the target reference traction / electric braking force range in the current cycle. The state transition strategy is: judge whether the state x(kT + T) after transition continues to use the traction condition 39 or choose to update the condition to the cruise condition / coasting condition according to the current state x(kT) - 39 traction condition / 41 coasting condition. The specific process is as follows: Define the state space (S), and the state space contains all possible conditions of the train. For example: S 39 : Traction condition; S 40 : Cruise condition; S 41 Coasting condition; S 42 Braking condition, and determine the possible actions that can be taken in each state, define the action space (A), and define the actions that the agent can take in each state. For example: A continue continue: Continue with the current operating condition, A switch : Switch to a new operating condition. Construct the state transition probability matrix (P), which defines the probability P(s'|s,a) of transitioning from state s to state s' by taking action a. Using the state transition probability matrix, the likelihood of taking a specific action is predicted. Define the reward function (R), where the reward function R(s,a) is defined based on the energy consumption when taking action a to transition from state s to state s'. For example: if the action results in a decrease in energy consumption, a positive reward is given. Combining the reward function designed to be inversely proportional to energy consumption, the state value function V π (s) (V π (s) represents the total expected return that can be obtained by following policy π in state s) and the action value function Q π (s,a) (Q π (s,a) represents the total expected return that can be obtained by taking action a in state s and following policy π). By iteratively evaluating the policy (calculating the value function under the current policy) and improving the policy (adjusting the policy based on the value function), the optimal policy π is gradually approximated * , which can maximize the long-term cumulative reward, i.e., minimize the energy consumption. Apply the optimal policy to guide the train's operating condition selection in different states, where

[0081] The traction force calculation formula under the traction operating condition is: F t = F r0 + K r1 ·v + K r2 ·v 2 + m·a

[0082] Traction force range: F t = F r0 , F t_max = F r0 + K r1 ·v + K r2 ·v 2 + m·a;

[0083] The traction force calculation formula under the cruise operating condition is: F cr = F r0 + K r1 ·v + K r2 ·v 2

[0084] Traction force range: F cr_min = F r0 , F cr_max = F r0 + K r1 ·v + K r2 ·v 2 ;

[0085] The traction force calculation formula under the coasting operating condition is: Fco = 0

[0086] Tractive force range: F co = 0;

[0087] The calculation formula for electric braking force under braking conditions is: F br = F r0 + K r1 ·v + K r2 ·v 2 - m·a

[0088] Electric braking force range: F br_min = F r0 - m·a, F br_max = F r0 + K r1 ·v + K r2 ·v 2 - m·a.

[0089] After selecting a specific working condition, the target reference traction / electric braking force range can be obtained according to the current train speed, and finally the target reference traction / electric braking force range will be output to the 33 rolling optimization link to accelerate the prediction process of the 33 rolling optimization link and achieve the minimum energy consumption in the overall operation process. During the rolling optimization control process, further prediction is carried out to output the target traction / electric braking force to the train automatic operation ATO system, and finally the control of the traction / electric braking force is realized, thereby achieving the purpose of energy saving.

[0090] In addition, an expert system 10 is designed to further constrain the predicted output target traction / electric braking force. The expert system 10 includes a knowledge base 43 and an inference engine 48. The knowledge base 43 consists of a constraint condition set 44, a historical data set 45, an energy-saving strategy library 46, and an environmental factor library 47; the inference engine consists of a 49 inference engine, a data processing layer 53, and an output layer 56. The inference engine 49 includes a rule applicator 50, a strategy evaluator 51, and an optimization processor 52, the data processing layer 53 includes a historical analyzer 54 and an energy monitor 55, and the output layer 56 output layer includes a result outputter 57.

[0091] Embodiment 3

[0092] Such as Figure 3As shown in the figure, Embodiment 2 of the present invention includes an energy-saving assisted driving system control module 1 established based on finite set model predictive control. This system module mainly consists of an input information processing module 64, a core control module 65, and an optimization output and feedback module 66. The system is divided into three parts. The first part is the input information processing module 64; the second part is the core control module 65; the third part is the optimization output and feedback module 66. The association relationship among the three parts is that the input information processing module integrates and inputs the processed actual operation information such as the train attributes 22, line data 23, vehicle load 24, environmental parameters 25, formation data 26, and actual train operation conditions (traction condition 39, cruise condition 40, coasting condition 41, braking condition 42) of the subway train to the core control module 65 for predictive optimization control output, and feeds back the train speed curve tracking situation and the energy consumption analysis result to the core control module 65.

[0093] The core control module 65 includes an energy-saving model 32 based on the traction / braking characteristics of train dynamics, system identification 35, finite set model predictive control 9, an expert system 10, a Markov optimization strategy 11, and power operation distribution 19. The finite set model predictive control module 16 uses the energy-saving model 32 based on the traction / electric braking characteristics of train dynamics as the prediction model, and mainly predicts the target traction force / target electric braking force in this module. During this process, the finite set model predictive control 9 is restricted by the expert system 10 to prevent abnormal operation of the subway train. During the rolling optimization process of the finite set model predictive control 9, the Markov optimization strategy 11 is introduced to judge the optimal rolling optimization direction, making the obtained target value better. In addition, a power operation distribution module 19 is set up, which judges the power operation distribution situation according to the energy consumption analysis result 71 and flexibly distributes the power of the subway train.

[0094] The optimization output and feedback module 66 includes the given target value of the traction force of the power car 68, the given target value of the electric braking force of the power car 69, the train speed curve tracking situation 70, and the energy consumption analysis result 71. The given target value of the traction force of the power car 68 and the given target value of the electric braking force of the power car 69 output by the core control module 65 are imported into the optimization output and feedback module 66, and then a control instruction is input to the TCMS system by the ATO system. Finally, the TCMS system executes the output of the traction force / electric braking force to achieve precise control of the traction force / electric braking force of the subway train.

[0095] During the train operation, the train speed curve tracking situation 70 and the energy consumption analysis result 71 are fed back to the core control module 65 in real time, which can enable the core control module 65 to obtain better optimization output and the power operation distribution module 19 to achieve better energy distribution control.

Claims

1. A traction characteristic control method for energy-saving auxiliary driving of a subway train, characterized in that, It includes the following steps: S1. Obtain the actual operation information and train attribute information of the train; S2. Construct an energy-saving model using the actual train operation information and train attribute information; E total = E t + E r ; Among them, E total is the total energy consumption of the subway vehicle, E t is the traction / electric braking energy consumption, and E r is the resistance energy consumption; S3. Define the optimization objective as minimizing the energy consumption within the prediction time window: Among them represents the traction / electric braking energy consumption of the train within the k-th time step represents the resistance energy consumption of the train within the k-th time step; N represents the total number of time steps within the prediction time window; set the constraint conditions for the optimization objective S4. Initialize the constraint conditions, traction / electric braking energy consumption, and resistance energy consumption; S5. Set the target reference traction / electric braking force range, calculate the optimization objective function value corresponding to the current traction / electric braking force sequence within the target reference traction / electric braking force range, obtain the gradient of the optimization objective function value with respect to each traction / electric braking force, and update the traction / electric braking force sequence using the gradient; S6. Determine whether the change rate of the optimization objective function value is less than the set threshold or whether the maximum number of iterations is reached. If so, output the updated traction / electric braking force sequence; S7. Obtain the traction / electric braking force at the k-th time step in the updated traction / electric braking force sequence, compare the traction / electric braking force at the k-th time step with the actual traction / electric braking force, and correct the energy-saving model according to the comparison error; S8. Replace the energy-saving model with the corrected energy-saving model and return to step S3; S9. End when the train reaches the destination or the parking platform.

2. The traction characteristic control method for energy-saving auxiliary driving of a subway train according to claim 1, characterized in that, The expression of the energy-saving model is: Among them, E total is the total energy consumption of the subway vehicle, which is the sum of the traction / electric braking energy consumption and the resistance energy consumption; E t is the traction / electric braking energy consumption, E r is the resistance energy consumption, K t is a coefficient related to the characteristics of the traction motor, v is the train speed, F r0 is the basic resistance, K r1 is the linear resistance coefficient, K r2 is the quadratic resistance coefficient, F t is the traction force, F r is the resistance, m is the mass of the subway vehicle, g is the acceleration due to gravity, and θ is the track gradient.

3. The traction characteristic control method for energy-saving auxiliary driving of a subway train according to claim 2, characterized in that, The parameter identification process of the energy-saving model includes: Initialize the basic running resistance F r0_Init , the initial resistance coefficient K r_Init , the initial traction motor related coefficient K t_Init , define that the three dimensions of each particle respectively correspond to the basic running resistance F r0_Init , the initial resistance coefficient K r_Init , the initial traction motor related coefficient K t_Init , set the ranges of the three initial parameter values and the initial velocity and position of the particles; Define the fitness function MSE as follows: Among them, E prep is the energy consumption output by the energy-saving model, and E actual is the actually output energy consumption; Evaluate the quality of the position of each particle in each dimension, update the velocity and position of each particle until the iteration reaches the predetermined maximum number of iterations or the change in the fitness function value converges, stop the iteration, and output the optimal values of the particle in three dimensions, that is, output three optimal parameter values.

4. The traction characteristic control method for energy-saving auxiliary driving of a subway train according to claim 1, characterized in that In step S4, the constraint conditions include speed constraints and traction / electric braking force constraints.

5. The traction characteristic control method for energy-saving auxiliary driving of a subway train according to claim 1, characterized in that, In step S5, the updated traction / electric braking sequence obtained in the (n + 1)-th iteration process The expression is: The updated traction / electric braking force sequence obtained for the n-th iteration At the n-th iteration, the objective function E(F t ) with respect to the traction / electric braking force at the k-th time step , where α is the learning rate; the objective function is represents the total energy consumption at the k-th time step, where is the traction energy consumption at the k-th time step, is the resistance energy consumption at the k-th time step.

6. The traction characteristic control method for energy-saving auxiliary driving of a subway train according to claim 1, characterized in that, The process of obtaining the target reference traction / electric braking force range includes: Based on the current speed of the train and the train-line-environment operation state, by constructing a state probability matrix and defining a reward function inversely proportional to the energy consumption, obtain the state value and action value functions, adjust the strategy according to the current action value function, and gradually approach the optimal strategy of minimizing energy consumption; use the optimal strategy to guide the train to select operating conditions in different states, and determine the target reference traction / electric braking force range through the selection of operating conditions.

7. A traction characteristic control system for energy-saving auxiliary driving of a subway train, characterized in that, It includes a memory, a processor, and a computer program stored on the memory; characterized in that the processor executes the computer program to implement the steps of the method according to any one of claims 1 to 6.

Images