A method, medium and equipment for overall optimization design of railway hump horizontal and vertical surfaces
The overall optimization design of the flat and vertical surface of the railway hump is solved through the multi-objective particle swarm algorithm, which solves the limitations of the individual optimization of the flat and vertical surfaces in the existing technology, realizes more efficient vehicle slipping and reduces construction costs, and improves the overall hump performance.
Patent Information
- Application Number
- CN202411131679.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-19
- Publication Date
- 2024-11-26
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing railway camel design mainly relies on empirical methods and semi-automated systems, resulting in separate optimization of the flat and vertical planes, excessive reduction in decision variables, small solution space, failure to form overall optimization, and difficult to meet the improvement of efficiency and safety.
The multi-objective particle swarm algorithm is used to optimize the overall optimization of the camel plane and longitudinal section model, combining the vehicle slip time and construction project costs as the objective function, the plane and longitudinal section scheme is optimized to achieve the overall optimization of the two-layer architecture. .
It improves the optimization design efficiency and effect, ensures the safe and fast vehicle release, reduces the total cost of construction projects, and enhances the disintegration ability and land utilization rate of Hump.
Smart Images

Figure CN118656905B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of hump design, and in particular to a method, medium and equipment for overall optimization design of the horizontal and vertical surfaces of a railway hump. Background Art
[0002] At present, the design and operation of railway yard humps mainly rely on traditional empirical methods and semi-automated systems. These methods have certain limitations in improving efficiency and ensuring safety. Domestic and foreign scholars have conducted long-term and in-depth research on this, but all of them adopt the idea of decomposing it into separate optimization of the horizontal and longitudinal sections. In the existing plane optimization methods, they mainly rely on experience and compactness principles, resulting in over-simplification of decision variables, too small solution space, and easy omission of important solutions. An overall optimization method has not been formed in the longitudinal section. The hump is a three-dimensional device, and this kind of horizontal and vertical separation optimization is essentially a sub-optimization under a finite goal. It is urgent to study the overall optimization method of the horizontal and longitudinal sections of the hump.
[0003] Therefore, comprehensive improvements are made to the deficiencies of previous designs. For any plane scheme, the optimal solution of the lower model is generated and then the overall optimization is performed to solve the problems existing in the prior art. Summary of the invention
[0004] The object of the present invention is to provide a method for overall optimization design of the plane and longitudinal section of a railway hump for comprehensively optimizing the plane and longitudinal section of the hump. The specific technical scheme is as follows:
[0005] A method for overall optimization design of railway hump horizontal and vertical surfaces comprises the following steps:
[0006] S1: Simplify and determine the parameters of the hump's plane design and longitudinal section design, and build the overall plane and longitudinal model of the hump, including:
[0007] Plane model, taking curve angle, curve radius, and straight line segment adjustment value as plane decision variables;
[0008] For the longitudinal section model, the mileage of the slope change point, the slope of the slope section, and the vertical curve radius are taken as the longitudinal section decision variables;
[0009] The shunting time of the train on the difficult line and the height difference between the uncoupling point of the long train group and the peak are used as the longitudinal section objective function, and the total length constraint of the push slope section and the acceleration slope constraint are used as the longitudinal section constraint conditions;
[0010] The position constraint of the slope change point is used as the constraint condition of the overall horizontal and vertical model;
[0011] The average shunting time of the train and the total cost of the construction project are used as the objective function of the overall horizontal and vertical model;
[0012] S2: Input the feasible plane scheme of the hump into the overall plane and longitudinal model, and use the multi-objective particle swarm algorithm to obtain the optimized longitudinal section scheme corresponding to the feasible plane scheme;
[0013] S3: Combine the optimized longitudinal section scheme with the feasible plane scheme to solve the objective function of the overall plane and longitudinal model, and use the multi-objective particle swarm algorithm to obtain the optimized plane scheme;
[0014] S4: Based on the optimized plane solution, optimize again to obtain the optimal solution of the corresponding longitudinal section solution and determine whether it converges. If it does not converge, return to S3. If it converges, output the optimized plane and longitudinal overall model.
[0015] Preferably, the plane model and the longitudinal section model are specifically:
[0016] The hump is set as a fixed design style before the first branch turnout according to the compactness principle and track matching requirements; the multiple harnesses connected to the shunting line after the first branch turnout are regarded as fixed modules, and the connection units between the turnouts are used as parameters to establish a plane model;
[0017] The straight line element is determined by the positions of two slope change points, and the circular curve is calculated based on the vertical curve radius at the given slope change point and is used to establish the longitudinal section model;
[0018] The decision variables of the overall model of the hump's horizontal and vertical planes are expressed as follows:
[0019] Curve Corner Vector : ;
[0020] Curve Radius Vector : ;
[0021] Line segment adjustment value vector : ;
[0022] Slope change point mileage column vector : ;
[0023] Slope column vector of the next slope section at the slope change point : ;
[0024] Vertical curve radius column vector : ;
[0025] in, is the curve corner, is the curve radius, Adjust the value for the straight line segment, is the mileage of the slope change point, is the slope of the next slope section at the slope change point, is the vertical curve radius, is the number of plane connection units, is the number of slope change points.
[0026] Preferably, the objective function of the overall planar and longitudinal model is:
[0027] The hump dismantling capacity is evaluated based on whether the hump can safely and quickly shun the train to the corresponding line; the feasibility of the hump scheme is evaluated based on the hump investment cost; the average shunting time of the train and the total cost of the construction project are used as the objective function of the overall horizontal and vertical model;
[0028] The objective function expression of average shunting time is: ;
[0029] in, is the minimum value of the average shunting time function, and the overall shunting time of any train is ; is the average sled time of the last vehicle in the sled route; The first car in the series The time between the vehicle and the next vehicle at the top of the mountain; is the number of vehicles in the train;
[0030] The total construction project cost includes track laying costs, turnout construction costs and speed reducer costs. Track laying costs include the laying costs of rails, sleepers, and roadbeds under sleepers, as well as related power facility costs. Excluding the turnout part, the expression is:
[0031] ;
[0032] In the formula, Total cost of laying track for the hump surface; The laying cost per linear meter of hump track; is the length of a single line in the hump plane design; is the total length of all lines in the hump plane design; The length of the line occupied by a single turnout arranged for the hump; The length of the line occupied by all the switches arranged for the hump;
[0033] The cost of turnout construction is related to the turnout model and is calculated by multiplying the number of turnouts by the construction cost of a single turnout of the corresponding model:
[0034] ;
[0035] Total cost of laying the hump turnout; Cost of laying individual turnouts for different types; is the number of turnouts of the corresponding type;
[0036] The reducer cost is related to the reducer model and the number of nodes set, and the expression is: ;
[0037] Total cost of building a hump reducer; Construction costs for different types of reducers per section; is the number of reducer sections of the corresponding type;
[0038] The objective function expression of the total cost of construction project is:
[0039] ;
[0040] in Represents the minimum value of the total cost of the construction project.
[0041] Preferably, the longitudinal section objective function is specifically:
[0042] Single vehicle release to any mileage Time Expressed as: ;
[0043] Where: is the function of the vehicle's release speed changing with mileage and elevation;
[0044] The objective function of the shunting time of the train on the difficult line is recorded as:
[0045] ;
[0046] The minimum time for a train to slack out on a difficult line; is the total release time;
[0047] Assume that the average unit length gravity of the train group is The unit resistance of the vehicle group is The total length of the train set is ,expression:
[0048] ;
[0049] Find:
[0050] ;
[0051] Where: is the slope of the hook slope; is the acceleration slope; The length of the straight line for the insulated sections of the difficult line where the train slips to the difficult line; The length of the straight line between the insulating sections of the easy-to-drive line and the difficult-to-drive adjacent line;
[0052] The height difference between the decoupling point and the peak platform is for: ;
[0053] The height difference objective function between the uncoupling point and the peak of the long train group is expressed as: .
[0054] Preferably, the longitudinal section constraint conditions are specifically:
[0055] In the longitudinal section, the total length constraint of the push slope segment and the acceleration slope constraint are recorded as ;
[0056] The total length constraint of the push slope section is expressed as: ;
[0057] Where: For the train leader; To push part The length of the slope section; The total number of slope sections for the push part;
[0058] The acceleration slope constraint of the sled section is expressed as: ;
[0059] Where: The mileage of the slope change point between the acceleration slope and the intermediate slope; is the length of the vertical curve tangent between the acceleration slope and the intermediate slope; It is the safety distance between the vertical curve and the turnout basic rail; The basic track mileage before the first branch turnout;
[0060] The position of the slope change point in the horizontal and vertical coupling is expressed as a constraint condition:
[0061] ;
[0062] Where: is the mileage of the current slope change point; The end mileage of the switch or reducer range before the current slope change point; The starting mileage of the switch or reducer range after the current slope change point; It is the tangent length of the vertical curve at the current slope change point.
[0063] Preferably, the step S2 specifically includes:
[0064] S2-1: The position and slope range of the slope change point meet the longitudinal section constraints, and a feasible plane solution is input to obtain the initial particle swarm of the feasible plane solution;
[0065] S2-2: Generate the initial longitudinal section plan, including:
[0066] Within the mileage range of the slope change point and the slope range of the longitudinal section constraint condition, the mileage array and the slope array of the slope change point are generated in 100 equal divisions, and the 100 first longitudinal section schemes are formed by correspondingly combining them one by one in the order from small to large;
[0067] Traverse these 100 first longitudinal section schemes, calculate the satisfaction of the constraint conditions of each scheme, record the lowest peak height scheme and the highest peak height scheme that meet the constraints, and use their slope and slope change point mileage as the new constraint range respectively;
[0068] Within the new constraint range, the slope section is divided into 50 equal parts to generate 50 second longitudinal section schemes;
[0069] The corresponding slope sections of the second longitudinal section scheme are randomly replaced to form 50 initial longitudinal section scheme groups;
[0070] S2-3: After the initial longitudinal section scheme group is determined, all particles are individually optimal, and the global optimal individual is determined by Parato dominance relationship and crowding ranking; iterate according to the particle swarm algorithm calculation formula, calculate the new initial longitudinal section scheme group position and randomly generate any number between (0,1) , calculate the acceptance probability of the new position ,like , update the position of the scheme, otherwise the scheme will not be updated, and finally the optimized longitudinal section scheme is obtained.
[0071] Preferably, the step of obtaining the optimized plane solution includes:
[0072] There are only three planar decision variables for a single-beam hump: the curve angle , Curve radius And the straight line segment adjustment value ; The turnout and reducer are determined by the position of the straight section of the track. When generating the plane plan, it is necessary to consider the straight line constraint between the last branch turnout and the curve and the line spacing constraint of the track in the harness;
[0073] Generate a single-line plan, including: taking the straight line segment adjustment value as =0m, the angle range is calculated with an iterative step length of 1°, and the step length is shortened by binary search at the critical point to obtain the critical range of the angle; the angle of the single-beam plane solution is uniformly taken in the critical range of the angle according to the number of particles, and the length of the straight line segment is uniformly taken in the range of 0~50m to generate 50 plane solution particles; the required energy height of each braking position is determined, and the maximum reducer energy height is taken to adjust the number of reducer sections at the braking position of the plane solution; according to the connection constraint processing method, the plane solution after adjusting the number of reducer sections at the braking position is constrained, and 20 solutions are randomly selected from the solutions that meet the constraints after the treatment as the initial plane solution for iteration;
[0074] A decision-making method incorporating the idea of a tournament is adopted. Based on the sorting of the objective function, the evaluation indicators of the non-inferior solutions are compared in turn. The priority of each solution is calculated according to the victory value of each indicator. All non-inferior solutions are sorted in this way, so as to select the optimal plane solution that combines multiple objectives. The calculation process is as follows:
[0075] Championship Formula ;
[0076] Sorting equation ;
[0077] ;
[0078] for The plan Project function and The victory value of the scheme compared with the corresponding objective function; is a set of non-inferior solutions; is the number of solutions in the non-inferior solution set; and For the plan and No. Item target function value; for The ranking value of the solution in the set of non-inferior solutions; For the The weight value corresponding to the project target function; The number of optimization objective functions of the current model; is a non-inferior solution set Medium Solution No. Project function and Other options in The success value compared with the project target function;
[0079] The congestion in MOPSO is also taken into consideration; the external archive in MOPSO is retained, and the single congestion sequence is changed to a double sequence of congestion and tournament, and the roulette algorithm is used to establish the probability decision equation of the double sequence:
[0080] ;
[0081] Where: For selection The probability that the solution is optimal as a whole; and For crowdedness sequence and tournament sequence The ranking value of the scheme.
[0082] The application of the technical solution of the present invention has the following beneficial effects:
[0083] The overall optimization design method of the plane and longitudinal surface of the railway hump provided by the present invention includes: S1: simplifying and determining the parameters of the plane design and longitudinal section design of the hump, and constructing the overall plane and longitudinal surface model of the hump; S2: inputting the feasible plane scheme of the hump into the overall plane and longitudinal surface model, and adopting the multi-objective particle swarm algorithm to obtain the optimized longitudinal section scheme corresponding to the feasible plane scheme; S3: combining the optimized longitudinal section scheme with the feasible plane scheme to solve the objective function of the overall plane and longitudinal surface model, and adopting the multi-objective particle swarm algorithm to obtain the optimized plane scheme; S4: optimizing again based on the optimized plane scheme to obtain the optimal solution of the corresponding longitudinal section scheme and judging whether it converges, if it does not converge, returning to S3, and if it converges, outputting the optimized plane and longitudinal surface overall model. The present invention performs overall optimization of the plane and longitudinal surface of the hump based on a double-layer architecture, constructs an overall optimization model with plane parameters and longitudinal section optimal solution as decision variables, shunting efficiency and construction project cost as overall objective functions, and plane and longitudinal section geometric line type and equipment function requirements as constraints, thereby improving the efficiency and effect of optimization design.
[0084] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned railway hump horizontal and vertical surface overall optimization design method is implemented.
[0085] The present invention also provides an electronic device, comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute the aforementioned railway hump horizontal and vertical surface overall optimization design method by executing the executable instructions.
[0086] In addition to the above-described purposes, features and advantages, the present invention has other purposes, features and advantages. The present invention will be further described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0087] The accompanying drawings constituting a part of the present invention are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the accompanying drawings:
[0088] Figure 1 A flowchart of a method according to an exemplary embodiment of the present invention;
[0089] Figure 2 A flowchart of a multi-objective particle swarm algorithm provided in an embodiment;
[0090] Figure 3 A comparison diagram of the planar solutions provided for the embodiments;
[0091] Figure 4 A comparison diagram of the longitudinal section scheme provided for the embodiment. DETAILED DESCRIPTION
[0092] The embodiments of the present invention are described in detail below with reference to the accompanying drawings, but the present invention can be implemented in many different ways as defined and covered by the claims.
[0093] refer to Figure 1 , a railway hump horizontal and vertical surface overall optimization design method, comprising the following steps:
[0094] S1: In the early stage, it is necessary to input basic parameters and determine the design scale of the hump, which mainly includes basic information and design scale, so as to simplify and determine the parameters of the plane design and longitudinal section design of the hump, and build the overall plane and longitudinal model of the hump; it includes the plane model and the longitudinal section model, taking the curve turning angle, curve radius, and straight section adjustment value as the plane decision variables; taking the slope change point mileage, slope of the slope section, and vertical curve radius as the longitudinal section decision variables, specifically:
[0095] The hump is set as a fixed design style before the first branch turnout according to the compactness principle and track matching requirements; the multiple harnesses connected to the shunting line after the first branch turnout are regarded as fixed modules, and the connection units between the turnouts are used as parameters to establish the plane model; the straight line element is determined by the positions of the two slope change points, and the circular curve is calculated according to the vertical curve radius at the given slope change point to establish the longitudinal section model;
[0096] The decision variables of the overall model of the hump's horizontal and vertical planes are expressed as follows:
[0097] Curve Corner Vector : ;
[0098] Curve Radius Vector : ;
[0099] Line segment adjustment value vector : ;
[0100] Slope change point mileage column vector : ;
[0101] Slope column vector of the next slope section at the slope change point : ;
[0102] Vertical curve radius column vector : ;
[0103] in, is the curve corner, is the curve radius, Adjust the value for the straight line segment, is the mileage of the slope change point, is the slope of the next slope section at the slope change point, is the vertical curve radius, is the number of plane connection units, is the number of slope change points.
[0104] The shunting time of the train on the difficult line and the height difference between the uncoupling point of the long train set and the peak are used as the longitudinal section objective function, and the total length constraint of the push slope section and the acceleration slope constraint are used as the longitudinal section constraint conditions; the position constraint of the slope change point is used as the constraint condition of the overall horizontal and vertical model; the average shunting time of the train and the total cost of the construction project are used as the objective function of the overall horizontal and vertical model; specifically:
[0105] The objective function of the overall model of the horizontal and vertical planes is:
[0106] The hump dismantling capacity is evaluated based on whether the hump can safely and quickly shun the train to the corresponding line; the feasibility of the hump scheme is evaluated based on the hump investment cost; the average shunting time of the train and the total cost of the construction project are used as the objective function of the overall horizontal and vertical model;
[0107] The objective function expression of average shunting time is:
[0108] ;
[0109] in, is the minimum value of the average shunting time function, and the overall shunting time of any train is ; is the average sled time of the last vehicle in the sled route; The first car in the series The time between the vehicle and the next vehicle at the top of the mountain; is the number of vehicles in the train;
[0110] The total construction project cost includes track laying costs, turnout construction costs and speed reducer costs. Track laying costs include the laying costs of rails, sleepers, and roadbeds under sleepers, as well as related power facility costs. Excluding the turnout part, the expression is:
[0111] ;
[0112] In the formula, Total cost of laying track for the hump surface; The laying cost per linear meter of hump track; is the length of a single line in the hump plane design; is the total length of all lines in the hump plane design; The length of the line occupied by a single turnout arranged for the hump; The length of the line occupied by all the switches arranged for the hump;
[0113] The cost of turnout construction is related to the turnout model and is calculated by multiplying the number of turnouts by the construction cost of a single turnout of the corresponding model:
[0114] ;
[0115] Total cost of laying the hump turnout; Cost of laying individual turnouts for different types; is the number of turnouts of the corresponding type;
[0116] The reducer cost is related to the reducer model and the number of nodes set, and the expression is:
[0117] ;
[0118] Total cost of building a hump reducer; Construction costs for different types of reducers per section; is the number of reducer sections of the corresponding type;
[0119] The objective function expression of the total cost of construction project is:
[0120] ;
[0121] in Represents the minimum value of the total cost of the construction project.
[0122] Longitudinal section objective function:
[0123] Single vehicle release to any mileage Time Expressed as: ;
[0124] Where: is the function of the vehicle's release speed changing with mileage and elevation;
[0125] The objective function of the shunting time of the train on the difficult line is recorded as:
[0126] ;
[0127] The minimum time for a train to slack out on a difficult line; is the total release time;
[0128] Assume that the average unit length gravity of the train group is The unit resistance of the vehicle group is The total length of the train set is ,expression:
[0129] ;
[0130] Find:
[0131] ;
[0132] Where: is the slope of the hook slope; is the acceleration slope; The length of the straight line for the insulated sections of the difficult line where the train slips to the difficult line; The length of the straight line between the insulating sections of the easy-to-drive line and the difficult-to-drive adjacent line;
[0133] The height difference between the decoupling point and the peak platform is for: ;
[0134] The height difference objective function between the uncoupling point and the peak of the long train group is expressed as:
[0135] .
[0136] Longitudinal section constraints:
[0137] In the longitudinal section, the total length constraint of the push slope segment and the acceleration slope constraint are recorded as ;
[0138] The total length constraint of the push slope section is expressed as: ;
[0139] Where: For the train leader; To push part The length of the slope section; The total number of slope sections for the push part;
[0140] The acceleration slope constraint of the sled section is expressed as: ;
[0141] Where: The mileage of the slope change point between the acceleration slope and the intermediate slope; is the length of the vertical curve tangent between the acceleration slope and the intermediate slope; It is the safety distance between the vertical curve and the turnout basic rail; The basic track mileage before the first branch turnout;
[0142] The position of the slope change point in the horizontal and vertical coupling is expressed as a constraint condition:
[0143] ;
[0144] Where: is the mileage of the current slope change point; The end mileage of the switch or reducer range before the current slope change point; The starting mileage of the switch or reducer range after the current slope change point; It is the tangent length of the vertical curve at the current slope change point.
[0145] S2: See Figure 2 , the feasible plane scheme of the hump is input into the overall plane and longitudinal model, and the multi-objective particle swarm algorithm is used to obtain the optimized longitudinal section scheme corresponding to the feasible plane scheme;
[0146] S2-1: The position and slope range of the slope change point meet the longitudinal section constraints, and a feasible plane solution is input to obtain the initial particle swarm of the feasible plane solution;
[0147] S2-2: Generate the initial longitudinal section plan, including:
[0148] Within the mileage range of the slope change point and the slope range of the longitudinal section constraint condition, the mileage array and the slope array of the slope change point are generated in 100 equal divisions, and the 100 first longitudinal section schemes are formed by correspondingly combining them one by one in the order from small to large;
[0149] Traverse these 100 first longitudinal section schemes, calculate the satisfaction of the constraint conditions of each scheme, record the lowest peak height scheme and the highest peak height scheme that meet the constraints, and use their slope and slope change point mileage as the new constraint range respectively;
[0150] Within the new constraint range, the slope section is divided into 50 equal parts to generate 50 second longitudinal section schemes;
[0151] The corresponding slope sections of the second longitudinal section scheme are randomly replaced to avoid the slope section change rule being too monotonous, improve the diversity of the scheme, and form 50 initial longitudinal section scheme groups;
[0152] S2-3: After the initial longitudinal section scheme group is determined, all particles are individually optimal, and the global optimal individual is determined by Parato dominance relationship and crowding ranking; iterate according to the particle swarm algorithm calculation formula, calculate the new initial longitudinal section scheme group position and randomly generate any number between (0,1) , calculate the acceptance probability of the new position ,like , update the position of the scheme, otherwise the scheme will not be updated, and finally the optimized longitudinal section scheme is obtained, which is the optimal solution of the current longitudinal section.
[0153] S3: Combine the optimized longitudinal section scheme with the feasible plane scheme to solve the objective function of the overall plane and longitudinal model, and use the multi-objective particle swarm algorithm to obtain the optimized plane scheme, including:
[0154] There are only three planar decision variables for a single-beam hump: the curve angle , Curve radius And the straight line segment adjustment value ; The turnout and reducer are determined by the position of the straight section of the track. When generating the plane plan, it is necessary to consider the straight line constraint between the last branch turnout and the curve and the line spacing constraint of the track in the harness;
[0155] Generate a single-line plan, including: taking the straight line segment adjustment value as =0m, the angle range is calculated with an iterative step length of 1°, and the step length is shortened by the binary method at the critical point to obtain the critical range of the angle; the angle of the single-beam plane solution is uniformly taken in the critical range of the angle according to the number of particles, and the length of the straight line segment is uniformly taken in the range of 0~50m, and 50 plane solution particles are generated; the required energy height of each braking position is judged, and the maximum reducer energy height is taken to adjust the number of brake position reducer sections of the plane solution; according to the connection constraint processing method, the plane solution after adjusting the number of brake position reducer sections is constrained, and 20 solutions are randomly selected from the solutions that meet the constraints after the treatment as the initial plane solution, and the initial solution particle group of the overall plane and longitudinal model is generated for iteration;
[0156] A decision-making method incorporating tournament thinking is adopted. Based on the ranking of objective functions, the evaluation indicators of non-inferior solutions are compared in turn. The priority of each solution is calculated according to the victory value of each indicator. All non-inferior solutions are ranked in this way, so as to select the optimal plane solution that combines multiple objectives.
[0157] The calculation process is as follows:
[0158] Championship Formula ;
[0159] Sorting equation ;
[0160] ;
[0161] for The plan Project function and The victory value of the scheme compared with the corresponding objective function; is a set of non-inferior solutions; is the number of solutions in the non-inferior solution set; and For the plan and No. Item target function value; for The ranking value of the solution in the set of non-inferior solutions; For the The weight value corresponding to the project target function; The number of optimization objective functions of the current model; is a non-inferior solution set Medium Solution No. Project function and Other options in The success value compared with the project target function;
[0162] In order to ensure the overall diversity of the scheme, the congestion in MOPSO is also taken into consideration; the external archive in MOPSO is retained, and the single congestion sequence is changed to a double sequence of congestion and tournament, and the roulette algorithm is used to establish the probability decision equation of the double sequence:
[0163] ;
[0164] Where: For selection The probability that the solution is optimal as a whole; and For crowdedness sequence and tournament sequence The ranking value of the scheme.
[0165] S4: Based on the optimized plane solution, optimize again to obtain the optimal solution of the corresponding longitudinal section solution and determine whether it converges. If it does not converge, return to S3. If it converges, output the optimized plane and longitudinal overall model.
[0166] The following is an example of the overall optimization design method of the horizontal and vertical planes of the railway hump in this embodiment, taking the design of a hump in a northern city as an example. The marshaling yard where it is located connects 5 surrounding railway lines and is a local marshaling yard on the road network. The early design plan considered a low marshaling capacity, a small-capacity hump with 12 shunting lines, a small number of vehicles disassembled per day, and a short effective length of the arrival and departure lines. However, with the increasing density of the surrounding road network, its marshaling capacity has gradually become saturated. According to the current density of the road network and the forecast of future traffic growth, the plan requires the renovation and expansion of the existing shunting yard, utilizing the existing 6+6 shunting yard, and adding 12 shunting lines and reserving 8 shunting lines to ensure the marshaling capacity requirements. Considering the symmetry of the hump design, the shunting yard is designed as a 4-line harness.
[0167] The hump of this design is a medium-capacity hump, the shunting part is equipped with a secondary brake position, and the shunting yard is a point-to-point speed regulation system; there are 2 push lines and 2 shunting lines, the angle between the front arrival yard and the center axis of the hump is 2°, and the center of the crossover line is 2m away from the center axis. Figure 4 The double-push single-slip operation mode is adopted, and the no-slip line and the detour line are configured. The maximum train weight after hump dismantling is 5,000 tons. The selection standards of locomotives, laying rails, switches, reducers and other related equipment all adopt the common standard models on the market.
[0168] Since only the straight sections of the shunting yard are used, the impact on the track laying is small, so they are not distinguished during optimization. The entire hump is calculated and compared according to the new construction mode. The hardware device used by the program is Intel Core i7-7700, 3.6 GHz processor, 8 GB RAM. The calculation time is 15 hours and 27 minutes.
[0169] This method combines the design characteristics of the hump, performs overall optimization of the hump's horizontal and vertical sections based on a double-layer architecture, and constructs an overall optimization model with plane parameters and the optimal solution of the longitudinal section as decision variables, shunting efficiency and construction project costs as the overall objective function, and horizontal and vertical section geometric line types and equipment functional requirements as constraints, thereby improving the efficiency and effect of the optimization design.
[0170] See also Figure 3 and Figure 4, which is a comparison between the plan and longitudinal section scheme of the overall optimization design method of the railway hump plane and longitudinal section in this embodiment and the traditional scheme, increases the straight line adjustment value of the connecting part, reduces the curve angle, and finally shortens the shunting distance from the peak to the last branch turnout, reduces the total shunting resistance, and makes the arrangement between the wire harnesses more compact, thereby improving the land utilization rate in front of the hump shunting yard. At the same time, the reduction of the total shunting resistance also reduces the energy required for the shunting of the vehicle, so that the overall elevation of the longitudinal section can be reduced, providing a good plane foundation for its optimization scheme. From the longitudinal section comparison diagram, it can be seen that the height difference between the uncoupling point of the long train and the peak is not too low, and the slope section model of the shunting part is more in line with the minimum deceleration curve, which not only saves the shunting time of the vehicle, but also reduces the engineering investment of the filling.
[0171] This embodiment also discloses a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for overall optimization design of the horizontal and vertical surfaces of railway humps described in this embodiment is implemented.
[0172] This embodiment also discloses an electronic device, including:
[0173] processor;
[0174] and a memory for storing executable instructions for the processor;
[0175] Wherein, the processor is configured to execute the railway hump horizontal and vertical surface overall optimization design method as described in this embodiment by executing the executable instructions.
[0176] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A method for overall optimization design of the horizontal and vertical surfaces of railway humps, characterized in that: The steps include: S1: Simplify and determine the parameters of the hump's plane design and longitudinal section design, and build the overall plane and longitudinal model of the hump, including: Plane model, taking curve angle, curve radius, and straight line segment adjustment value as plane decision variables; For the longitudinal section model, the mileage of the slope change point, the slope of the slope section, and the vertical curve radius are taken as the longitudinal section decision variables; Among them, the shunting time of the train on the difficult line and the height difference between the uncoupling point of the long train group and the peak are used as the longitudinal section objective function, and the total length constraint of the push slope section and the acceleration slope constraint are used as the longitudinal section constraint conditions; The position constraint of the slope change point is used as the constraint condition of the overall horizontal and vertical model; The average shunting time of the train and the total cost of the construction project are used as the objective function of the overall horizontal and vertical model; S2: Input the feasible plane scheme of the hump into the overall plane and longitudinal model, and use the multi-objective particle swarm algorithm to obtain the optimized longitudinal section scheme corresponding to the feasible plane scheme; S3: Combine the optimized longitudinal section scheme with the feasible plane scheme to solve the objective function of the overall plane and longitudinal model, and use the multi-objective particle swarm algorithm to obtain the optimized plane scheme; S4: Based on the optimized plane solution, optimize again to obtain the optimal solution of the corresponding longitudinal section solution and determine whether it converges. If it does not converge, return to S3. If it converges, output the optimized plane and longitudinal overall model; The step S2 specifically includes: S2-1: The position and slope range of the slope change point meet the longitudinal section constraints, and a feasible plane solution is input to obtain the initial particle swarm of the feasible plane solution; S2-2: Generate the initial longitudinal section plan, including: Within the mileage range of the slope change point and the slope range of the longitudinal section constraint condition, the mileage array and the slope array of the slope change point are generated in 100 equal divisions, and the 100 first longitudinal section schemes are formed by correspondingly combining them one by one in the order from small to large; Traverse these 100 first longitudinal section schemes, calculate the satisfaction of the constraint conditions of each scheme, record the lowest peak height scheme and the highest peak height scheme that meet the constraints, and use their slope and slope change point mileage as the new constraint range respectively; Within the new constraint range, the slope section is divided into 50 equal parts to generate 50 second longitudinal section schemes; The corresponding slope sections of the second longitudinal section scheme are randomly replaced to form 50 initial longitudinal section scheme groups; S2-3: After the initial longitudinal section scheme group is determined, all particles are individually optimal, and the global optimal individual is determined by Parato dominance relationship and crowding ranking; iterate according to the particle swarm algorithm calculation formula, calculate the new initial longitudinal section scheme group position and randomly generate any number x between (0,1), calculate the acceptance probability p of the new position, if x∈[0,p], update the scheme position, otherwise do not update, and finally obtain the optimized longitudinal section scheme; The steps to obtain the optimized plane solution include: The plan decision variables of the single-beam hump include: curve angle α, curve radius R and straight segment adjustment value L; the turnout and reducer are determined by the position of the straight segment of the track where they are located. When generating the plan plan, it is necessary to consider the straight line constraint between the last branch turnout and the curve and the line spacing constraint of the track within the beam; Generate a single-beam plane solution, including: taking the straight line segment adjustment value as L=0m, calculating the angle range with an iterative step length of 1°, using the binary method to shorten the step length at the critical point to obtain the critical range of the angle; the angle of the single-beam plane solution is uniformly taken in the critical range of the angle according to the number of particles, and the straight line segment length is uniformly taken in the range of 0 to 50m, and 50 plane solution particles are generated; the required energy height of each braking position is determined, and the maximum reducer energy height is taken to adjust the number of reducer sections at the braking position of the plane solution; according to the connection constraint processing method, the plane solution after adjusting the number of reducer sections at the braking position is constrained, and 20 solutions are randomly selected from the solutions that meet the constraints after the treatment as the initial plane solution for iteration; A decision-making method incorporating the idea of a tournament is adopted. Based on the sorting of the objective function, the evaluation indicators of the non-inferior solutions are compared in turn. The priority of each solution is calculated according to the victory value of each indicator. All non-inferior solutions are sorted in this way, so as to select the optimal plane solution that combines multiple objectives. The calculation process is as follows: Championship Formula Sorting equation t (c) (a,b) is the victory value of the c-th objective function of solution a compared with the corresponding objective function of solution b; Φ is the set of non-inferior solutions; |Φ| is the number of solutions in the set of non-inferior solutions; f (c) (a) and f (c) (b) is the c-th objective function value of scheme a and b; R(a) is the ranking value of scheme a in the non-inferior solution set; μ c is the weight value corresponding to the cth objective function; d is the number of optimized objective functions of the current model; T (c) (a, Φ) is the victory value of the c-th objective function of solution a in the non-inferior solution set Φ compared with the c-th objective functions of other solutions in Φ; The congestion in MOPSO is also taken into consideration; the external archive in MOPSO is retained, and the single congestion sequence is changed to a double sequence of congestion and tournament, and the roulette algorithm is used to establish the probability decision equation of the double sequence: Where: P a is the probability of selecting solution a as the overall optimal solution; R1(a) and R2(a) are the ranking values of solution a in the congestion sequence and the tournament sequence.
2. The method for overall optimization design of the horizontal and vertical surfaces of a railway hump according to claim 1 is characterized in that: The plane model and the longitudinal section model are specifically: The hump is set as a fixed design style before the first branch turnout according to the compactness principle and track matching requirements; the multiple harnesses connected to the shunting line after the first branch turnout are regarded as fixed modules, and the connection units between the turnouts are used as parameters to establish a plane model; The straight line element is determined by the positions of two slope change points, and the circular curve is calculated based on the vertical curve radius at the given slope change point and is used to establish the longitudinal section model; The decision variables of the overall model of the hump's horizontal and vertical planes are expressed as follows: Curve rotation vector α: α=[α1,α2,…,α m ] T ; Curve radius vector R: R = [R1, R2, ..., R m ] T ; Line segment adjustment value vector L: L = [L1, L2, ..., L m ] T ; Slope change point mileage column vector k: k = [k1, k2, ..., k n ] T ; The column vector i of the slope of the next slope section at the slope change point: i=[i1,i2,…,i n ] T ; Vertical curve radius column vector R v : R v =[R v1 ,R v2 ,…,R vn ] T ; Among them, α m is the curve angle, R m is the curve radius, L m is the straight line segment adjustment value, k n is the mileage of the slope change point, i n R is the slope of the next slope section at the slope change point, vn is the vertical curve radius, m is the number of plane connection units, and n is the number of slope change points.
3. The method for overall optimization design of the horizontal and vertical surfaces of a railway hump according to claim 2 is characterized in that: The objective function of the overall planar and longitudinal model is specifically: The hump dismantling capacity is evaluated based on whether the hump can safely and quickly shun the train to the corresponding line; the feasibility of the hump scheme is evaluated based on the hump investment cost; the average shunting time of the train and the total cost of the construction project are used as the objective function of the overall horizontal and vertical model; The objective function expression of average shunting time is: Among them, minF t (α,L,R,k,i,R v ) is the minimum value of the average shunting time function. The overall shunting time of any train is is the average shunting time of the last vehicle in the shunting line; t by is the shunting interval between the yth vehicle and the next vehicle in the train at the peak; v is the number of vehicles in the train; The total construction project cost includes track laying costs, turnout construction costs and speed reducer costs. Track laying costs include the laying costs of rails, sleepers, and roadbeds under sleepers, as well as related power facility costs. Excluding the turnout part, the expression is: C L =μ L (∑L t -SL s ); In the formula, C L is the total cost of track laying on the hump surface; μ L L is the laying cost per linear meter of hump track; t is the length of a single line in the hump plane design; ΣL t is the total length of all lines in the hump plane design; L s The length of the line occupied by a single turnout arranged for the hump; ΣL s The length of the line occupied by all the switches arranged for the hump; The cost of turnout construction is related to the turnout model and is calculated by multiplying the number of turnouts by the construction cost of a single turnout of the corresponding model: C S =Sm Si N i ; C S is the total cost of laying hump turnout; μ Si Cost of laying individual turnouts of different types; N i is the number of turnouts of the corresponding type; The reducer cost is related to the reducer model and the number of sections set, and the expression is: C P =Σμ pi N pi ; C P is the total construction cost of the hump reducer; μ pi The construction cost of a single section of different types of reducers; N pi is the number of reducer sections of the corresponding type; The objective function expression of the total cost of construction project is: minF C (α,L,R,k,i,R v )=C L +C S +C P ; Where minF C (α,L,R,k,i,R v ) represents the minimum value of the total cost of the construction project.
4. The method for overall optimization design of the horizontal and vertical surfaces of railway humps according to claim 3 is characterized in that: The specific longitudinal section objective function is: The time T for a single vehicle to be released to any mileage k m Expressed as: Where: v(k,h) is the function of the vehicle's slipping speed changing with mileage and elevation; The objective function of the shunting time of the train on the difficult line is recorded as: minf t (α,L,R,k,i,R v ) = T; minf t (α,L,R,k,i,R v ) the minimum time for trains to be shunted on difficult lines; T is the total release time; Assume that the average unit length gravity of the train set is q, the unit resistance of the train set is r, and the total length of the train set is l cs ,expression: L c1 qi ph +l cs qr=(l cs -L c1 -L c2 )qi su Find: Where: i ph is the slope of the hook slope; i su is the acceleration slope; l c1 The length of the straight line for the difficult-to-travel vehicle to slip to the insulated sections of the difficult-to-travel line; c2 The length of the straight line between the insulating sections of the easy-to-drive line and the difficult-to-drive adjacent line; Then the height difference Δh between the decoupling point and the peak platform is: Δh = l c1 i ph ; The height difference objective function between the uncoupling point of the long train group and the peak is expressed as: minΔh=f Δh (k,i,R v ) minΔh is the minimum height difference between the uncoupling point of the long train set and the peak; f Δh (k,i,R v ) is the height difference function between the uncoupling point of the long train set and the peak.
5. The method for overall optimization design of the horizontal and vertical surfaces of railway humps according to claim 4 is characterized in that: The specific longitudinal section constraints are: In the longitudinal section, the total length constraint of the push slope section and the acceleration slope constraint are recorded as c V (α,L,R,k,i,R v ); The total length constraint of the push slope section is expressed as: Where: L cs is the train length; j is the length of the jth slope section of the push part; s is the total number of slope sections in the push part; The acceleration slope constraint of the sled section is expressed as: k su +T su +l sa ≤k sbt ; Where: k su T is the mileage of the slope change point between the acceleration slope and the intermediate slope; su l is the length of the vertical curve tangent between the acceleration slope and the intermediate slope; sa k is the safety distance between the vertical curve and the turnout track; sbt The basic track mileage before the first branch turnout; The position of the slope change point in the horizontal and vertical coupling is expressed as a constraint condition: E j +T ver <k j j+1 -T ver ; Where: k j is the mileage of the current slope change point; E j S is the end mileage of the switch or reducer range before the current slope change point; j+1 The starting mileage of the switch or reducer range after the current slope change point; T ver It is the tangent length of the vertical curve at the current slope change point.
6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for overall optimization design of the horizontal and vertical surfaces of railway humps as described in any one of claims 1 to 5 is implemented.
7. An electronic device, characterized in that: include: processor; and a memory for storing executable instructions for the processor; Wherein, the processor is configured to execute the railway hump horizontal and vertical surface overall optimization design method as described in any one of claims 1-5 by executing the executable instructions.
