A rack rail car multi-drive unit traction force self-balancing torque distribution method and system

Abstract

This invention relates to the field of control technology for mining transportation equipment, and particularly to a method and system for condition perception and self-balancing traction force distribution in a multi-drive system of a rack railcar. The method includes acquiring the environmental status of the section ahead of the rack railcar and its current real-time position. This invention significantly improves the feedforward response speed of traction force distribution and shortens system control delay through environmental perception and suggestion correction of the section ahead; it effectively improves the traction force coordination consistency among multiple drive units by fusing local allocation weights with suggestion correction results; it significantly improves the adhesion utilization rate of the rack rail meshing surface and suppresses single-wheel slippage through section environment vector-driven adhesion coefficient feedforward correction; and it effectively balances the load of each drive unit through weighted collaborative game and dynamic redistribution mechanisms, avoiding load concentration and mechanical impact. Overall, it enhances the operational stability, control accuracy, and traffic safety of the rack railcar under complex roadway conditions such as slope changes, slippery conditions, and contamination.

Description

Technical Field

[0001] This invention relates to the field of control technology for mining transportation equipment, and in particular to a method and system for condition sensing and self-balancing traction force distribution in a multi-drive system of a rack railcar. Background Technology

[0002] The rack railcar is a key piece of equipment for underground mine roadway transportation. Its operational safety and transportation efficiency are highly dependent on the coordinated control performance of the multi-drive system. Existing rack railcars mostly adopt fixed ratio, master-slave synchronization or centralized command torque distribution strategies, which are difficult to respond in a timely manner to the impact of typical underground working conditions such as continuous track slope changes, local wet and slippery conditions or coal slurry contamination, rack rail surface deterioration, load changes and communication delays on the traction capacity of each drive unit. This can easily lead to problems such as uneven torque distribution among drive units, increased rack rail meshing impact, overheating of local drive units or protective derating.

[0003] At the feedforward control level, existing systems can usually only make closed-loop corrections based on the current state of the vehicle. They lack advance perception and reliability judgment of environmental conditions such as slopes, curves, wetness, pollution, deterioration of the toothed rail surface, and obstacle risks in the target section ahead. They cannot complete the pre-distribution of traction force before the vehicle enters the unfavorable section, which makes it easy for instantaneous load transfer and torque shock to occur at the working condition switching node.

[0004] At the level of allocation constraints, traditional algorithms generally rely on static thresholds and preset parameters, and do not uniformly incorporate the predicted adhesion capability of the target section, the local torque allocation results, and the proposed correction results of the preceding section into the equilibrium solution. When the proposed data times out, lacks credibility, or the target section is inconsistent, there is also a lack of weight reduction or cancellation mechanism, which can easily lead to unreliable proposals being directly applied to torque allocation, affecting the stability of on-site operation.

[0005] The above-mentioned problems are common in existing technologies, for example:

[0006] Although Chinese patent CN223279096U, "Electric Gear Rail Car", discloses an integrated electric drive structure and basic braking energy recovery function, its torque distribution still adopts a master-slave synchronous mode and does not set up a drive unit-level adhesion constraint and feedforward correction mechanism.

[0007] Chinese patent CN223014613U, "Power Unit and Gear Rail Cart", focuses on the explosion-proof structure and thermal management design of a single power unit. Its control logic does not involve the dynamic game and equilibrium solution of traction forces between multiple units.

[0008] Although Chinese patent CN119370129A, "Electrical Control System for Explosion-proof Lithium Battery Gear Rail Car and Explosion-proof Lithium Battery Gear Rail Car", proposes a distributed controller architecture, the actual control strategy is still that each drive unit responds independently to the upper-level command, and no traction force self-balancing distribution model based on the fusion of front environment perception and local state has been established. Summary of the Invention

[0009] (a) Technical problems to be solved

[0010] To address the key technical problems in existing multi-drive systems for rack railcars, such as unpredictable forward conditions, lack of adhesion constraints in traction force distribution, and inability to dynamically rebalance torque after amplitude limiting, this invention provides a self-balancing torque distribution method and system for traction force of multi-drive units in rack railcars.

[0011] (II) Technical Solution

[0012] This invention is achieved through the following technical solution: This invention proposes a method for self-balancing torque distribution of traction force in multiple drive units of a rack railcar, comprising the following steps:

[0013] S100: Obtain the environmental status of the section ahead of the rack railcar and the current real-time position of the rack railcar, identify the target section that the rack railcar is about to enter, and generate a section adaptive weight correction amount that characterizes the difference in risk exposure of each drive unit in the target section.

[0014] The adaptive weight correction amount for the section is generated by the edge computing nodes (EMUs) deployed in the roadway;

[0015] The environmental status data of the preceding section collected by the edge computing node EMU includes: the slope status and curve status obtained by matching the basic data of the line, as well as the wet and slippery status, pollution status, toothed rail surface status and obstacle risk status obtained by real-time sensing.

[0016] The segment adaptive weight correction amount includes the generation time, target segment number, effective time window, and confidence identifier;

[0017] The edge computing node (EMU) performs structured quantization of the environmental state of the preceding segment to obtain the segment environmental vector:

[0018] :

[0019] In the formula, For the first The target segment in the first The segmental environmental vector for each acquisition cycle, As an indicator of the unfavorable degree of the slope, As an indicator of the degree of unfavorability of the curve, This is an indicator of slippery condition. As a pollution status indicator, This is an indicator of the surface deterioration of the gear track. For obstacle risk indicators;

[0020] All the above indicators are normalized to the [0, 1] interval, and a comprehensive risk index is calculated based on the environmental vector of the aforementioned segment:

[0021]

[0022] In the formula, For the first The target segment in the first Comprehensive risk indicators for each data collection cycle , , , , , These are the weighting coefficients for the risks of ramps, curves, slipperiness, contamination, gear track surface deterioration, and obstacles, respectively. And all of the above weight coefficients are ≥0;

[0023] Based on the spatial configuration of the rack railcar within the target section and the installation position of each drive unit, the exposure coefficient of each drive unit relative to the target section is determined, and the exposure coefficient is normalized to [0, 1]. The suggested correction amount for each drive unit is then calculated.

[0024]

[0025] In the formula,

[0026] ; For the number of drive units, For the first The driving unit is relative to the first The first target segment Exposure coefficient for each collection cycle, The average exposure coefficient. For the first The first drive unit in the... Suggested correction amount for each data collection cycle To suggest a correction to the strength coefficient, This serves as a credibility identifier for the segment's environmental vector.

[0027] The segment adaptive weight correction amount includes the proposed correction amount;

[0028] S200: Determine the local allocation weight based on the local target output torque of each drive unit in the current control cycle;

[0029] The local allocation weights are determined as follows:

[0030]

[0031] In the formula, For the first The first drive unit in the... Locally allocated weights within each control cycle The local target output torque of the i-th drive unit in the k-th control cycle is... Let the local target output torque of the m-th drive unit be the value in the k-th control cycle; when the denominator is zero, let... , This refers to the number of drive units;

[0032] S300: Perform a validity check on the adaptive weight correction amount of the section. If the check passes, merge it with the local allocation weight to obtain the traction force suggested allocation weight.

[0033] The proposed weighting of traction force is obtained through the following method:

[0034] Calculate the intermediate weights: ;

[0035] Normalize all intermediate weights:

[0036]

[0037] In the formula, As the intermediate weight, It is recommended to assign weights to traction forces. The valid correction component after validity verification. The intermediate weight of the m-th drive unit in the k-th control cycle;

[0038] When the sum of all intermediate weights is zero, or ;

[0039] The validity check includes:

[0040] Determine whether the time difference between the current moment and the moment when the segment adaptive weight correction is generated is not greater than the preset maximum effective time window;

[0041] Determine whether the confidence indicator attached to the adaptive weight correction amount of the segment is not lower than a preset confidence threshold;

[0042] If any condition is not met, cancel the adaptive weight correction of the segment or apply a decay factor to it for weight reduction.

[0043] S400: Based on the target segment environment vector, the local adhesion coefficient of each driving unit is feedforward corrected to obtain the predicted adhesion coefficient;

[0044] The predicted adhesion coefficient is obtained in the following way:

[0045] Extracting the wet and slippery state of the target section Pollution status and the degree of unfavorable slope The three indicators;

[0046] Calculate the adhesion correction factor:

[0047]

[0048] In the formula, The adhesion correction factor for the target section. The minimum adhesion retention factor, , , The weighting coefficients are respectively for wet and slippery conditions, contaminated conditions, and unfavorable slope conditions. All weighting coefficients are non-negative and their sum is 1. The adhesion correction coefficient is not less than the minimum adhesion retention coefficient and not greater than 1, wherein the minimum adhesion retention coefficient is a preset constant that is greater than 0 and less than 1.

[0049] Multiply the local adhesion coefficient of each drive unit by the adhesion correction coefficient respectively:

[0050]

[0051] In the formula, For the first The predicted adhesion coefficient of the driving unit in the k-th control cycle For the first The local adhesion coefficient of the driving unit in the kth control cycle

[0052] The local adhesion coefficient of each drive unit characterizes the ratio of the tangential force to the normal force of the gear teeth at the gear meshing point of the drive unit. Its value is equal to the ratio of the output torque of the drive wheel of the drive unit to the product of the drive wheel radius and the normal force of the gear tooth contact.

[0053] S500: Using the local allocation weight, the traction force suggested allocation weight, and the predicted adhesion coefficient as inputs, construct a weighted collaborative game model and solve for the equilibrium allocation weight that satisfies the constraints of total weight normalization, non-negativity, and adhesion ability.

[0054] The equilibrium solution of the weighted collaborative game model satisfies the following three constraints:

[0055] ;

[0056] ;

[0057]

[0058] ; The total traction torque requirement for the rack rail chock vehicle. To allocate weights to achieve the equilibrium;

[0059]

[0060] ; The normal force at the tooth contact point of the i-th drive unit during the k-th control cycle. Let the radius of the drive wheel corresponding to the i-th drive unit be . The average intermediate weight of each drive unit during the k-th control cycle. This represents the maximum traction torque of the i-th drive unit of the rack railcar during the k-th control cycle.

[0061] S600: Generate the feedforward target torque according to the balanced allocation weight, and after limiting and torque redistribution, output the feedforward pre-allocated torque command vector of each drive unit.

[0062] The torque redistribution includes:

[0063] The amount of torque reduction caused by the limiting is determined as the torque difference to be redistributed;

[0064] Select the remaining drive units that still have a safe output margin, excluding the over-limit drive units;

[0065] The torque difference is allocated according to the proportion of the remaining torque margin of each of the other drive units;

[0066] The allocation result is superimposed on the initial limiting command of the corresponding drive unit to form the final feedforward pre-allocated torque command vector;

[0067] S700: When the sum of the remaining available capacity of all drive units is insufficient to cover the traction limit difference, output a traction demand reduction command or a safety risk warning.

[0068] On the other hand, the present invention also provides a self-balancing torque distribution system for traction force of multiple drive units of a rack railcar, used to execute the self-balancing torque distribution method for traction force of multiple drive units of a rack railcar described in any of the above claims: including: a section adaptive weight correction amount generation module, configured to collect the environmental status and current real-time position of the section in front of the rack railcar, identify the target section, and generate a section adaptive weight correction amount characterizing the difference in risk exposure of each drive unit in the target section; the section adaptive weight correction amount includes the generation time, target section number, effective time window, and confidence indicator;

[0069] The local allocation weight determination module is configured to receive the local target output torque of each drive unit in the current control cycle, calculate and output the local allocation weight;

[0070] The traction force suggestion allocation weight determination module is configured to receive the segment adaptive weight correction amount and the local allocation weight, perform validity verification, and merge the two after the verification is passed to output the traction force suggestion allocation weight.

[0071] The predicted adhesion coefficient acquisition module is configured to receive the target segment environmental vector and the local adhesion coefficient of each driving unit, perform feedforward correction, and output the predicted adhesion coefficient.

[0072] The equilibrium weight allocation acquisition module is configured to receive the local allocation weight, the traction force suggested allocation weight and the predicted adhesion coefficient, construct a weighted collaborative game model, and solve for the equilibrium allocation weight that satisfies the constraints of total weight normalization, non-negativity and adhesion ability.

[0073] The feedforward pre-allocation torque command generation module is configured to receive the equalization allocation weight, generate the feedforward target torque, perform amplitude limiting processing and torque redistribution, and output a traction demand reduction command or safety risk warning when the sum of the remaining available capacity of all drive units is insufficient to bear the amplitude limiting difference; the feedforward pre-allocation torque command generation module outputs the feedforward pre-allocation torque command vector of each drive unit.

[0074] The output of the segment adaptive weight correction generation module is communicatively connected to the first input of the traction force suggestion allocation weight determination module.

[0075] The output of the local allocation weight determination module is communicatively connected to the second input of the traction force suggestion allocation weight determination module.

[0076] The output of the traction force suggestion allocation weight determination module is communicatively connected to the first input of the balanced allocation weight acquisition module;

[0077] The output of the predicted adhesion coefficient acquisition module is communicatively connected to the second input of the balanced weight allocation acquisition module.

[0078] The output of the balanced weight allocation acquisition module is communicatively connected to the input of the feedforward pre-allocated torque command generation module.

[0079] On the other hand, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the self-balancing torque distribution method for traction force of the multi-drive unit of the rack railcar described above.

[0080] On the other hand, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for the self-balancing torque distribution of traction force of multiple drive units in a rack railcar.

[0081] (III) Beneficial Effects

[0082] This invention significantly improves the feedforward response speed of traction distribution and shortens system control delay by perceiving and suggesting corrections of the environment in the forward section.

[0083] By integrating locally assigned weights with suggested correction results, the traction coordination consistency among multiple drive units is effectively improved.

[0084] By using segment environment vector-driven adhesion coefficient feedforward correction, the adhesion utilization rate of the toothed track meshing surface is significantly improved, and single wheel slippage is suppressed.

[0085] Through weighted collaborative game and dynamic redistribution mechanism, the load of each drive unit is effectively balanced to avoid load concentration and mechanical shock.

[0086] This enhances the overall stability, control precision, and traffic safety of the rack railcar under complex roadway conditions such as slope changes, wetness, and pollution. Attached Figure Description

[0087] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0088] Figure 1 This is a schematic flowchart of the self-balancing torque distribution method described in this invention.

[0089] Figure 2 This is a framework diagram of the self-balancing torque distribution system described in this invention. Detailed Implementation

[0090] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure; rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0091] The following method is executed as a feedforward pre-allocation method based on vehicle-side collaboration and target section environment prediction. Its processing objects are the local target output torque output by the vehicle-side local controller and the forward section suggestion correction results generated by the edge computing node EMU. The following method does not take the online parameter identification of the original vehicle monitoring signals such as original motor current, voltage, and vibration as a necessary step, so as to distinguish it from the vehicle local closed-loop dynamic control method.

[0092] The K collection period below refers to the collection period number of the edge computing node;

[0093] The control cycle k below refers to the control cycle number of the on-board controller;

[0094] That is, refer to Figure 1 The present invention proposes a self-balancing torque distribution method for the traction force of multiple drive units in a rack railcar, comprising the following steps:

[0095] S100: Obtain the environmental status of the section ahead of the rack railcar and the current real-time position of the rack railcar, identify the target section that the rack railcar is about to enter, and generate a section adaptive weight correction amount that characterizes the difference in risk exposure of each drive unit in the target section.

[0096] The adaptive weight correction amount for the section is generated by the edge computing nodes (EMUs) deployed in the roadway;

[0097] The edge computing nodes (EMUs) are deployed at key locations such as roadway slope change points, curve entrances, intersections, and areas with high incidence of wet or slippery conditions or coal slime pollution. They are used to periodically collect or update the slope status, curve status, wet and slippery conditions, pollution status, rack surface status, and obstacle risk status of the covered sections, and generate corresponding environmental status data of the section in front of the rack railcar.

[0098] The environmental status data of the preceding section collected by the edge computing node EMU includes: the slope status and curve status obtained by matching the basic data of the line, as well as the wet and slippery status, pollution status, toothed rail surface status and obstacle risk status obtained by real-time sensing.

[0099] The segment adaptive weight correction amount includes the generation time, target segment number, effective time window, and confidence identifier;

[0100] The edge computing node (EMU) performs structured quantization of the environmental state of the preceding segment to obtain the segment environmental vector:

[0101] :

[0102] In the formula, For the first The target segment in the first The segmental environmental vector for each acquisition cycle, As an indicator of the unfavorable degree of the slope, As an indicator of the degree of unfavorability of the curve, This is an indicator of slippery condition. As a pollution status indicator, This is an indicator of the surface deterioration of the gear track. For obstacle risk indicators;

[0103] All the above indicators are normalized to the [0, 1] interval, and a comprehensive risk index is calculated based on the environmental vector of the aforementioned segment:

[0104]

[0105] In the formula, For the first The target segment in the first Comprehensive risk indicators for each data collection cycle , , , , , These are the weighting coefficients for the risks of ramps, curves, slipperiness, contamination, gear track surface deterioration, and obstacles, respectively. And all of the above weight coefficients are ≥0;

[0106] Based on the spatial configuration of the rack railcar within the target section and the installation position of each drive unit, the exposure coefficient of each drive unit relative to the target section is determined, and the exposure coefficient is normalized to [0, 1]. The suggested correction amount for each drive unit is then calculated.

[0107]

[0108] In the formula,

[0109] ; For the number of drive units, For the first The driving unit is relative to the first The first target segment Exposure coefficient for each collection cycle, The average exposure coefficient. For the first The first drive unit in the... Suggested correction amount for each data collection cycle To suggest a correction to the strength coefficient, This serves as a credibility identifier for the segment's environmental vector.

[0110] The segment adaptive weight correction amount includes the proposed correction amount;

[0111] S200: Determine the local allocation weight based on the local target output torque of each drive unit in the current control cycle; the local target output torque is the current segment torque result formed by the vehicle local closed-loop control or the previous level traction control, which is used to reflect the current driving state of the vehicle.

[0112] The local allocation weights are determined as follows:

[0113]

[0114] In the formula, For the first The first drive unit in the... Locally allocated weights within each control cycle The local target output torque of the i-th drive unit in the k-th control cycle is... Let the local target output torque of the m-th drive unit be the value in the k-th control cycle; when the denominator is zero, let... , This refers to the number of drive units;

[0115] S300: Perform a validity check on the adaptive weight correction amount of the section. If the check passes, merge it with the local allocation weight to obtain the traction force suggested allocation weight.

[0116] The proposed weighting of traction force is obtained through the following method:

[0117] Calculate the intermediate weights: ;

[0118] Normalize all intermediate weights:

[0119]

[0120] In the formula, As the intermediate weight, It is recommended to assign weights to traction forces. The valid correction component after validity verification. The intermediate weight of the m-th drive unit in the k-th control cycle;

[0121] When the sum of all intermediate weights is zero, or ;

[0122] The validity check includes:

[0123] I. Determine the current time of execution S200. Generation time of segment adaptive weight correction Is the time difference no greater than the preset maximum effective time window? ;

[0124] II. Determine whether the confidence indicator attached to the adaptive weight correction amount of the segment is not lower than a preset confidence threshold. The credibility identifier is a quantitative value ranging from 0 to 1, determined based on the data quality of the environmental conditions of the section in front of the rack railcar, the communication quality between the vehicle and the information provider, and the consistency between the suggestions and the actual execution results under similar historical conditions.

[0125] III. If any condition is not met, cancel the adaptive weight correction amount of the segment or apply a decay factor to it for weight reduction.

[0126] Conditions I and II specifically include:

[0127] like The vehicle-side local controller (vehicle-side control system) determines the suggested correction results of the weight allocation for each drive unit received from the EMU. Effective, among which, For dynamic analysis of coordinated control of rack rail chocks and railcars. The timestamp corresponding to the collaborative suggestion result (i.e., the suggestion correction result). To suggest the maximum effective time window, For the first The credibility identifier corresponding to each target segment The threshold for recommending adoption;

[0128] The trustworthiness identifier provided by the edge computing node EMU :

[0129]

[0130] In the formula, ,in, The environmental perception data quality coefficient for the target segment is used to characterize the current environmental vector of the target segment. The integrity and effectiveness; The communication state quality coefficient between the rack railcar and the EMU is used to characterize the link latency, packet loss rate and connection stability between the rack railcar and the EMU. The historical recommendation consistency coefficient is used to characterize the degree of consistency between EMU recommendations and the actual execution results of the rack rail clamping vehicle under similar historical working conditions. Let the weighting coefficients satisfy: , , The reliability of the proposed results is determined by the EMU based on a comprehensive assessment of the current section's environmental perception data quality, the communication status of the rack railcar, and the historical implementation results of the proposed solutions. Specifically, the more complete the environmental perception data, the more stable the communication status, and the more consistent the proposed solutions are with the actual implementation results under similar historical conditions, the higher the reliability. The larger the value, the better; conversely, if there are missing sensor data, unstable communication, or significant deviations in historical recommendations, then... The smaller;

[0131] S400: Based on the target segment environment vector, the local adhesion coefficient of each driving unit is feedforward corrected to obtain the predicted adhesion coefficient; the predicted adhesion coefficient characterizes the expected available adhesion capability of each driving unit in the target segment, and is used to distinguish it from the local adhesion estimate in the current segment.

[0132] The predicted adhesion coefficient is obtained in the following way:

[0133] Extracting the wet and slippery state of the target section Pollution status and the degree of unfavorable slope The three indicators;

[0134] Calculate the adhesion correction factor:

[0135]

[0136] In the formula, The adhesion correction factor for the target section. The minimum adhesion retention factor, , , The weighting coefficients are respectively for wet and slippery conditions, contaminated conditions, and unfavorable slope conditions. All weighting coefficients are non-negative and their sum is 1. The adhesion correction coefficient is not less than the minimum adhesion retention coefficient and not greater than 1, wherein the minimum adhesion retention coefficient is a preset constant that is greater than 0 and less than 1.

[0137] Multiply the local adhesion coefficient of each drive unit by the adhesion correction coefficient respectively:

[0138]

[0139] In the formula, For the first The predicted adhesion coefficient of the driving unit in the k-th control cycle For the first The local adhesion coefficient of each drive unit in the kth control cycle;

[0140] The local adhesion coefficient of each drive unit characterizes the ratio of the tangential force to the normal force of the gear teeth at the gear meshing point of the drive unit. Its value is equal to the ratio of the output torque of the drive wheel of the drive unit to the product of the drive wheel radius and the normal force of the gear tooth contact.

[0141] S500: Using the local allocation weight, the traction force suggested allocation weight, and the predicted adhesion coefficient as inputs, construct a weighted collaborative game model and solve for the equilibrium allocation weight that satisfies the constraints of total weight normalization, non-negativity, and adhesion ability.

[0142] The equilibrium solution of the weighted collaborative game model satisfies the following three constraints: First:

[0143]

[0144] Second item:

[0145]

[0146] Third item:

[0147]

[0148] In the formula, The total traction torque requirement for the rack rail chock vehicle. To allocate weights to achieve the equilibrium;

[0149]

[0150] ; The normal force at the tooth contact point of the i-th drive unit during the k-th control cycle. Let the radius of the drive wheel corresponding to the i-th drive unit be . The average intermediate weight of each drive unit during the k-th control cycle. This represents the maximum traction torque of the i-th drive unit of the rack railcar during the k-th control cycle. It is the upper limit of feedforward torque obtained by predicting and correcting the adhesion capability based on the defined current maximum allowable torque and the previous working conditions.

[0151] The above The action path is clear: it first corrects the predicted adhesion capability of each driving unit facing the forward section, and then affects the available torque boundary in the subsequent game equilibrium.

[0152] Furthermore, the equilibrium solution of the weighted cooperative game is equivalent to solving the weighted cost minimization problem with linear constraints mentioned above. That is, under the condition of satisfying the three constraints, the weight allocation vector that minimizes the cost function is the equilibrium allocation weight of the i-th driving unit in the k-th control cycle. Since the objective function is quadratic and the constraints are linear, the problem can be transformed into a quadratic programming problem. The equilibrium allocation weight can be obtained in real time through KKT conditions or a standard quadratic programming solver.

[0153] The first constraint is a local consistency term, used to ensure that the equilibrium result does not deviate from the current allocation trend already formed by the vehicle-side local closed-loop control or the previous level traction control; the second is a suggestion adoption term, used to reflect the feedforward correction effect of the EMU suggestion on the vehicle-side allocation, and its influence strength is weighted by credibility; the third is a feedforward adhesion penalty term, used to suppress the excessive traction ratio assigned to drive units with poor predicted adhesion conditions before entering the target section. All weight coefficients are non-negative.

[0154] S600: Generate the feedforward target torque according to the balanced allocation weight, and after limiting and torque redistribution, output the feedforward pre-allocated torque command vector of each drive unit.

[0155] The torque redistribution includes:

[0156] The amount of torque reduction caused by the limiting is determined as the torque difference to be redistributed;

[0157] Select the remaining drive units that still have a safe output margin, excluding the over-limit drive units;

[0158] The torque difference is allocated according to the proportion of the remaining torque margin of each of the other drive units;

[0159] The allocation result is superimposed on the initial limiting command of the corresponding drive unit to form the final feedforward pre-allocated torque command vector;

[0160] Specific implementations include:

[0161] Based on the balanced allocation of weights among the driving units Generate the feedforward target torque of each drive unit before it enters the target section. ;

[0162] The feedforward target torque before the i-th drive unit enters the target segment is: :

[0163] , For the first The weights are evenly distributed across each driving unit.

[0164] This indicates the total traction torque required by the rack rail chock to overcome the track gradient and running resistance, and to meet the traction requirements given by the driver within the current control cycle.

[0165] If the target feedforward torque is greater than the preset upper limit threshold, the feedforward pre-allocated torque command vector of the corresponding drive unit will be... The amplitude limit is set to a preset upper limit threshold;

[0166] Specifically, if the Each drive unit satisfies Then its amplitude is limited as follows: ;

[0167] The unachieved torque difference is determined based on the feedforward target torque and the preset upper limit threshold.

[0168] Specifically, define the first The torque difference that the drive unit failed to complete is :

[0169] And redistribute according to the remaining capacity of the remaining drive units;

[0170] Determine the remaining capacity of the remaining drive units that meet the preset conditions;

[0171] Specifically, the drive units other than the first drive unit are referred to as the remaining units. For the remaining units that have not yet reached their own safe output limit and are not in the protection exit state, the remaining capacity is determined according to their rated capacity, the currently allocated torque and the protection state. If the vehicle-side local controller has provided a health or available capacity coefficient, the coefficient can be used as a derating factor in the calculation of the remaining capacity to participate in the allocation.

[0172] S700: When the sum of the remaining available capacity of all drive units is insufficient to cover the limiting difference, output a traction demand reduction command or a safety risk warning.

[0173] Specifically:

[0174] Based on the remaining capacity and the unfulfilled torque difference, obtain the compensation torque borne by the remaining drive units and the updated feedforward pre-allocated torque command vector and execute it;

[0175] Specifically, no. The compensation torque borne by the remaining units within each control cycle is : And update its final feedforward output instruction as follows:

[0176] ; T pre,out =[ T 1 pre,out , T 2 pre,out ,…, T n pre,out ] This vector serves as the pre-allocated torque command before the rack railcar enters the target section. After the rack railcar enters the section, it continues to be dynamically corrected by local feedback and local self-balancing mechanism, thus forming a multi-drive collaborative control closed loop that combines feedforward and feedback.

[0177] This embodiment enables the rack railcar to obtain forward working condition suggestions before entering the target section by deploying edge computing nodes (EMUs) in the roadway and establishing communication with the vehicle-end controller. The EMU does not directly issue torque control commands, nor does it replace the vehicle-end local closed-loop control. Instead, it provides feedforward suggestions with the target section number, generation time, effective time window, and confidence index, which the vehicle-end controller uses to pre-allocate traction force.

[0178] Compared with the existing centralized fixed-proportion torque distribution scheme, this invention incorporates the vehicle-side local target output torque, target section environmental suggestions, and predicted adhesion capability into the weighting solution. Positioning and communication can be achieved using one or more combinations of UWB, inertial navigation, encoder odometer, and mine-use wireless or wired communication, without making a specific communication standard a necessary limitation, thus better meeting the feasible conditions of underground sites.

[0179] The above method constitutes a complete feedforward cooperative control chain: Through forward section environmental perception and real-time position identification of the rack railcar, suggested correction results for the drive unit allocation weights corresponding to each drive unit within the target section are generated. Combined with the local allocation weights determined by the local target output torque, suggested traction force allocation weights are formed. The local adhesion coefficient is feedforward corrected using the section environmental vector to obtain the predicted adhesion coefficient. A weighted cooperative game model is constructed using the local allocation weights, suggested traction force allocation weights, and predicted adhesion coefficients as inputs. The equilibrium allocation weights for each drive unit are obtained through equilibrium solution. A feedforward target torque is generated based on the equilibrium allocation weights, and amplitude limiting is performed. The amplitude limiting difference is dynamically redistributed based on the available capacity of each drive unit, ultimately generating a feedforward pre-allocated torque command vector. This method implements feedforward control on a section-by-section basis, using adhesion capacity as a hard constraint, cooperative game as an optimization mechanism, and dynamic redistribution as a safety net, thereby improving the balance of traction force allocation, adhesion utilization, and cooperative consistency of multiple drive units under complex roadway conditions.

[0180] In the above embodiments, the local target output torque of each drive unit is not directly given by the EMU, but is formed by the on-board local closed-loop control or the previous level traction control within the current control cycle; this method performs feedforward pre-allocation correction for the target section based on the result.

[0181] On the other hand, refer to Figure 2 As shown, the present invention also provides a self-balancing torque distribution system for traction force of multiple drive units of a rack railcar, used to execute the self-balancing torque distribution method for traction force of multiple drive units of a rack railcar described in any of the above claims: including: a section adaptive weight correction amount generation module, configured to collect the environmental status and current real-time position of the section in front of the rack railcar, identify the target section, and generate a section adaptive weight correction amount characterizing the difference in risk exposure of each drive unit in the target section; the section adaptive weight correction amount includes the generation time, target section number, effective time window, and confidence indicator;

[0182] The local allocation weight determination module is configured to receive the local target output torque of each drive unit in the current control cycle, calculate and output the local allocation weight;

[0183] The traction force suggestion allocation weight determination module is configured to receive the segment adaptive weight correction amount and the local allocation weight, perform validity verification, and merge the two after the verification is passed to output the traction force suggestion allocation weight.

[0184] The predicted adhesion coefficient acquisition module is configured to receive the target segment environmental vector and the local adhesion coefficient of each driving unit, perform feedforward correction, and output the predicted adhesion coefficient.

[0185] The equilibrium weight allocation acquisition module is configured to receive the local allocation weight, the traction force suggested allocation weight and the predicted adhesion coefficient, construct a weighted collaborative game model, and solve for the equilibrium allocation weight that satisfies the constraints of total weight normalization, non-negativity and adhesion ability.

[0186] The feedforward pre-allocation torque command generation module is configured to receive the equalization allocation weight, generate the feedforward target torque, perform amplitude limiting processing and torque redistribution, and output a traction demand reduction command or safety risk warning when the sum of the remaining available capacity of all drive units is insufficient to bear the amplitude limiting difference; the feedforward pre-allocation torque command generation module outputs the feedforward pre-allocation torque command vector of each drive unit.

[0187] The output of the segment adaptive weight correction generation module is communicatively connected to the first input of the traction force suggestion allocation weight determination module.

[0188] The output of the local allocation weight determination module is communicatively connected to the second input of the traction force suggestion allocation weight determination module.

[0189] The output of the traction force suggestion allocation weight determination module is communicatively connected to the first input of the balanced allocation weight acquisition module;

[0190] The output of the predicted adhesion coefficient acquisition module is communicatively connected to the second input of the balanced weight allocation acquisition module.

[0191] The output of the balanced weight allocation acquisition module is communicatively connected to the input of the feedforward pre-allocated torque command generation module.

[0192] Based on the same concept, corresponding to any of the above embodiments, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements any of the above-described methods for the self-balancing torque distribution of traction force of the multi-drive unit of the rack railcar, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0193] Based on the same concept, corresponding to any of the above embodiments, the present invention also provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it implements any of the above-described self-balancing torque distribution methods for the traction force of the multi-drive unit of the rack railcar, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0194] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of this application is limited to these examples; under the concept of this application, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of the embodiments of this application as described above, which are not provided in detail for the sake of brevity.

[0195] Additionally, to simplify the description and discussion, and to avoid obscuring the embodiments of this application, the well-known power / ground connections to integrated circuit (IC) chips and other components may or may not be shown in the provided drawings. Furthermore, the apparatus may be shown in block diagram form to avoid obscuring the embodiments of this application, and this also takes into account the fact that the details of the implementation of these block diagram apparatuses are highly dependent on the platform on which the embodiments of this application will be implemented (i.e., these details should be fully understood by those skilled in the art). While specific details (e.g., circuits) have been set forth to describe exemplary embodiments of this application, it will be apparent to those skilled in the art that the embodiments of this application can be implemented without these specific details or with variations thereof. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0196] Although this application has been described in conjunction with specific embodiments thereof, many substitutions, modifications, and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed.

[0197] The embodiments of this application are intended to cover all such substitutions, modifications, and variations that fall within the broad scope of this application. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the embodiments of this application should be included within the protection scope of this application.

Claims

1. A method for self-balancing torque distribution of traction force among multiple drive units of a rack railcar, characterized in that: Includes the following steps: The system acquires the environmental status of the section ahead of the rack railcar and the current real-time position of the rack railcar, identifies the target section that the rack railcar is about to enter, and generates a section adaptive weight correction amount that characterizes the difference in risk exposure of each drive unit in the target section. The local allocation weight is determined based on the local target output torque of each drive unit in the current control cycle; The effectiveness of the adaptive weight correction for the segment is verified. If the verification is successful, it is merged with the locally allocated weight to obtain the traction force suggested allocation weight. Based on the target segment environment vector, the local adhesion coefficient of each driving unit is feedforward corrected to obtain the predicted adhesion coefficient; Using the local allocation weight, the traction suggested allocation weight, and the predicted adhesion coefficient as inputs, a weighted collaborative game model is constructed, and the equilibrium allocation weight that satisfies the constraints of total weight normalization, non-negativity, and adhesion ability is solved. The feedforward target torque is generated based on the balanced allocation weight, and after amplitude limiting and torque redistribution, the feedforward pre-allocated torque command vector of each drive unit is output. When the sum of the remaining available capacity of all drive units is insufficient to cover the traction limit difference, a traction demand reduction command or a safety risk warning will be output.

2. The self-balancing torque distribution method for traction force of multiple drive units in a rack railcar according to claim 1, characterized in that: The adaptive weight correction amount for the section is generated by the edge computing nodes (EMUs) deployed in the roadway; The environmental status data of the preceding section collected by the edge computing node EMU includes: the slope status and curve status obtained by matching the basic data of the line, as well as the wet and slippery status, pollution status, toothed rail surface status and obstacle risk status obtained by real-time sensing. The segment adaptive weight correction amount includes the generation time, target segment number, effective time window, and confidence index.

3. The self-balancing torque distribution method for traction force of multiple drive units in a rack railcar according to claim 2, characterized in that: The edge computing node (EMU) performs structured quantization of the environmental state of the preceding segment to obtain the segment environmental vector: ： In the formula, For the first The target segment in the first The segmental environmental vector for each acquisition cycle, As an indicator of the unfavorable degree of the slope, As an indicator of the degree of unfavorability of the curve, This is an indicator of slippery condition. As a pollution status indicator, This is an indicator of the surface deterioration of the gear track. For obstacle risk indicators; All the above indicators are normalized to the [0, 1] interval, and a comprehensive risk index is calculated based on the environmental vector of the aforementioned segment: ； In the formula, For the first The target segment in the first Comprehensive risk indicators for each data collection cycle. , , , , , These are the weighting coefficients for the risks of ramps, curves, slipperiness, contamination, gear track surface deterioration, and obstacles, respectively. And all of the above weight coefficients are ≥0; Based on the spatial configuration of the rack railcar within the target section and the installation position of each drive unit, the exposure coefficient of each drive unit relative to the target section is determined, and the exposure coefficient is normalized to [0, 1]. The suggested correction amount for each drive unit is then calculated. ； In the formula, ; For the number of drive units, For the first The driving unit is relative to the first The first target segment Exposure coefficient for each collection cycle, The average exposure coefficient. For the first The first drive unit in the... Suggested correction amount for each data collection cycle To suggest a correction to the strength coefficient, This serves as a credibility identifier for the segment's environmental vector. The segment adaptive weight correction amount includes the proposed correction amount.

4. The self-balancing torque distribution method for traction force of multiple drive units in a rack railcar according to claim 1, characterized in that: The validity check includes: Determine whether the time difference between the current moment and the moment when the segment adaptive weight correction is generated is not greater than the preset maximum effective time window; Determine whether the confidence indicator attached to the adaptive weight correction amount of the segment is not lower than a preset confidence threshold; If any condition is not met, cancel the adaptive weight correction of the segment or apply a decay factor to reduce its weight.

5. The self-balancing torque distribution method for traction force of multiple drive units in a rack railcar according to claim 1, characterized in that: The local allocation weights are determined as follows: ； In the formula, For the first The first drive unit in the... Locally allocated weights within each control cycle The local target output torque of the i-th drive unit in the k-th control cycle is... Let the local target output torque of the m-th drive unit be the value in the k-th control cycle; when the denominator is zero, let... , This refers to the number of drive units; The proposed weighting of traction force is obtained through the following method: Calculate the intermediate weights: ; Normalize all intermediate weights: ; In the formula, As the intermediate weight, It is recommended to assign weights to traction forces. The valid correction component after validity verification. The intermediate weight of the m-th drive unit in the k-th control cycle; When the sum of all intermediate weights is zero, or .

6. The self-balancing torque distribution method for traction force of multiple drive units in a rack railcar according to claim 1, characterized in that: The predicted adhesion coefficient is obtained in the following way: Extracting the wet and slippery state of the target section Pollution status and the degree of unfavorable slope The three indicators; Calculate the adhesion correction factor: ; In the formula, The adhesion correction factor for the target section. The minimum adhesion retention factor, , , The weighting coefficients are respectively for wet and slippery conditions, contaminated conditions, and unfavorable slope conditions. All weighting coefficients are non-negative and their sum is 1. The adhesion correction coefficient is not less than the minimum adhesion retention coefficient and not greater than 1, wherein the minimum adhesion retention coefficient is a preset constant that is greater than 0 and less than 1. Multiply the local adhesion coefficient of each drive unit by the adhesion correction coefficient respectively: ; In the formula, For the first The predicted adhesion coefficient of the driving unit in the k-th control cycle For the first The local adhesion coefficient of a driving unit in the kth control cycle.

7. The self-balancing torque distribution method for traction force of multiple drive units in a rack railcar according to claim 6, characterized in that: The local adhesion coefficient of each drive unit characterizes the ratio of the tangential force to the normal force of the gear teeth at the meshing point of the drive unit. Its value is equal to the ratio of the output torque of the drive wheel of the drive unit to the product of the drive wheel radius and the normal force of the gear tooth contact.

8. The self-balancing torque distribution method for traction force of multiple drive units in a rack railcar according to claim 1, characterized in that: The equilibrium solution of the weighted collaborative game model satisfies the following three constraints: ； ； ； The total traction torque requirement for the rack rail chock vehicle. To allocate weights to achieve the equilibrium; ； The normal force at the tooth contact point of the i-th drive unit during the k-th control cycle. Let the radius of the drive wheel corresponding to the i-th drive unit be . The average intermediate weight of each drive unit during the k-th control cycle. This represents the maximum traction torque of the i-th drive unit of the rack railcar during the k-th control cycle.

9. The self-balancing torque distribution method for traction force of multiple drive units in a rack railcar according to claim 1, characterized in that: The torque redistribution includes: The amount of torque reduction caused by the limiting is determined as the torque difference to be redistributed; Select the remaining drive units that still have a safe output margin, excluding the over-limit drive units; The torque difference is allocated according to the proportion of the remaining torque margin of each of the other drive units; The allocation result is superimposed on the initial limiting command of the corresponding drive unit to form the final feedforward pre-allocated torque command vector.

10. A self-balancing torque distribution system for the traction force of multiple drive units in a rack railcar, used to execute the self-balancing torque distribution method for the traction force of multiple drive units in a rack railcar as described in any one of claims 1-9, characterized in that: include: The section adaptive weight correction quantity generation module is configured to collect the environmental status and current real-time position of the section in front of the rack railcar, identify the target section, and generate a section adaptive weight correction quantity that characterizes the difference in risk exposure of each drive unit in the target section; the section adaptive weight correction quantity is accompanied by the generation time, target section number, effective time window and confidence index. The local allocation weight determination module is configured to receive the local target output torque of each drive unit in the current control cycle, calculate and output the local allocation weight; The traction force suggestion allocation weight determination module is configured to receive the segment adaptive weight correction amount and the local allocation weight, perform validity verification, and merge the two after the verification is passed to output the traction force suggestion allocation weight. The predicted adhesion coefficient acquisition module is configured to receive the target segment environmental vector and the local adhesion coefficient of each driving unit, perform feedforward correction, and output the predicted adhesion coefficient. The equilibrium weight allocation acquisition module is configured to receive the local allocation weight, the traction force suggested allocation weight and the predicted adhesion coefficient, construct a weighted collaborative game model, and solve for the equilibrium allocation weight that satisfies the constraints of total weight normalization, non-negativity and adhesion ability. The feedforward pre-allocation torque command generation module is configured to receive the equalization allocation weight, generate the feedforward target torque, perform amplitude limiting processing and torque redistribution, and output a traction demand reduction command or safety risk warning when the sum of the remaining available capabilities of all drive units is insufficient to cover the amplitude limiting difference. The feedforward pre-allocated torque command generation module outputs the feedforward pre-allocated torque command vector for each drive unit. The output of the segment adaptive weight correction generation module is communicatively connected to the first input of the traction force suggestion allocation weight determination module. The output of the local allocation weight determination module is communicatively connected to the second input of the traction force suggestion allocation weight determination module. The output of the traction force suggestion allocation weight determination module is communicatively connected to the first input of the balanced allocation weight acquisition module; The output of the predicted adhesion coefficient acquisition module is communicatively connected to the second input of the balanced weight allocation acquisition module. The output of the balanced weight allocation acquisition module is communicatively connected to the input of the feedforward pre-allocated torque command generation module.

Images