CN-121998205-A - Intelligent algorithm-based energy-saving optimization method for rail train

CN121998205ACN 121998205 ACN121998205 ACN 121998205ACN-121998205-A

Abstract

The invention discloses an intelligent algorithm-based energy-saving optimization method for a railway train, which comprises the steps of constructing an energy consumption objective function under the condition of continuity based on an electric traction force, a mechanical braking force and a traction chain nonlinear loss model, constructing constraint conditions such as electric traction force, mechanical braking force and line speed limit, performing interval discretization processing on the energy consumption objective function, uniformly dividing a train running line into N nodes according to a total distance S in a space domain to form discretized speed vectors so as to obtain discretized representation of the energy consumption objective function, optimizing the speed vectors by adopting an improved differential evolution algorithm to obtain optimal speeds of all the nodes by using self-adaptive mutation operators of at least two mutation strategies according to the energy consumption objective function and the constraint conditions, and finally outputting the optimal speeds of all the nodes as train energy-saving running control instructions. According to the invention, a collaborative optimization mechanism of a high-precision energy consumption model and an improved differential evolution algorithm is constructed, and the optimization efficiency and the solution quality are obviously improved on the premise of ensuring engineering constraint.

Inventors

CHEN DEWANG
CHEN XIUBIN
MENG ZHENYU
LIU LIN
ZHOU QIANG
LIN RONGRONG

Assignees

福建理工大学

Dates

Publication Date: 20260508
Application Date: 20260408

Claims (9)

1. The intelligent algorithm-based energy-saving optimization method for the rail train is characterized by comprising the following steps of: constructing an energy consumption objective function under continuous conditions, namely constructing the energy consumption objective function under continuous conditions based on an electric traction force, a mechanical braking force and a traction chain nonlinear loss model, , , wherein, Is the total distance of the line and, Is a state variable that is a function of the state, , Is already driving to the distance Is used for the time period of (a), Is to travel to distance Speed at time Is defined by the square of (a), Is a control variable which is used to control the operation of the device, , For positive, traction, and vice versa, Is regenerative braking; the mechanical braking force is always positive; for travelling to distance Sum of losses of all elements at the time; Constructing constraint conditions, namely constructing constraint conditions of train operation according to the capability of train equipment and the characteristics of a line, wherein the constraint conditions at least comprise electric traction force constraint, mechanical braking force constraint, line speed limit constraint, running time constraint and loss of a traction chain; the energy consumption objective function is subjected to interval discretization under the continuous condition, namely, a train running line is uniformly divided into N nodes based on the total distance S of the line on the space domain, and discretized speed vectors are formed based on the speeds of the nodes Wherein, the method comprises the steps of, Representing the speed of the kth node, The energy consumption objective function is expressed in discrete space as ; As the power tractive effort of the kth node, Total concrete loss for the kth node; is the distance between two adjacent nodes; According to the energy consumption objective function and constraint conditions, adopting an improved differential evolution algorithm to optimize a speed vector by using adaptive mutation operators of at least two mutation strategies to obtain the optimal speed of each node, wherein a train driving distance s is used as an individual code, a new individual is generated in a feasible space through binomial crossover, feasibility processing is carried out on individuals which do not meet the constraint conditions, a mutation vector is dynamically selected to generate for each individual in the evolution stage by the corresponding mutation strategy, the optimal speed of each node is obtained according to individual fitness distribution optimization, and a discretized speed vector is formed based on the speed of each node As a population, introducing a perturbation mechanism based on a history optimal track to generate new individuals to replace stagnant individuals when the population falls into local optimal stagnation; And outputting the optimal speed of each node as a control instruction for energy-saving operation of the train.
2. The intelligent algorithm-based energy-saving optimization method for the rail train, as set forth in claim 1, is characterized in that the total loss of all the elements is calculated as follows: , wherein, Is the loss of the transformer and, Is the loss of the motor and the converter, Is the loss of the auxiliary equipment and, Is the loss of the gearbox.
3. The intelligent algorithm-based energy-saving optimization method for the rail train, as set forth in claim 1, wherein the total concrete loss of the kth node is The total specific loss is based on corresponding regenerative braking Sum speed of Is calculated by nonlinear relation of total concrete loss The expression of (2) is as follows: ; Wherein, the And Constant efficiencies for traction and regenerative braking, respectively.
4. The intelligent algorithm-based energy-saving optimization method for rail trains as set forth in claim 3, wherein the method comprises the following steps: And The values of (2) are all 0.73.
5. The intelligent algorithm-based energy-saving optimization method for the rail train is characterized in that mutation strategies comprise a DE/rand/1 strategy and a DE/best/1 strategy, a judgment method for population diversity is set, when the fitness value of a child population is larger than that of a parent population, the current population is judged to be in stagnation, when the stagnation algebra of the population in the evolution process is larger than 10, the population diversity is judged to be low, otherwise, the population diversity is judged to be high, when the population diversity is high, the DE/rand/1 strategy is adopted to enhance global searching capability, and when the population gradually converges, the weight of the DE/best/1 strategy is increased to accelerate local convergence speed.
6. The intelligent algorithm-based energy-saving optimization method for the rail train, as set forth in claim 5, wherein the expression of DE/rand/1 is: ; The expression of DE/best/1 is: , Wherein, the Is the first The variation vector of the individual is calculated, 、 And Is three individual vectors which are randomly selected from the population and are different from each other, Is the optimal individual in the current population; Is a differential scaling factor.
7. The intelligent algorithm-based energy-saving optimization method for rail trains according to claim 1, wherein new individuals are generated based on a perturbation mechanism of a historical optimal track The expression of (2) is: ; Wherein, the Is the first Globally optimal individuals in generation evolution History of tracks of (a); 、 And Is three individual vectors which are randomly selected from the population and are different from each other, Is a differential scaling factor that is used to scale the image, Is a historical track Weight coefficient of (c) in the above-mentioned formula (c).
8. The intelligent algorithm-based energy-saving optimization method for the rail train, as set forth in claim 1, wherein the constraint conditions include: (1) Electric traction constraint: , wherein, The electric traction force of the train is; is the line distance; is the maximum value of the electric traction force, and is determined by the minimum value in the limits of the traction system, the power limit and the acceleration constraint; The minimum value of the electric traction force is determined by the maximum value of the brake system limit, the regeneration power limit and the comfort degree constraint; (2) Mechanical braking force constraints: , wherein, The mechanical braking force of the train is; a maximum value of the mechanical braking force that can be provided for the brake; (3) Line speed limit constraint: Or is , wherein, Is the train speed; For train speed Is the square of (2); setting an upper limit of line speed according to infrastructure conditions of track section speed limit, curve radius and gradient; (4) Resistance constraint: , wherein, The total resistance of the train; Is the rolling resistance of the roller, and the roller is the rolling resistance, Is the air resistance, the air resistance is high, Is the curve of the additional resistance force, Is gradient resistance; (5) Loss of the traction chain: , wherein, Is the loss of the traction chain; Is the loss of the motor and the current transformer, Is the loss of the transformer and, Is the mechanical loss of the gearbox, Is auxiliary equipment loss; (6) And (3) driving time constraint: , wherein, For the total time spent by the train completing the line, Setting an upper limit of total consumed time for the train to reach the end position; (7) Boundary condition constraints: , , the square of the speeds of the start point and the end point of the train meets the boundary requirement, wherein ; The time of the start is indicated as such, To start speed Is defined by the square of (a), Is the reaching of the end point speed Square of (d).
9. The intelligent algorithm-based energy-saving optimization method for the rail train, as set forth in claim 1, is characterized in that individuals which do not meet the constraint condition are processed in a punishment function mode, namely, the individuals which do not meet the constraint condition are assigned infinity so as to mark elimination.

Description

Intelligent algorithm-based energy-saving optimization method for rail train Technical Field The invention relates to the technical field of traffic management, in particular to an intelligent algorithm-based energy-saving optimization method for rail trains. Background With the rapid development of high-speed railways and urban rail transit, the problem of energy consumption in the running process of trains is increasingly focused. In order to improve the transportation efficiency and reduce the operation cost, the existing research is mainly developed around three types of methods, namely 1) an energy consumption analysis method based on a classical dynamic equation, wherein the method generally adopts an idealized electric traction force, resistance and speed relation and simplifies the energy consumption into the product integral of the electric traction force and the speed, 2) a track planning method based on an optimal control theory is solved through Pontryagin maximum principle, convex optimization or dynamic programming, but traction chain loss is assumed or ignored by using 'fixed efficiency', so as to generate an idealized operation mode such as Bang-Coast-Bang, and 3) an energy-saving scheduling method based on an intelligent optimization algorithm, such as differential evolution, genetic algorithm, particle swarm, and the like, realizes high-dimensional track optimization through global search, but most of the methods still adopt linear or ideal energy consumption models, and cannot reflect the real nonlinear loss characteristics of an engineering system. In addition, some studies focus on regenerative braking and energy feedback mechanisms, but the actual feedback amount is limited by the motor, inverter and grid absorption capacity, and regenerative braking exhibits significant saturation and nonlinear characteristics, which are often not fully incorporated by conventional models. In general, while existing methods provide beneficial exploration for energy-efficient train control, it is generally difficult to accurately reflect the actual energy consumption characteristics of the traction chain and the braking system, depending on excessively simplified energy consumption assumptions, thereby limiting engineering applicability of the optimization results. However, the prior art has the following technical disadvantages in practical engineering application: Firstly, the existing model generally adopts the assumption of 'fixed efficiency' or 'ideal lossless', and ignores nonlinear losses (such as copper loss, iron loss and mechanical loss) of a traction motor, a converter, a gear box and auxiliary equipment in a traction chain, so that traction energy consumption is systematically underestimated, and the energy consumption prediction of an optimized track deviates from a true value, and particularly, the energy consumption difference caused by the nonlinear losses cannot be reflected under high traction/high regeneration working conditions. Secondly, many studies do not consider the coupling relationship of regenerative braking to mechanical braking force, either neglecting mechanical braking force or assuming that braking is entirely borne by electrical regeneration, and do not consider regenerative braking to be limited by the absorption capacity of the inverter, motor and grid, resulting in excess braking to be generated by mechanical braking force and its energy not being recoverable. Third, existing methods often treat conditions such as speed, acceleration, line speed limit, etc. in a soft constraint or empirical correction manner, which may cause the track to violate the line speed limit, and the infeasible acceleration to occur in a slope or curve section, and thus cannot be directly used in a train control system. Fourth, many methods still construct models in the time domain, and infrastructure parameters such as resistance, speed limit, gradient and the like of the high-speed railway are defined in the space domain, so that the time domain modeling is difficult to accurately describe dynamics constraint and coupling characteristics of a train-line, and numerical errors are easy to be introduced in the variable transformation process. Finally, the traditional optimization algorithm has obvious limitations when processing high-dimensional, strong-constraint and nonlinear energy consumption models, the optimal control method is difficult to obtain an analytic solution after adding a complex loss model, while the intelligent optimization algorithm has flexibility, but lacks an effective feasibility restoration mechanism, is difficult to ensure the continuity and feasibility of a speed track, is easy to sink into local optimum and causes slow convergence speed, and in addition, the traditional differential evolution, genetic algorithm and particle swarm algorithm are easy to generate problems of rapid decline of population diversity, slow sink into local optimum and convergence speed and