Locomotive adhesion control method and device based on model-free adaptive sliding mode control and medium

Abstract

The invention relates to a locomotive adhesion control method and device based on model-free self-adaptive sliding mode control and a medium, and the method comprises the steps: collecting the traction torque and the wheel rotating speed of a locomotive at the current moment, and calculating the current creep speed; on the basis of the current creep speed and the adhesion coefficient obtained through the full-dimensional state observer, the optimal reference creep speed under the current rail surface condition is estimated online; the optimal reference creep speed is used as a tracking target, a model-free self-adaptive sliding mode controller constructed based on a partial format dynamic linearization data model and sliding mode control is utilized, and the traction torque adjusting amount is obtained through calculation; and the locomotive traction torque is adjusted in real time according to the traction torque adjusting quantity, so that the actual creep speed tracks the optimal reference creep speed, and the locomotive runs near the optimal adhesion point. Compared with the prior art, the method has the advantages of being high in practicability, improving user experience, being high in safety and the like.

Description

Technical Field

[0001] This invention relates to the field of rail transit locomotive control technology, and in particular to a locomotive adhesion control method, equipment and medium based on model-free adaptive sliding mode control. Background Technology

[0002] The effective traction of electric locomotives relies on the transmission of traction motor torque to the rails via wheel-rail adhesion. Wheel-rail adhesion characteristics exhibit strong nonlinearity, time-varying nature, and randomness, making them susceptible to sudden changes in rail surface conditions such as moisture, oil, ice, and snow. These changes can cause a sharp drop in the usable adhesion coefficient, leading to wheel slippage or wheel roll. This not only results in traction loss and energy waste but also exacerbates wheel-rail wear, seriously threatening operational safety and efficiency.

[0003] Currently, locomotive adhesion control mainly employs model-based control methods such as PID control, linearized model-based design, and rule-based logic control methods such as acceleration thresholding. However, the locomotive traction system is a highly nonlinear, large-inertial system encompassing multiple mass blocks, transmission clearances, and complex wheel-rail contact relationships, making it difficult to establish an accurate, universal mathematical model. Controllers based on inaccurate or simplified model designs often exhibit insufficient robustness and dynamic performance under severe disturbances such as sudden changes in rail surface conditions. This often manifests as response lag, excessive overshoot, or frequent control actions, leading to drastic fluctuations in traction output and impacting operational stability and adhesion utilization efficiency.

[0004] While traditional sliding mode control offers some robustness to parameter perturbations and disturbances, its control performance heavily relies on prior knowledge of the system model and disturbance upper bounds. Furthermore, its inherent high-frequency chattering problem can adversely affect the transmission mechanism, limiting its application in precision control systems. On the other hand, model-free adaptive control, as a data-driven method, does not require a precise mechanistic model of the controlled object, relying solely on online modeling based on system input and output data. It exhibits good adaptability to nonlinear and time-varying systems, but its dynamic response speed and anti-interference capability remain insufficient when dealing with disturbances of large amplitude and rapid changes. Therefore, how to achieve a locomotive adhesion control method that combines fast response, strong robustness, and effective suppression of control chatter without requiring a precise system model is a technical problem that needs to be solved. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the existing technology by providing a locomotive adhesion control method, device and medium based on model-free adaptive sliding mode control. By integrating the data-driven modeling capability of model-free adaptive control with the strong robustness of sliding mode control, and by using a higher-order sigmoid function to improve the reaching law to suppress chattering, fast, stable and highly anti-interference tracking control of the optimal adhesion point is achieved without the need for a precise locomotive mechanism model.

[0006] The objective of this invention can be achieved through the following technical solutions: According to one aspect of the present invention, a locomotive adhesion control method based on model-free adaptive sliding mode control is provided, the specific steps of which include: S1. Collect the locomotive's traction torque and wheel speed at the current moment, and calculate the current creep speed; S2. Based on the current creep speed and the adhesion coefficient obtained through the full-dimensional state observer, estimate the optimal reference creep speed under the current track surface conditions online; S3. Using the optimal reference creep speed as the tracking target, the traction torque adjustment amount is calculated by using a model-free adaptive sliding mode controller constructed based on a partial scheme dynamic linearization data model and sliding mode control. S4. Adjust the locomotive's traction torque in real time according to the traction torque adjustment amount, so that the actual creep speed tracks the optimal reference creep speed, and the locomotive runs near the optimal adhesion point.

[0007] Furthermore, the model-free adaptive sliding mode controller is obtained based on the partial scheme dynamic linearized data model, and its expression is: , in, In order to be in The creep speed at any given moment; To control the input increment vector, In order to be in The output traction torque adjustment of the model-free adaptive sliding mode controller at any given time; In order to be in The uncertainty function at time; It is a pseudo-partial derivative matrix. For the output of the adaptive sliding mode controller; To control the length of the input constant.

[0008] Furthermore, the pseudo-partial derivative matrix The estimation expression is: , in, Let be the estimated value of the pseudopartial derivative matrix at time k. This is an estimate of the pseudopartial derivative matrix at time k-1. Let μ be the step size factor, and μ be the weighting factor. , .

[0009] Furthermore, the sliding mode control function is composed of a linear combination of the creep speed tracking error at the current moment and the creep speed tracking error at the previous moment, where the combination coefficients are the sliding surface design parameters that satisfy the convergence condition.

[0010] Furthermore, the sliding mode control employs an improved discrete reaching law, which uses a higher-order sigmoid function instead of the traditional sign function. The dynamic behavior is jointly determined by the reaching speed parameter, the robust gain parameter, and the sampling period.

[0011] Furthermore, the output of the model-free adaptive sliding mode controller for: , in, Design a parameter vector for the sliding surface, where To approximate the velocity parameters, The system sampling period is Design parameters for the sliding surface; This is the output of the model-free adaptive sliding mode controller at the current moment, i.e., the traction torque adjustment amount; This is the output from the previous time step; In order to be in The creep speed at any given moment; This is the optimal reference creep velocity for the next time step, k+1. , The pseudo-partial derivative vector The first two components are used to describe the dynamic response characteristics of the system output to changes in the control input; In order to be in The uncertainty function at time; For tracking error vector; This refers to the robust gain parameter in the reaching law; The sampling period; It is a higher-order Sigmoid function used to replace the traditional sign function to suppress and control chattering.

[0012] Furthermore, the specific steps for obtaining the optimal reference creep velocity include: Based on the adhesion coefficient estimate obtained from the full-dimensional state observer and the calculated current creep velocity, the slope of the wheel-rail adhesion characteristic curve at the current working point is estimated online using the recursive least squares method with forgetting factor. The gradient search decision logic is implemented based on the slope estimate: if the slope is greater than zero, the optimal reference creep speed is increased; if the slope is less than zero, the optimal reference creep speed is decreased; if the slope is equal to zero, the optimal reference creep speed is kept basically stable.

[0013] Furthermore, when using the recursive least squares method with a forgetting factor to estimate the slope of the wheel-rail adhesion characteristic curve at the current operating point online, the forgetting factor ranges from 0.95 to 0.99.

[0014] According to a second aspect of the present invention, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the program to implement the method described thereon.

[0015] According to a third aspect of the present invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the method described thereon.

[0016] Compared with the prior art, the present invention has the following beneficial effects: (1) This invention uses a partial scheme dynamic linearized data model to perform data-driven modeling of the locomotive adhesion system and designs an online estimation algorithm for pseudo-partial derivatives based on projection operators. This allows the controller to be updated only by the input and output data of the system, without relying on the precise mechanism model of the locomotive-track system. This enables the controller to have a strong adaptive learning capability and to automatically track the dynamic changes of the system. It eliminates the dependence on precise mathematical modeling of complex, nonlinear and time-varying controlled objects and enhances the versatility and engineering feasibility of the control method.

[0017] (2) This invention combines the model-free adaptive mechanism with sliding mode control and uses a higher-order sigmoid function to improve the discrete reaching law to replace the traditional sign function. This makes the controller inherit the strong robustness of sliding mode variable structure to parameter perturbation, unmodeled dynamics and external disturbances. It also effectively suppresses the high-frequency chattering of the control signal through the continuous switching function. When dealing with strong disturbances such as sudden changes in rail surface conditions, the system can still maintain a fast and stable dynamic response, ensuring the stability and smoothness of traction output, reducing the adverse effects on the transmission mechanism, and improving driving comfort and equipment life.

[0018] (3) This invention constructs a two-layer adaptive structure that includes feedforward target optimization and feedback tracking control. By estimating the slope of the adhesion curve online based on the adhesion coefficient and creep speed, and dynamically generating the optimal reference creep speed according to the gradient search logic, it realizes the active and real-time perception and decision-making of the optimal working point under the current rail surface condition. Combined with the precise tracking of the subsequent modelless adaptive sliding mode controller, a complete perception-decision-execution intelligent control closed loop is formed, which enables the locomotive to automatically adapt to different rail surface conditions such as dry, wet, and oily conditions, and always run near the optimal adhesion point, thereby fundamentally improving the adhesion utilization efficiency and operational safety. Attached Figure Description

[0019] Figure 1The flowchart shows the locomotive adhesion control method based on model-free adaptive sliding mode control. Figure 2 A schematic diagram of the dynamic model structure of a locomotive's single-wheelset traction transmission system; Figure 3 The diagram shows the internal structure and signal flow of a model-free adaptive sliding mode controller. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0021] To overcome the shortcomings of existing technologies, such as insufficient controller adaptability and lag in dynamic response due to reliance on precise mathematical models under complex conditions like abrupt changes in track surface conditions, and the impact of chattering on control stability caused by the reliance on prior model knowledge in traditional sliding mode control, this embodiment provides a locomotive adhesion control method based on model-free adaptive sliding mode control (MFASMC). By integrating the online learning capability of model-free adaptive control with the strong robustness of sliding mode control, and designing an improved reaching law to suppress chattering, an adhesion control strategy that does not rely on a precise object model and possesses both fast tracking and strong anti-interference capabilities is constructed. This scheme enables the locomotive to automatically identify and track the optimal adhesion point under different track surface conditions, thereby effectively improving adhesion utilization efficiency and operational stability, while ensuring the smoothness and reliability of the control process.

[0022] like Figure 1 As shown in this embodiment, a locomotive adhesion control method based on model-free adaptive sliding mode control is provided. The specific steps include: S1. Collect the locomotive's traction torque and wheel speed at the current moment, and calculate the current creep speed; S2. Based on the current creep velocity and the adhesion coefficient obtained through the full-dimensional state observer, estimate the optimal reference creep velocity under the current track surface conditions online; S3. Using the optimal reference creep speed as the tracking target, the traction torque adjustment amount is calculated by using a model-free adaptive sliding mode controller constructed based on a partial scheme dynamic linearization data model and sliding mode control. S4. Adjust the locomotive's traction torque in real time according to the traction torque adjustment amount, so that the actual creep speed tracks the optimal reference creep speed, and the locomotive runs near the optimal adhesion point.

[0023] like Figure 2The diagram shown is a schematic representation of the dynamic model of the locomotive's single-wheelset traction transmission system in this embodiment. The traction torque output by the motor is marked with the motor shaft as the starting point of the power source. T m With motor shaft speed ω m This data serves as the core input for calculating creep velocity and achieving adhesion control. Power is transmitted from the motor shaft via a gear transmission mechanism consisting of a drive gear and a driven gear. The torque transmission parameters at the gear meshing point are also relevant. T wm The drive gear, or wheel, is the node in a mechanical structure where power is transmitted, reflecting the influence of mechanical characteristics such as transmission clearance and gear ratio on power transmission. The driven gear is directly connected to the wheel, ultimately generating traction through wheel-rail contact. F t The wheel-rail contact interface is the core area where adhesion occurs. Changes in rail surface conditions, such as moisture, oil, ice, and snow, directly affect the adhesion coefficient at this point. Furthermore, wheel-related motion parameters, including wheel radius, also play a role. R Vehicle speed v and acceleration a Creep speed, as a core indicator reflecting the wheel-rail adhesion state, is calculated based on the synergistic relationship between speed, rotational speed, and mechanical structure parameters.

[0024] The viscosity coefficient, which cannot be directly measured, is obtained online through a full-dimensional state observer to measure the output torque of the traction motor. and wheelset speed As the input signal, by appropriately configuring the two poles of the observer, its response speed is made much faster than the dynamics of the controlled object itself. Typically, the poles are set on the left half-real axis of the complex plane, far from the origin, to ensure fast convergence of the state estimation and a certain degree of noise filtering. The observer outputs an estimated value of the motor load torque online. Then, the wheel-rail adhesion coefficient is calculated according to the formula. The estimated value provides a key input for subsequent optimal reference creep velocity decision-making.

[0025] like Figure 3 The diagram shows the internal structure and signal flow of a model-free adaptive sliding mode controller. The tracking error of the creep speed is defined as the control input, and the output is the increment of the motor torque command. The locomotive adhesion control system is converted into a nonlinear discrete model, expressed as: , in, For locomotive adhesion control system conversion in Output creep speed at any time In order to be in The output of the time controller In order to be in The uncertainty function at time includes uncertainty disturbances, parameter estimation errors, etc.

[0026] make: , The expression for the nonlinear discrete model can then be rewritten as: .

[0027] In this embodiment, the multidimensional function is pre-converted into a nonlinear discrete model. The system's inputs and outputs have continuous bounded partial derivatives; after being transformed into a nonlinear discrete model, it satisfies the generalized Lipschitz condition, that is, for any time and ,have In the formula , , It is a constant greater than zero.

[0028] The locomotive adhesion control system is controllable when two preset conditions are met. hour, ,in, Let be a partial derivative matrix, defined as: , . It is a constant that controls the length of the input.

[0029] Therefore, the model-free adaptive sliding mode controller is derived from the partial scheme dynamic linearized data model, and its expression is: , in, In order to be in The creep speed at any given moment; To control the input increment vector, In order to be in The output traction torque adjustment of the model-free adaptive sliding mode controller at any given time; In order to be in The uncertainty function at time; It is a pseudo-partial derivative matrix. For the output of the adaptive sliding mode controller; To control the length of the input constant.

[0030] pseudo-partial derivative matrix The estimation expression is: , in, Let be the estimated value of the pseudopartial derivative matrix at time k. This is an estimate of the pseudopartial derivative matrix at time k-1. Let μ be the step size factor, and μ be the weighting factor. , .

[0031] The sliding mode control function is a linear combination of the creep velocity tracking error at the current moment and the creep velocity tracking error at the previous moment, where the combination coefficients are the sliding surface design parameters that satisfy the convergence condition. The expression is: , in, for k The creep velocity tracking error at any given moment. for k The optimal reference creep velocity at any given time. Design parameters for the sliding surface. , Design parameter vectors for the sliding surface. .

[0032] Sliding mode control employs an improved discrete reaching law, which uses a higher-order sigmoid function instead of the traditional sign function. The dynamic behavior is jointly determined by the reaching speed parameter, the robust gain parameter, and the sampling period. The expression for the discrete reaching law is: , in, , The sampling period is, and ; This is a higher-order sigmoid function used to replace the sign function to suppress chattering.

[0033] Output of Model-Free Adaptive Sliding Mode Controller for: , in, Design a parameter vector for the sliding surface, where To approximate the velocity parameters, The system sampling period is Design parameters for the sliding surface; This is the output of the model-free adaptive sliding mode controller at the current moment, i.e., the traction torque adjustment amount; This is the output from the previous time step; In order to be in The creep speed at any given moment; This is the optimal reference creep velocity for the next time step, k+1. , The pseudo-partial derivative vector The first two components are used to describe the dynamic response characteristics of the system output to changes in the control input; In order to be in The uncertainty function at time; For tracking error vector; This refers to the robust gain parameter in the reaching law; The sampling period; It is a higher-order Sigmoid function used to replace the traditional sign function to suppress and control chattering; q The meaning is the approximation velocity parameter.

[0034] The specific steps for obtaining the optimal reference creep velocity include: Based on the viscosity coefficient estimate obtained from the full-dimensional state observer and the calculated current creep velocity The recursive least squares method with a forgetting factor is used to estimate the slope of the wheel-rail adhesion characteristic curve at the current operating point online; the slope of the wheel-rail adhesion characteristic curve is based on the estimated value of the adhesion coefficient. and current creep speed The differential signal is obtained. Implement gradient search decision logic based on slope estimates: If the slope If the value is greater than 0, the running point is determined to be to the left of the optimal point, and a fixed step size or the same as the minimum step size is used. The optimal reference creep velocity is increased proportionally with the step size; If the slope If the value is less than 0, the running point is determined to be to the right of the optimal point, and the optimal reference creep speed is reduced by a fixed step size. If the slope If the value is 0, it is determined that the running point is close to the optimal value, and the optimal reference creep speed is kept basically stable.

[0035] When estimating the slope of the wheel-rail adhesion characteristic curve at the current operating point online using the recursive least squares method with a forgetting factor, the forgetting factor... The value is close to 1, and the range is from 0.95 to 0.99, which enables the algorithm to both track time-varying parameters and maintain sufficient estimation smoothness.

[0036] The adhesion coefficient for slope estimation is obtained from a full-dimensional state observer.

[0037] During implementation, for the model-free adaptive sliding mode controller, the pseudo-partial derivative vector is first initialized. (Typically, this can be set to a small non-zero vector), control input increment vector, and controller state.

[0038] Calculate the error and sliding surface, and calculate the tracking error. Calculate the current sliding surface function value. .

[0039] In each control cycle, the pseudo-partial derivative vector is updated using the latest input and output data. Among them, the step size factor and weighting factors It needs to be selected according to the subsequent adjustment principles.

[0040] The updated pseudopartial derivative vector ,error Slip mode surface value And the parameters of the reaching law Substituting the terms of the higher-order sigmoid function into the control law formula given in claim 5, the incremental traction torque command at the current moment is calculated. .

[0041] Control output: The integrated torque command It is sent to the traction control system to complete closed-loop control.

[0042] In this embodiment, the key adjustable parameters of the MFASMC controller can be adjusted according to the dynamic characteristics of the specific controlled object: Among them, the partial scheme dynamic linearization PFDL parameters Adjustments will be made: Step size factor The adjustment directly affects the update speed of the pseudopartial derivative. If the value is too small, convergence will be slow; if it is too large, estimation oscillation may occur. Start by trying values ​​around 1. Regarding the weighting factor... The adjustment is used to prevent the denominator of the algorithm from being too small and to ensure numerical stability. A small positive number (such as between 1e-5 and 1) is selected.

[0043] Design parameters for sliding surfaces The adjustment determines the dynamic convergence characteristics of the error. Increase the design parameters of the sliding surface It can accelerate the error decay rate, but may increase the initial amplitude of the control quantity.

[0044] For the reaching law parameter Adjustments will be made: Approach velocity The adjustment needs to meet the following requirements. , The larger the value, the faster the state tends to the sliding surface, but the amplitude of the control variable must also be considered; this affects the robustness gain. The adjustments are used to combat disturbances and unmodeled dynamics. The larger the value, the stronger the robustness, but it may exacerbate chattering and needs to be adjusted in conjunction with a higher-order sigmoid function.

[0045] For parameters of higher-order sigmoid functions The adjustment is to use a higher-order sigmoid function. hour, The larger the value, the closer the function is to the sign function, which enhances robustness but increases the risk of chattering. The smaller the value, the smoother the output, but the robustness decreases slightly. A trade-off needs to be struck between smoothness and robustness.

[0046] This embodiment is set with reference to typical values ​​of a certain type of electric locomotive, such as: wheelset radius. =0.625 m, axle load =25000kg, gearbox transmission ratio =5.64, transmission efficiency =0.95. To fully verify the adaptive and robust performance of the controller, a severe test scenario with abrupt changes in track surface conditions was designed. During the total simulation time, the locomotive was arranged to run sequentially on a high-adhesion dry track surface, a low-adhesion wet track surface, and then return to a high-adhesion track surface, simulating actual operation conditions such as weather changes or track surface contamination. A traditional PI controller, widely used in industry and simple in design, and a basic sliding mode controller (SMC) with unoptimized parameters were selected as performance comparison benchmarks. The PI controller parameters were set according to typical engineering tuning methods, and the basic SMC adopted a constant velocity reaching law.

[0047] By conducting multiple tests on the simulation model, and comprehensively analyzing and comparing the performance of the method in this embodiment from multiple dimensions such as time-domain response curves, performance indicators, and control signal quality: In terms of dynamic response speed and convergence, the MFASMC can quickly detect changes in the system's optimal operating point at the instant of a step change in track surface conditions. Because its embedded online reference speed generation module can rapidly adjust the control target, and the controller itself updates model parameters based on real-time data, it can drive the actual creep speed to converge quickly and smoothly to the new optimal value. Simulation waveforms show that its settling time is significantly shorter than that of a traditional PI controller; compared to the basic SMC, it exhibits a better convergence trend and smaller initial deviation during the dynamic adjustment process after a sudden change.

[0048] In terms of tracking accuracy and steady-state performance, throughout the simulation process including abrupt changes, the creep speed under MFASMC control closely fluctuates around the dynamically changing optimal reference value. After entering steady state, the magnitude of its steady-state error is significantly smaller than that of the PI controller. This is attributed to the continuous learning and compensation of the system dynamics by the model-free adaptive mechanism.

[0049] In terms of robustness and anti-interference capability, the system's equivalent gain decreases in the low-adhesion track surface stage, equivalent to facing continuous strong disturbances. The sliding mode control inherently provides robustness in MFASMC, while its pseudo-partial derivatives... The online updates can adapt to such parameter perturbations to a certain extent. Therefore, compared with PI controllers that mainly rely on fixed parameter feedback, MFASMC can maintain more stable tracking under low adhesion conditions, with smaller fluctuations in creep speed, demonstrating excellent ability to cope with large-scale parameter changes and external disturbances.

[0050] In terms of control signal quality and engineering practicality, direct observation of the torque command signal output by the MFASMC reveals a reasonable trend, with continuous and smooth signal output. This effectively avoids the high-frequency chattering phenomenon, which is detrimental to the actuator, caused by the use of sign functions in the basic SMC. This is attributed to the adoption of a higher-order sigmoid function. The continuous approximation of discontinuous switching terms demonstrates that the present invention maintains strong robustness while possessing good engineering application value.

[0051] In terms of overall performance comparison, by comparing the response curves of PI, basic SMC and MFASMC under the same simulation conditions, it can be intuitively concluded that MFASMC has achieved a better balance and comprehensive improvement in the control performance indicators that have inherent contradictions, such as response speed, tracking accuracy, anti-interference and control smoothness.

[0052] The method in this embodiment achieves closed-loop adaptive sensing and control. Through an online optimal reference speed acquisition module, the system can proactively sense track surface changes and adjust the control target. The subsequent model-free adaptive sliding mode controller accurately executes tracking, forming a complete intelligent adaptive closed loop. Furthermore, this method combines strong robustness with excellent dynamic performance. Sliding mode control provides inherent robustness to unmodeled dynamics and disturbances, while the model-free adaptive mechanism and optimized convergence law jointly ensure rapid dynamic response and smooth control, effectively suppressing chattering. The entire scheme, from target generation to control law execution, is based on data-driven and online estimation, requiring no precise analytical model of the locomotive-track system. It is highly versatile, easy to implement in engineering, and eliminates dependence on precise physical models. Simultaneously, the controller's modular design parameters have clear physical or regulatory meanings, facilitating debugging and optimization by engineers based on the actual system.

[0053] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the described module can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0054] The electronic device of this invention includes a central processing unit (CPU), which can perform various appropriate actions and processes according to computer program instructions stored in read-only memory (ROM) or loaded from a storage unit into random access memory (RAM). The RAM may also store various programs and data required for device operation. The CPU, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0055] Multiple components in the device are connected to an I / O interface, including: input units such as a keyboard, mouse, etc.; output units such as various types of displays, speakers, etc.; storage units such as disks, optical disks, etc.; and communication units such as network interface cards, modems, wireless transceivers, etc. The communication unit allows the device to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks. The processing unit performs the various methods and processes described above, such as the method of the present invention. For example, in some embodiments, the method of the present invention may be implemented as a computer software program tangibly contained in a machine-readable medium, such as a storage unit. In some embodiments, part or all of the computer program may be loaded and / or installed on the device via ROM and / or the communication unit. When the computer program is loaded into RAM and executed by the CPU, one or more steps of the method of the present invention described above may be performed. Alternatively, in other embodiments, the CPU may be configured to execute the method of the present invention by any other suitable means (e.g., by means of firmware).

[0056] The functions described above in this document can be performed, at least in part, by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: Field Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), Application Standard Products (ASSPs), System-on-Chip (SoCs), Complex Programmable Logic Devices (CPLDs), and so on.

[0057] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0058] In the context of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0059] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A locomotive adhesion control method based on model-free adaptive sliding mode control, characterized in that, The specific steps include: S1. Collect the locomotive's traction torque and wheel speed at the current moment, and calculate the current creep speed; S2. Based on the current creep speed and the adhesion coefficient obtained through the full-dimensional state observer, estimate the optimal reference creep speed under the current track surface conditions online; S3. Using the optimal reference creep speed as the tracking target, the traction torque adjustment amount is calculated by using a model-free adaptive sliding mode controller constructed based on a partial scheme dynamic linearization data model and sliding mode control. S4. Adjust the locomotive's traction torque in real time according to the traction torque adjustment amount, so that the actual creep speed tracks the optimal reference creep speed, and the locomotive runs near the optimal adhesion point.

2. The locomotive adhesion control method based on model-free adaptive sliding mode control according to claim 1, characterized in that, The model-free adaptive sliding mode controller is obtained based on the partial scheme dynamic linearized data model, and its expression is: ， in, In order to be in The creep speed at any given moment; To control the input increment vector, In order to be in The output traction torque adjustment of the model-free adaptive sliding mode controller at any given time; In order to be in The uncertainty function at time; It is a pseudo-partial derivative matrix. For the output of the adaptive sliding mode controller; To control the length of the input constant.

3. The locomotive adhesion control method based on model-free adaptive sliding mode control according to claim 2, characterized in that, The pseudo-partial derivative matrix The estimation expression is: ， in, Let be the estimated value of the pseudopartial derivative matrix at time k. This is an estimate of the pseudopartial derivative matrix at time k-1. Let μ be the step size factor, and μ be the weighting factor. , .

4. The locomotive adhesion control method based on model-free adaptive sliding mode control according to claim 1, characterized in that, The sliding mode control function is composed of a linear combination of the creep speed tracking error at the current moment and the creep speed tracking error at the previous moment, where the combination coefficients are the sliding surface design parameters that satisfy the convergence condition.

5. The locomotive adhesion control method based on model-free adaptive sliding mode control according to claim 4, characterized in that, The sliding mode control employs an improved discrete reaching law, which uses a higher-order sigmoid function instead of the traditional sign function. The dynamic behavior is jointly determined by the reaching speed parameter, the robust gain parameter, and the sampling period.

6. The locomotive adhesion control method based on model-free adaptive sliding mode control according to claim 1, characterized in that, The output of the model-free adaptive sliding mode controller for: ， in, Design a parameter vector for the sliding surface, where To approximate the velocity parameters, The system sampling period is Design parameters for the sliding surface; This is the output of the model-free adaptive sliding mode controller at the current moment, i.e., the traction torque adjustment amount; This is the output from the previous time step; In order to be in The creep speed at any given moment; This is the optimal reference creep velocity for the next time step, k+1. , The pseudo-partial derivative vector The first two components are used to describe the dynamic response characteristics of the system output to changes in the control input; In order to be in The uncertainty function at time; For tracking error vector; This refers to the robust gain parameter in the reaching law; The sampling period; It is a higher-order Sigmoid function used to replace the traditional sign function to suppress and control chattering.

7. The locomotive adhesion control method based on model-free adaptive sliding mode control according to claim 1, characterized in that, The specific steps for obtaining the optimal reference creep velocity include: Based on the adhesion coefficient estimate obtained from the full-dimensional state observer and the calculated current creep velocity, the slope of the wheel-rail adhesion characteristic curve at the current working point is estimated online using the recursive least squares method with forgetting factor. The gradient search decision logic is implemented based on the slope estimate: if the slope is greater than zero, the optimal reference creep speed is increased; if the slope is less than zero, the optimal reference creep speed is decreased; if the slope is equal to zero, the optimal reference creep speed is kept basically stable.

8. The locomotive adhesion control method based on model-free adaptive sliding mode control according to claim 7, characterized in that, When estimating the slope of the wheel-rail adhesion characteristic curve at the current operating point online using the recursive least squares method with a forgetting factor, the forgetting factor ranges from 0.95 to 0.

99.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the program, it implements the method as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1 to 8.

Images