Railway vehicle traction process wheel-rail adhesion state online sensing method and system

Abstract

The application relates to the field of rail vehicle wheel-rail adhesion technology and discloses a rail vehicle traction process wheel-rail adhesion state online sensing method and system. The method is executed by a vehicle-mounted sensing system and comprises the following steps: acquiring a motor output torque signal, a wheel pair angular velocity signal and a vertical load signal, and calculating wheel-rail adhesion force calculation values of each dynamic axle; a traction control unit takes the minimum value in the wheel-rail adhesion force calculation values of each dynamic axle in the same traction bogie as a constraint, corrects the equivalent target traction force of a single dynamic axle, generates a corrected bogie target traction force and converts the corrected bogie target traction force into an equivalent target output torque of a single traction motor, and the equivalent target output torque is used as a torque instruction limiting value of a corresponding traction inverter unit to perform an anti-idling torque limiting control. The application directly uses the online calculation result of the adhesion force in the bogie-level anti-idling control, can actively constrain the traction output before the wheel idles, and reduces the idling risk.

Description

Technical Field

[0001] This application relates to the field of rail vehicle wheel-rail adhesion technology, specifically to an online sensing method and system for the wheel-rail adhesion state during the rail vehicle traction process. Background Technology

[0002] Rail vehicles rely on the adhesion between wheelsets and rails to achieve traction, braking, and guidance. The wheel-rail adhesion state directly determines the effective transmission of traction and braking forces, and is a key foundation for ensuring train operation safety, improving transportation efficiency, and optimizing adhesion utilization. Under complex environmental conditions such as rain, snow, frost, fallen leaves, and oil contamination, the usable adhesion coefficient between wheels and rails decreases significantly, easily leading to wheel slippage or coasting. This not only reduces operating efficiency but can also cause wheel tread abrasion, abnormal wheel-rail wear, and longitudinal impact on the vehicle, and in severe cases, even endanger train operation safety. Therefore, real-time acquisition of the wheel-rail adhesion state is crucial for safe train operation.

[0003] Currently, the wheel-rail adhesion coefficient cannot be directly measured by sensors. Existing online sensing methods for wheel-rail adhesion state mainly employ indirect acquisition methods, including state observer-based methods, parameter identification methods, and dynamic model methods. Most existing research is based on the longitudinal dynamics of the vehicle, using parameters such as wheelset speed, train speed, and braking force to obtain the wheel-rail adhesion state. Extended Kalman filtering, sliding mode observers, and neural networks are common technical approaches.

[0004] However, existing online wheel-rail adhesion state sensing methods still have room for expansion in terms of application conditions and data sources. On the one hand, existing research and related technologies mostly focus on braking conditions, typically estimating the wheel-rail adhesion state during braking based on the dynamic relationship of vehicle braking conditions, combined with information such as wheelset speed, vehicle speed, braking system output parameters, and braking actuator status. For example, Chinese invention patent CN119538600B discloses a wheel-rail adhesion sensing scheme for braking conditions: this scheme is based on the wheelset rotation dynamic equation under braking conditions, transforms the time-domain differential operation to the complex frequency domain through Laplace transform, and introduces a first-order low-pass filter to reconstruct the adhesion expression, achieving real-time estimation of wheel-rail adhesion and adhesion coefficient during braking while avoiding the amplification of noise from direct differentiation of angular velocity. This type of method is of great significance for braking anti-skid control, braking safety assessment, and low adhesion state identification, and has formed a relatively systematic technical foundation. However, since the formation of braking force during braking involves the coupling of multiple factors such as brake cylinder pressure, brake transmission components, and brake shoe or pad friction characteristics, the relevant parameters usually need to be obtained through model conversion or indirect estimation. The estimation results are affected to some extent by changes in the braking system state, friction pair characteristics, and operating environment. Therefore, wheel-rail adhesion state perception methods established for braking conditions focus more on adhesion utilization assessment and anti-skid control requirements during braking. For continuous perception of wheel-rail adhesion state during traction operation, it is still necessary to establish a method that matches the dynamic characteristics under traction conditions.

[0005] On the other hand, there is still room for improvement in the online sensing and control of wheel-rail adhesion under traction conditions. Although both traction and braking follow the laws of wheelset rotational dynamics, they differ in the direction of force, actuators, available parameters, and control objectives. Obtaining the adhesion state under braking conditions usually requires calculations based on parameters such as brake cylinder pressure, brake transmission mechanism status, and brake shoe or pad friction characteristics. However, under traction conditions, the traction control system can provide the motor output torque in real time, or calculate the motor output torque online based on motor current, voltage, speed, and the motor's mathematical model. After conversion using gear ratio and transmission efficiency, the driving torque acting on the wheelset can be obtained, thus providing a more direct data source for online calculation of wheel-rail adhesion under traction conditions based on wheelset rotational dynamics.

[0006] In existing technologies, wheelset rotational dynamics modeling, Laplace transform, and first-order low-pass filtering have been used to calculate wheel-rail adhesion states. However, these technologies primarily focus on adhesion estimation under braking conditions, rail surface identification in group train operation control, or data-driven adhesion state identification. For the traction process of a single-track rail vehicle, it remains necessary to design a specific online calculation link for adhesion states specific to traction conditions, utilizing existing onboard traction motor output torque and wheelset angular velocity signals. Furthermore, it is essential to convert the real-time calculated wheel-rail adhesion states into the target output torque of the traction motor for traction anti-slip control. Summary of the Invention

[0007] To address the challenges of obtaining wheel-rail adhesion status in real-time and accurately under existing traction conditions, and the fact that existing online adhesion status sensing methods primarily focus on braking conditions or complex observation models, making it difficult to directly serve active control against wheel slippage during traction, this application provides an online sensing method and system for wheel-rail adhesion status during the traction process of rail vehicles.

[0008] In a first aspect, this application provides an online sensing method for wheel-rail adhesion state during the traction process of a rail vehicle. The method is executed by an onboard sensing system installed on the rail vehicle. The onboard sensing system is communicatively connected to the traction control unit of the rail vehicle and is connected to a speed measuring device and a load acquisition device. The rail vehicle has multiple axles, each driven by a corresponding traction motor. Traction motors within the same traction bogie are driven by the same traction inverter unit. The traction control unit generates and distributes the original target traction force to each traction bogie. The method includes: The traction control unit acquires the motor output torque signal from the traction motor in real time, the speed measuring device acquires the wheel angular velocity signal of the corresponding wheelset of the moving axle, and the load acquisition device acquires the vertical load signal. The on-board sensing system calculates the wheel-rail adhesion value of each moving axle based on the motor output torque signal and the wheel-pair angular velocity signal. The onboard sensing system provides the traction control unit with the calculated wheel-rail adhesion values ​​of each moving axle within the same traction bogie; The traction control unit determines the equivalent target traction force corresponding to a single moving axle in the traction bogie based on the allocated original target traction force, and uses the minimum value among the calculated wheel-rail adhesion values ​​of each moving axle in the traction bogie as a constraint to correct the equivalent target traction force, thus obtaining the corrected equivalent target traction force for a single moving axle. The traction control unit generates the corrected target traction force of the traction bogie based on the corrected equivalent target traction force of a single driving axle, and converts the corrected target traction force into the equivalent target driving torque of the wheelset side, and then into the equivalent target output torque of a single traction motor. This torque command limit value is used as the torque command limit value of the traction inverter unit corresponding to the traction bogie, so as to perform anti-idle torque limiting control on the traction motor of the traction bogie.

[0009] The rail vehicle includes at least one motor car, the motor car includes at least one traction bogie, the traction bogie includes multiple driving axles, and each driving axle is equipped with wheelsets; the original target traction force is generated by the traction control unit according to the driver control level, ATO operation requirements, vehicle operating status, traction unit configuration, traction capacity of each motor car, load status, traction system availability status and fault clearing status and is distributed to each traction bogie.

[0010] The calculated wheel-rail adhesion values ​​for each axle, obtained by the onboard sensing system based on the motor output torque signal and the wheelset angular velocity signal, specifically include: The onboard sensing system converts the motor output torque signal into a physical quantity of driving torque acting on the wheelset based on the preset gear ratio and preset transmission efficiency. The wheelset rotation dynamics equation is established by the vehicle-mounted sensing system to characterize the relationship between the physical quantity of driving torque, the wheelset angular velocity signal and the physical quantity of wheel-rail adhesion to be calculated; After the vehicle-mounted sensing system performs a Laplace transform on the dynamic equations, the equations are filtered and reconstructed in the complex frequency domain using a first-order low-pass filter. The reconstructed expression is then subjected to an inverse Laplace transform to calculate the wheel-rail adhesion value, thus avoiding the introduction of noise by directly differentiating the wheel-pair angular velocity signal.

[0011] The vehicle-mounted sensing system converts the motor output torque signal into a driving torque physical quantity acting on the wheelset based on a preset gear ratio and a preset transmission efficiency, specifically including: , Where T(t) is the physical quantity of the driving torque acting on the wheelset. For transmission efficiency, R g T is the gear ratio. m (t) represents the motor output torque signal.

[0012] The process of establishing the wheelset rotation dynamics equations using the onboard sensing system, and performing Laplace transform, filtering reconstruction, and inverse Laplace transform to calculate the wheel-rail adhesion value specifically includes: The rotational dynamics equations of the wheelset are established as follows: , Where J is a preset equivalent rotational inertia parameter for the wheelset. This is the wheel-to-wheel angular velocity signal. For the wheel's angular acceleration, F a (t) represents the physical quantity of wheel-rail adhesion to be calculated, R w Here are the preset wheelset radius parameters, and T(t) is the physical quantity of driving torque; Taking the Laplace transform of this dynamic equation, we get: , Among them, F a (p) is F a The Laplace transform of T(t), where T(p) is the Laplace transform of T(t). for Laplace transform, p represents the initial wheel-pair angular velocity signal at the moment the system begins calculation, where p is the complex frequency; A first-order low-pass filter is introduced into the vehicle-mounted perception system, and its transfer function is: , This is the cutoff angular frequency of a first-order low-pass filter; The viscosity expression is filtered using a first-order low-pass filter to obtain the filtered expression: , Laplace transform of the calculated wheel-rail adhesion force; Reconstruct the filtered expression and define the complex frequency domain auxiliary variable Z(p) to satisfy: ; Make ; Taking the inverse Laplace transform of Z(p), we obtain the time-domain auxiliary state variable z(t) which satisfies the first-order differential equation: , Let z(t) be the rate of change of the time-domain auxiliary state variable with respect to time. The first-order differential equation is solved online to obtain z(t), and the calculated value of the wheel-rail adhesion force is then determined. .

[0013] Cutoff angular frequency of a first-order low-pass filter The value range is 5 s −1 up to 100 s −1 Furthermore, the vehicle-mounted sensing system obtains the current vehicle speed from the speed measuring device, and the vehicle-mounted sensing system determines the speed based on at least one of the following: the current vehicle speed, the fluctuation intensity of the wheel-pair angular velocity signal, and the rate of change of the motor output torque signal. Perform dynamic segmentation adjustments.

[0014] Initial wheel angular velocity signal Take the first set of valid wheel-pair angular velocity signal samples at the start of system calculation or the average value over several sampling periods after startup; the initial value z(0) of the time-domain auxiliary state variable is determined according to the following formula: ,in, Take zero at the initial stage of vehicle traction establishment, or take T(0) / R based on the initial driving torque. w T(0) is the physical quantity of the driving torque at the initial moment of calculation; the vehicle-mounted sensing system performs smooth transition processing on the calculated wheel-rail adhesion value within a preset time after startup.

[0015] The method also includes: The wheel-rail adhesion coefficient is calculated by the onboard sensing system based on the calculated wheel-rail adhesion force and vertical load signal, specifically according to the following formula: ,in, The wheel-rail adhesion coefficient is . Q(t) is the calculated value of the adhesion force between the wheel and rail, and Q(t) is the vertical load signal. The on-board sensing system determines the current wheel-rail adhesion state level based on the wheel-rail adhesion coefficient and drives the on-board audible and visual warning device to output prompt information.

[0016] The operations performed by the traction control unit are specifically for the j-th motor car, including: Obtain the original target traction force F of the b-th traction bogie in the j-th train. tr,j,b,req (t), and determine the equivalent target traction force corresponding to a single moving axle within the traction bogie: , where subscript j is the train index, subscript b is the traction bogie index, subscript req represents the original target value, subscript eq represents the equivalent value, and subscript tr represents the traction force; Receive the calculated wheel-rail adhesion values ​​of the two driving axles within the same traction bogie sent by the onboard sensing system. and ,in Let be the calculated value of the wheel-rail adhesion force of the first moving axle in the b-th traction bogie of the j-th motor. The value of wheel-rail adhesion for the second moving axle in the b-th traction bogie of the j-th EMU is given by the subscript a, where a indicates adhesion. Using the minimum value of the calculated wheel-rail adhesion force of each driving axle in the traction bogie as a constraint, the corrected equivalent target traction force of a single driving axle is determined: ; Generate the corrected target traction force for the traction bogie: ; The corrected bogie target traction force is converted into an equivalent wheelset side target driving torque: , where R w The preset wheelset radius parameter, with the subscript d indicating the driving torque; According to the preset gear ratio R g and transmission efficiency The equivalent target driving torque on the opposite wheels is converted into the equivalent target output torque of a single traction motor: The subscript m indicates the traction motor side; The equivalent target output torque of a single traction motor is used as the torque command limit value for the corresponding traction inverter unit of the traction bogie, specifically including: When the traction control unit already has the original equivalent motor torque requirement T corresponding to the traction bogie m,j,b,eq,req When (t), the final equivalent motor torque command is limited to: , among which, T m,j,b,eq,cmd (t) represents the equivalent motor torque command ultimately output to the traction inverter unit corresponding to the b-th traction bogie in the j-th EMU, T m,j,b,eq,req (t) represents the original equivalent motor torque demand generated by the traction control unit, and the subscript cmd indicates the final command value; this is used to perform anti-idle torque limiting control on the traction motor of the traction bogie.

[0017] In a second aspect, this application provides an online sensing system for wheel-rail adhesion state during the traction process of a rail vehicle, wherein the online sensing system for implementing the method as described in any one of the first aspects includes an operation information acquisition module, an online adhesion calculation module, a target value conversion module, a motor target output torque calculation module, and a traction anti-idle control module; The operation information acquisition module is used to acquire the motor output torque signal, the wheel pair angular velocity signal, the vertical load signal, the vehicle speed signal, and vehicle parameters, including the wheelset equivalent moment of inertia parameter, wheelset radius parameter, gear ratio, and transmission efficiency. The online adhesion calculation module receives the motor output torque signal, the angular velocity signal of each driving wheel pair, the vertical load signal, and vehicle parameters sent by the operation information acquisition module. Based on the gear transmission ratio and transmission efficiency, it converts the motor output torque signal into a physical quantity of driving torque acting on the wheelset. Based on the rotational dynamics relationship of the wheelset, it calculates the adhesion value between each driving axle and the rail and the wheel-rail adhesion coefficient, and outputs the adhesion value between each driving axle and the rail and the wheel-rail adhesion coefficient to the target value conversion module. The target value conversion module receives the original target traction force of each traction bogie and the calculated wheel-rail adhesion force of each moving axle within the same traction bogie. It uses the minimum of the calculated wheel-rail adhesion force of each moving axle within the same traction bogie as the constraint of the equivalent target traction force of a single moving axle. Based on this constraint, it corrects the equivalent target traction force of a single moving axle determined by the original target traction force, generates the corrected equivalent target traction force of a single moving axle, and further generates the corrected bogie target traction force. It then converts the bogie target traction force into the equivalent wheelset side target driving torque and outputs it to the motor target output torque calculation module. The motor target output torque calculation module is used to receive the equivalent wheel-side target driving torque, convert the equivalent wheel-side target driving torque into the equivalent target output torque of a single traction motor according to the gear ratio and transmission efficiency, and output it to the traction anti-idle control module. The traction anti-slip control module is used to receive the equivalent target output torque of a single traction motor, use it as the torque command limit value of the corresponding traction inverter unit, and output anti-slip control commands to the traction motor and transmission system to perform anti-slip torque limiting control on the traction motor of the traction bogie. The traction motor and transmission system receive anti-idle control commands to execute traction output, and feed back the motor output torque signal, the wheel-to-wheel angular velocity signal, the vertical load signal and the vehicle speed signal to the operation information acquisition module to form a closed-loop control circuit.

[0018] In summary, compared with the prior art, this application has the following advantages: (1) This application establishes an online sensing system for wheel-rail adhesion state under traction conditions, which can utilize the existing motor output torque, wheel-pair angular velocity, vehicle speed and load information in the traction control system and speed measurement system, reducing the reliance on external sensors and parameters that are difficult to accurately determine.

[0019] (2) This application uses the output torque of the traction motor as the core input, and converts it into the driving torque acting on the wheelset through the gear transmission ratio and transmission efficiency, forming a parameter conversion link suitable for traction conditions. Compared with the method of indirectly obtaining braking force by combining the brake cylinder pressure, the state of the brake transmission mechanism and the friction characteristics of the brake shoe or brake pad under braking conditions, the input parameters of this application are more direct and easier to apply in real time on the vehicle.

[0020] (3) Based on the rotational dynamics of wheelsets, this application links the driving torque acting on the wheelsets, the angular velocity of the wheelsets, and the adhesion between the wheels and rails, so that the wheel-rail adhesion state, which is difficult to measure directly, can be calculated online through the existing on-board signals. The physical meaning is clear and the calculation link is clear.

[0021] (4) To address the problem that direct differentiation of wheel-rail angular velocity can easily amplify speed measurement noise and pulse quantization error, this application introduces complex frequency domain reconstruction and first-order low-pass filtering in the process of solving adhesion, and provides the selectable range of filter parameters and calibration method, thereby improving the stability of the online calculation results of wheel-rail adhesion and wheel-rail adhesion coefficient.

[0022] (5) This application supplements the adhesion state classification threshold, data preprocessing rules, gear transmission ratio value method, transmission efficiency value method and initialization assignment method, so that the system can better adapt to the real-time computing needs of the vehicle and reduce the impact of abnormal sampling, communication delay, initialization transient and parameter uncertainty on the calculation results.

[0023] (6) This application can also apply the wheel-rail adhesion force and wheel-rail adhesion coefficient obtained online to traction anti-slip control. Among them, the wheel-rail adhesion force is used to determine the longitudinal force transmission state of two driving shafts in the same traction bogie, and the wheel-rail adhesion coefficient is used for adhesion state discrimination and auxiliary correction of control strategy. The system converts the original target traction force of the traction bogie into the equivalent target traction force corresponding to a single driving shaft, and uses the driving shaft with poor adhesion state in the same traction bogie as the basis for traction output constraint to form the equivalent target traction force correction value of a single driving shaft, thereby obtaining the corrected target traction force of the traction bogie. The corrected target traction force is further converted into the equivalent wheelset side target driving torque, and combined with the gear transmission ratio and transmission efficiency, it is converted into the equivalent target output torque of a single traction motor in the traction bogie, which serves as the torque control target, correction value or upper limit constraint of the corresponding traction inverter unit. Therefore, when any moving axle in a traction bogie experiences a decrease in adhesion or a tendency to spin freely, the system can limit the overall torque of that traction bogie, while the other traction bogies that have not experienced adhesion limitation or a tendency to spin freely can continue to output the corresponding original target traction force. This actively constrains the traction output before the wheels spin freely, reduces the risk of wheelset spin, and improves the level of adhesion utilization and vehicle running stability under low adhesion conditions. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of the overall structure of the online sensing system for wheel-rail adhesion status during the traction process of rail vehicles in this application. Figure 2 This is a schematic diagram of the wheelset dynamics model under traction conditions; Figure 3 This is a block diagram illustrating the principle of the online sensing method for wheel-rail adhesion state in this application. Figure 4 This is a schematic diagram illustrating a typical application of the wheel-rail adhesion state calculation results of this application in an on-board display and vehicle-to-ground communication system; Figure 5 This is a schematic diagram illustrating the application of the wheel-rail adhesion state calculation results of this application to traction anti-slip control. Detailed Description of Embodiments

[0025] The present application is further described below with reference to the accompanying drawings. The structure and principle of the present application are very clear to those skilled in the art. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0026] The English abbreviations involved in the present application and their interpretations of standard English names are as follows: MVB, whose full name is Multifunction Vehicle Bus, i.e., multifunction vehicle bus, is a commonly used communication bus for rail transit on-board control systems; ATO, whose full name is Automatic Train Operation, i.e., automatic train operation, is used to realize automatic train driving control; 4G and 5G refer to the fourth-generation and fifth-generation mobile communication technologies respectively, Wi-Fi is a wireless local area network communication technology, and all three are common communication methods for train-ground data transmission; EN 15595 is a European railway industry standard, its full English name is *Railway applications - Braking - Wheel slide protection*, and the corresponding Chinese name is 《铁路应用 制动 车轮防滑保护》.

[0027] It should be noted that the "on-board perception system" in the embodiments of the present application mentioned below is not an abstract software function or thinking rule, but is specifically implemented as an on-board control device including a processor, a memory and a communication interface. The processor (which may adopt a general processor in the prior art) is configured to call and execute the computer program stored in the memory, so as to implement the functions described in the present embodiment such as signal acquisition and preprocessing, driving torque conversion, online calculation of wheel-rail adhesion force, calculation of wheel-rail adhesion coefficient, adhesion state discrimination, target value conversion, calculation of motor target output torque, and traction anti-slip control; the communication interface is configured to perform data interaction with the traction control unit, speed measurement device and load acquisition device, and acquire physical signals such as motor output torque, wheelset angular velocity and vertical load in real time. All the functional modules can be implemented by the processor executing corresponding program instructions.

[0028] In the online calculation of adhesion, the Laplace transform, first-order low-pass filtering, and inverse Laplace transform operations involved are performed by the processor through discretized numerical calculations. Specifically, the first-order low-pass filter in the continuous domain is converted into a discrete-domain difference equation using the bilinear transform method or the zero-order hold method. The differential equation is then discretized into an iterative formula using numerical integration methods such as the forward Euler method or the trapezoidal rule, and recursively solved within each onboard control cycle to obtain the time-domain auxiliary state variables and the calculated wheel-rail adhesion value. The aforementioned discretization and numerical solution methods are common techniques in the field of digital signal processing; therefore, unnecessary prior art will not be elaborated upon here.

[0029] Therefore, each step executed by the on-board sensing system in the method described in this embodiment is based on real-time collected physical signals and preset vehicle parameters to estimate the physical state of wheel-rail adhesion during the traction process of the rail vehicle. The estimation results are directly used to control the output torque of the traction inverter unit, thereby changing the actual output torque of the traction motor, which ultimately affects the physical contact state between the wheelset and the rail. The entire process constitutes a technical closed loop of perception, calculation, decision-making, control, and execution, and belongs to a specific control method for the physical system of rail vehicles.

[0030] In actual vehicle deployment, the vehicle-mounted sensing system can be integrated with the vehicle's existing traction control unit on the same hardware platform, or it can be used as an independent controller to interact with data through the vehicle's communication network. The specific hardware selection and communication protocol can be determined according to the requirements of different rail vehicle platforms. These are all common technical means in the field of rail vehicle on-board equipment, and will not be elaborated here.

[0031] The online wheel-rail adhesion state sensing system for rail vehicles described in this application embodiment is deployed in the on-board control equipment of the rail vehicle. It establishes data connections with the vehicle's traction control unit, speed measuring device, load acquisition device, on-board display, and vehicle-to-ground communication equipment. It can independently perform online sensing of wheel-rail adhesion state and can also be linked with the traction control unit to achieve closed-loop control to prevent traction slippage. The rail vehicle includes at least one motor car, each motor car includes at least one traction bogie, each traction bogie includes multiple drive axles, each drive axle is equipped with wheelsets and is driven by a corresponding traction motor, and the traction motors within the same traction bogie are driven by the same traction inverter unit.

[0032] Figure 1 This is a schematic diagram of the overall structure of the online sensing system for wheel-rail adhesion state during the traction process of rail vehicles in this application. The system input signals can be obtained through three methods: MVB signal, Ethernet transmission, or sensor acquisition. The input signals include the motor output torque T. m (t), wheel-to-wheel angular velocity Vertical load Q, vehicle speed V t And vehicle parameters J, Rw R g , The system internally consists of a data acquisition module, a data preprocessing module, a traction drive parameter conversion module, a wheelset dynamics modeling module, an online adhesion calculation module, an adhesion coefficient calculation module, and an adhesion state discrimination module. The data acquisition module receives input signals from various sources; the data preprocessing module performs filtering, outlier removal, and unit conversion; and the traction drive parameter conversion module completes T... m The conversion from wheel-rail adhesion (t) to wheelset-side driving torque T(t) is achieved by the online adhesion calculation module, which calculates the adhesion between the wheel and rail online, and the adhesion state discrimination module completes the state classification and early warning. The system finally outputs four types of output signals: wheel-rail adhesion coefficient, wheel-rail adhesion, adhesion state and early warning information.

[0033] It should be noted that the online sensing method for wheel-rail adhesion state during the traction process of rail vehicles provided in this application is executed by an onboard sensing system configured on the rail vehicle. This onboard sensing system is implemented as an onboard control device including a processor, memory, and a communication interface. The processor calls the computer program stored in the memory to implement the following method steps. The communication interface is used for data interaction with the traction control unit, speed measuring device, load acquisition device, and onboard display and communication equipment.

[0034] The specific methods and procedures include: I. Parameter Acquisition.

[0035] The system collects various operating signals and vehicle-specific parameters in real time under traction conditions through the operation information acquisition module, specifically including: 1. Motor output torque signal T m (t): Outputted in real time by the traction control unit, or calculated online based on the current, voltage, and speed parameters of the traction motor combined with the mathematical model of the motor; 2. Wheelset angular velocity signal The angular velocity of the wheelset corresponding to each moving axle is acquired by a speed measuring device installed at the end of the wheel axle or the end of the traction motor shaft. 3. Vertical load signal Q(t): Acquired by air spring pressure sensor, axle box load sensor or vehicle weighing system, corresponding to the vertical load of each wheelset; 4. Vehicle speed signal: The vehicle speed is provided by the vehicle speed measurement system; 5. Vehicle inherent parameters: Pre-stored in the onboard system, including the equivalent rotational inertia J of the wheelset and the wheelset radius R. w The gear ratio Rg and transmission efficiency of the traction drive system .

[0036] II. Data Preprocessing.

[0037] The system first performs data preprocessing on the real-time signals such as motor output torque, wheel angular velocity, vertical load and vehicle speed to eliminate the impact of sampling asynchrony, pulse interference, unit inconsistency and abnormal fluctuations on subsequent calculations. The preprocessing includes four steps: time synchronization, outlier removal, unit conversion and basic smoothing.

[0038] 1. Time synchronization: Using a unified timestamp alignment, interpolation, or zero-order hold method, signals from different sources are time-aligned within the same calculation cycle, ensuring that the input signals for dynamic calculations are time-matched.

[0039] 2. Outlier Removal: Outliers are determined based on the physical allowable range of vehicle operating parameters, the rate of change of adjacent sampling points, and signal continuity. When the wheel-set angular velocity, motor output torque, vertical load, or vehicle speed exceeds the physical range allowed by the vehicle design, or when a sudden change occurs in adjacent sampling periods that does not conform to the vehicle's dynamic characteristics, it is determined to be an outlier. Outliers are corrected by maintaining the previous valid value, interpolating the nearest valid value, or replacing the value within a sliding window.

[0040] As an optional implementation, the effective range of the vehicle speed signal is determined based on the vehicle's maximum designed operating speed; the effective range of the motor output torque is determined based on the maximum allowable output torque of the traction motor; the effective range of the wheelset angular velocity is determined based on the maximum allowable angular velocity of the wheelset or the maximum speed of the traction motor converted through gear transmission relationship; the effective range of the vertical load is determined based on the nominal vertical load of the wheelset and the range of vehicle load variation; the above thresholds can be calibrated based on vehicle design parameters, sampling period and line test data.

[0041] 3. Unit conversion: Unify the units of physical quantities from different sources to the International System of Units (SI) to ensure dimensional consistency in subsequent dynamic calculations.

[0042] 4. Basic smoothing processing: Differentiated filtering methods are used to address the noise characteristics of different signals. For signals such as wheel angular velocity that are easily affected by the quantization error of the speed measurement pulse, median filtering is used to remove isolated pulse interference. The median filtering window length can be 3, 5 or 7 sampling points. For relatively slowly varying signals such as motor output torque, vertical load, and vehicle speed, a moving average filter is used for short-term smoothing.

[0043] When the calculation results are used as input for traction anti-slip control, the moving average window length can be 3 to 7 sampling points to ensure real-time control; when the calculation results are used for vehicle display, low adhesion warning or vehicle-to-ground communication output, the moving average window length can be 5 to 20 sampling points to improve the stability of the results.

[0044] It should be noted that the basic smoothing process in this step is only used to remove obvious pulse interference and abnormal sampling points, and does not replace the first-order low-pass filtering reconstruction process in the subsequent online viscosity calculation. The viscosity calculation is still based on the wheelset rotation dynamics relationship. The viscosity expression is reconstructed through Laplace transform and first-order low-pass filter to avoid directly differentiating the wheelset angular velocity in the time domain.

[0045] III. Conversion of traction transmission parameters.

[0046] The system, based on the gear transmission relationship between the traction motor and the wheelset, outputs the motor torque signal T. m The conversion of (t) into the physical quantity T(t) of the driving torque acting on the wheelset takes into account the gear ratio, transmission efficiency, and direction of torque. The conversion formula is as follows: , In the formula, T(t) is the physical quantity of the driving torque acting on the wheelset. For transmission efficiency, R g T is the gear ratio. m (t) represents the motor output torque signal.

[0047] Among them, the gear ratio R g The transmission ratio, characterizing the speed and torque on the traction motor side and the speed and torque on the wheelset side, is determined by the gearbox structural design. If the gear transmission ratio is defined as the ratio of the motor-side angular velocity to the wheelset-side angular velocity, then... The gear ratio is typically greater than 1. It can be determined by the ratio of the number of teeth on the traction motor side pinion to the number of teeth on the wheelset side gear, or it can be obtained from vehicle design documents, gearbox nameplate parameters, or traction system calibration parameters; as an optional implementation, R... g The value ranges from 2.0 to 7.0, and the specific value is determined based on the vehicle's maximum operating speed, wheelset radius, traction motor rated speed, traction motor maximum output torque, and vehicle traction characteristics requirements.

[0048] Transmission efficiency Characterizes the mechanical losses during torque transmission through the gearbox and transmission system; as an optional implementation, The value range is 0.94~0.99, with a preferred range of 0.96~0.98. Transmission efficiency can be fixed as a vehicle design parameter or test calibration parameter, or it can be determined based on gearbox efficiency tests, vehicle traction drive system design parameters, or operation and maintenance calibration data. For scenarios requiring improved calculation accuracy, transmission efficiency can be set as a correction parameter related to traction torque, speed, lubrication status, gearbox temperature, or operating mileage, and dynamically corrected through table lookup or piecewise functions.

[0049] The converted driving torque T(t) is used as the input for wheelset dynamics calculation.

[0050] IV. Online calculation of wheel-rail adhesion.

[0051] Figure 2 This is a schematic diagram of the wheelset dynamics model under traction conditions, illustrating the torque transmission path and the force relationship between the wheel and rail: traction motor output torque T m (t), a gear transmission pair consisting of a driving gear and a driven gear (gear transmission ratio R). g After transmission, the driving torque T(t) acts on the wheelset; the wheelset has a wheelset angular velocity. Wheelset radius R w The wheelset has an equivalent moment of inertia J, and the contact point between the wheelset and the rail bears a vertical load Q, generating an adhesive force F between the wheel and rail. a (t), the vehicle's speed is V t .

[0052] Figure 3 The diagram illustrates the principle of an online wheel-rail adhesion state sensing method. The system uses the motor output torque T... m (t), wheel-to-wheel angular velocity Vertical load Q, vehicle speed V t And vehicle parameters J, R w 、Rg、 As the input signal, it first passes through the traction transmission parameter conversion module, based on the formula... The motor output torque is converted into the wheelset-side driving torque; then the modeling module for wheelset dynamics is entered, based on the wheelset rotational dynamics equations. The process sequentially performs Laplace transform, first-order low-pass filtering, expression reconstruction, and inverse Laplace transform to avoid amplifying noise by directly differentiating the angular velocity; then, the calculated wheel-rail adhesion value is output by the online adhesion calculation module. Then proceed to the adhesion coefficient calculation module, based on the formula... The wheel-rail adhesion coefficient was calculated. Finally, the adhesion state discrimination module completes the state classification; the system ultimately outputs the calculated value of the adhesion force between the wheel and rail. Wheel-rail adhesion coefficient Adhesion status and early warning information.

[0053] 4.1 Modeling of wheel rotation dynamics.

[0054] Based on the wheelset rotational dynamics under traction conditions, the system establishes a quantitative relationship between driving torque, wheelset angular velocity, and wheel-rail adhesion. The wheelset rotational dynamics equation is as follows: , In the formula, J is the equivalent rotational inertia of the wheelset. This is the wheel-to-wheel angular velocity signal. For the wheel's angular acceleration, F a (t) represents the physical quantity of wheel-rail adhesion to be calculated, R w Here are the preset wheelset radius parameters, and T(t) is the physical quantity of driving torque.

[0055] The theoretical expression for the adhesion between the wheel and rail can be obtained by rearranging: , This formula shows that the adhesion between the wheel and rail can be calculated from the driving torque, the rate of change of the wheelset angular velocity, and the wheelset parameters. However, in practical vehicle applications, directly differentiating the wheelset angular velocity in the time domain amplifies speed measurement noise and pulse quantization errors, leading to drastic fluctuations in the adhesion calculation results. Therefore, this embodiment performs complex frequency domain processing on the dynamic equation and introduces a first-order low-pass filter to filter and reconstruct the adhesion expression, avoiding the direct calculation of adhesion using the differential term of angular velocity.

[0056] 4.2 Complex frequency domain transformation and filtering reconstruction.

[0057] Applying a Laplace transform to the wheelset rotation dynamics equations yields the viscosity expression in the complex frequency domain: , In the formula, where F a (p) is F a The Laplace transform of T(t), where T(p) is the Laplace transform of T(t). for Laplace transform, p represents the initial wheel-pair angular velocity signal at the moment the system begins calculation, where p is the complex frequency.

[0058] A first-order low-pass filter is introduced into the vehicle-mounted perception system, and its transfer function is: , In the formula, This is the cutoff angular frequency of a first-order low-pass filter, corresponding to the time constant. . The smaller the value, the larger the filtering time constant, the stronger the noise resistance, but the slower the response speed; The larger the value, the smaller the filter time constant and the faster the response speed, but the ability to suppress speed measurement noise and pulse quantization error is weakened.

[0059] As an optional implementation method The value range is 5 s −1 up to 100 s −1 The corresponding time constant is approximately 0.01~0.2s; the preferred value range is 10 s. −1 up to 50 s−1 The corresponding time constant is approximately 0.02~0.1s, which is a good compromise between the response to changes in wheel-rail adhesion during traction and the suppression of speed measurement noise. The specific value can be calibrated based on the sampling frequency of the speed measurement signal, the noise level of the wheel-set angular velocity signal, the vehicle's operating speed, and the rate of change of traction torque. When the vehicle is running at low speed, the quantization error of the speed measurement pulse is significant, or the noise of the angular velocity signal is large, the value can be appropriately reduced. To enhance the filtering effect; when the vehicle is under high-speed traction, the traction torque changes rapidly, or the adhesion state changes rapidly, the filter can be appropriately increased. To improve computational response speed. It can be used as a fixed calibration parameter, or it can be dynamically adjusted in segments according to the current vehicle speed, the fluctuation intensity of the wheel set angular velocity signal, or the rate of change of the motor output torque signal.

[0060] The adhesion expression is filtered by a first-order low-pass filter to obtain the complex frequency domain expression of the wheel-rail adhesion calculation value after filtering: , In the formula, This is the Laplace transform of the calculated wheel-rail adhesion force.

[0061] 4.3 Expression Refactoring.

[0062] Reconstructing the filtered expression yields: , Subsequently, an inverse Laplace transform is performed on the above expression to convert the complex frequency domain expression to the time domain. The complex frequency domain auxiliary variable Z(p) is defined to satisfy:

[0063] The complex frequency domain expression for the calculated wheel-rail adhesion force can be simplified to: , The differential properties of the Laplace transform are as follows: Where z(t) is the time-domain auxiliary state variable corresponding to Z(p), This is the Laplace transform operator.

[0064] Based on this differential property, performing an inverse Laplace transform on the auxiliary variable Z(p) yields the first-order differential equation satisfied by the time-domain auxiliary state variable z(t): , In the formula, Let z(t) be the rate of change of the time-domain auxiliary state variable with respect to time.

[0065] definition Then the above formula can be further written as: , Solving the above differential equation yields the time-domain expression for the auxiliary state variables: .

[0066] 4.4 Smooth transition.

[0067] When the system starts calculating, the initial wheel-pair angular velocity signal Take the first set of valid wheel-pair angular velocity signal samples at the moment the system starts calculating; if there is short-term signal jitter during the startup phase, the average value of the wheel-pair angular velocity over several sampling periods after startup can be used as the average value. .

[0068] The initial value of the time-domain auxiliary state variable \(z(0)\) is determined according to the following formula: In the formula, This is the calculated value of the wheel-rail adhesion at the initial moment; when the vehicle is in the initial stage of traction establishment, with a small or nearly stable traction torque, it can be set as follows: The value is 0, or T(0) / R can be approximated based on the initial driving torque. w Based on this, z(0) is determined. In practical vehicle applications, the stable calculation results within several sampling periods after startup can also be used to initialize z(0) to reduce the impact of startup transients.

[0069] To further reduce the impact of the initialization phase on the calculation results, the system adopts limited output, sliding average, or gradual release output methods within 0.1~1.0s after startup, so that the calculation results of wheel-rail adhesion and wheel-rail adhesion coefficient smoothly enter the normal online calculation state; the above time range can be calibrated according to the sampling period, vehicle control period, and traction system response time.

[0070] V. Calculation of wheel-rail adhesion coefficient and condition classification.

[0071] After obtaining the calculated wheel-rail adhesion values ​​for each moving axle, the system calculates the wheel-rail adhesion coefficient based on the wheel-rail adhesion definition and the real-time vertical load signal. The calculation formula is as follows: , In the formula, This is the calculated value of the wheel-rail adhesion coefficient. Q(t) represents the calculated value of the adhesion force between the wheel and rail, and Q(t) represents the vertical load signal.

[0072] The system classifies the current wheel-rail adhesion state based on the real-time calculated wheel-rail adhesion coefficient. The classification threshold is set with reference to the definitions of low adhesion and extremely low adhesion in EN15595 "Railway applications - Braking - Wheelslide protection" and GB / T44853-2024 "Electro-pneumatic braking systems for urban rail transit vehicles". The specific classification rules are as follows: When the wheel-rail adhesion coefficient is greater than 0.08, it is judged to be in a normal adhesion state; When the wheel-rail adhesion coefficient is greater than 0.05 and less than or equal to 0.08, it is judged as a low adhesion state; When the wheel-rail adhesion coefficient is greater than or equal to 0.03 and less than or equal to 0.05, it is judged as an extremely low adhesion state; When the wheel-rail adhesion coefficient is less than 0.03, it is judged as an ultra-low adhesion state, indicating that the adhesion conditions are below the definition range of ultra-low adhesion, and the system will handle it according to a more stringent low adhesion warning and traction torque limiting strategy.

[0073] When a low adhesion level or below is detected, the system activates the vehicle's audio-visual warning device to output corresponding prompts, reminding the driver to take appropriate driving measures.

[0074] VI. Implementation of traction anti-slip control.

[0075] Figure 4 This diagram illustrates a typical application of wheel-rail adhesion state calculation results in an onboard display and vehicle-to-ground communication system. The system is divided into two parts: the onboard unit and the ground unit. Both units communicate via 4G / 5G / Wi-Fi to achieve bidirectional data exchange. The onboard unit provides the driver's console with onboard display and warning functions. The console interface displays the current wheel-rail adhesion coefficient (example value μ=0.05) and adhesion state level (example: extremely low adhesion). A graded progress bar displays four adhesion levels: ultra-low, extremely low, low, and normal. When a low adhesion condition is detected, an extremely low adhesion warning is issued via an audio-visual device. The ground unit is the ground dispatch center, which performs four core functions based on uploaded adhesion data: adhesion state monitoring, low adhesion section identification, operational organization optimization, and safety decision support.

[0076] Figure 5 This is a schematic diagram of the application of the wheel-rail adhesion state calculation results of this application to traction anti-slip control (the formulas in the figure are explained in the full text of the specific implementation method of the specification, and will not be repeated here).

[0077] The system as a whole forms a closed-loop control loop, which, from left to right, includes an operation information acquisition module, an online adhesion calculation module, a target value conversion module, a motor target output torque calculation module, a traction anti-slip control module, and a traction motor and transmission system. The operation information acquisition module collects the motor output torque, wheel-to-wheel angular velocity, vertical load, vehicle speed, and vehicle parameters, and outputs them to the online adhesion calculation module. The online adhesion calculation module performs torque conversion through gear transmission relationships and calculates based on the wheelset rotational dynamics, outputting the calculated wheel-rail adhesion value and adhesion coefficient for each driving axle. The target value conversion module first converts the original target traction force into an equivalent single driving axle target traction force, and then uses the minimum value of the adhesion force between two driving axles within the same bogie. As a constraint, a corrected equivalent target traction force for a single driving axle is generated and further converted into the target driving torque of the bogie. The target output torque calculation module of the motor converts the target driving torque of the bogie into the equivalent target output torque of a single traction motor based on the gear ratio and transmission efficiency. The traction anti-slip control module receives the equivalent target output torque and uses it as the target value, correction value, or upper limit constraint for the torque control of the traction motor. When there is an original equivalent motor torque requirement, the final torque command takes the minimum of the two values. The traction motor and transmission system execute the torque command and drive the wheelset to run. At the same time, the motor output torque, wheelset angular velocity, vertical load, vehicle speed, and vehicle parameters are fed back to the operation information acquisition module as feedback information to form a complete closed loop.

[0078] Specifically, the online sensing system includes an operation information acquisition module, an online adhesion calculation module, a target value conversion module, a motor target output torque calculation module, and a traction anti-idle control module; The operation information acquisition module is used to acquire the motor output torque signal, the wheel pair angular velocity signal, the vertical load signal, the vehicle speed signal, and vehicle parameters, including the wheelset equivalent moment of inertia parameter, wheelset radius parameter, gear ratio, and transmission efficiency. The online adhesion calculation module receives the motor output torque signal, the angular velocity signal of each driving wheel pair, the vertical load signal, and vehicle parameters sent by the operation information acquisition module. Based on the gear transmission ratio and transmission efficiency, it converts the motor output torque signal into a physical quantity of driving torque acting on the wheelset. Based on the rotational dynamics relationship of the wheelset, it calculates the adhesion value between each driving axle and the rail and the wheel-rail adhesion coefficient, and outputs the adhesion value between each driving axle and the rail and the wheel-rail adhesion coefficient to the target value conversion module. The target value conversion module receives the original target traction force of each traction bogie and the calculated wheel-rail adhesion force of each moving axle within the same traction bogie. It uses the minimum of the calculated wheel-rail adhesion force of each moving axle within the same traction bogie as the constraint of the equivalent target traction force of a single moving axle. Based on this constraint, it corrects the equivalent target traction force of a single moving axle determined by the original target traction force, generates the corrected equivalent target traction force of a single moving axle, and further generates the corrected bogie target traction force. It then converts the bogie target traction force into the equivalent wheelset side target driving torque and outputs it to the motor target output torque calculation module. The motor target output torque calculation module is used to receive the equivalent wheel-side target driving torque, convert the equivalent wheel-side target driving torque into the equivalent target output torque of a single traction motor according to the gear ratio and transmission efficiency, and output it to the traction anti-idle control module. The traction anti-slip control module is used to receive the equivalent target output torque of a single traction motor, use it as the torque command limit value of the corresponding traction inverter unit, and output anti-slip control commands to the traction motor and transmission system to perform anti-slip torque limiting control on the traction motor of the traction bogie. The traction motor and transmission system receive anti-idle control commands to execute traction output, and feed back the motor output torque signal, the wheel-to-wheel angular velocity signal, the vertical load signal and the vehicle speed signal to the operation information acquisition module to form a closed-loop control circuit.

[0079] This application applies the wheel-rail adhesion force and adhesion coefficient calculated online to traction anti-slip control. Using the traction bogie as the basic control unit, the most unfavorable moving axle constraint principle is employed to limit the bogie-level traction output, actively constraining the traction output before significant wheel slippage. All motor cars and traction bogies in the entire train are controlled in the same manner. The specific implementation process is as follows: 6.1 Original target traction force distribution and conversion.

[0080] During the entire train traction process, the train control and management system or the vehicle-level traction control unit allocates traction requirements to each motor car based on the driver's control level, ATO operation requirements, vehicle operating status, traction unit configuration, traction capacity of each motor car, load status, traction system availability status, and fault clearing status. Each motor car's traction control unit further generates the original target traction force corresponding to each traction bogie based on the existing traction force allocation strategy, traction inverter unit status, traction motor status, load status, and fault torque limiting status.

[0081] Taking the b-th traction bogie in the j-th train as an example, let its original target traction force be F. tr,j,b,req (t), where the subscript j is the train index, the subscript b is the traction bogie index, the subscript req represents the original target value, and the subscript tr represents the traction force.

[0082] To facilitate comparison between the bogie-level traction requirements and the wheel-rail adhesion state of the driving axle, the original target traction force of this traction bogie is converted into the equivalent target traction force corresponding to a single driving axle: , In the formula, F tr,j,b,eq,req (t) represents the equivalent target traction force corresponding to a single moving axle in the b-th traction bogie of the j-th motor, with the subscript eq indicating the equivalent value.

[0083] 6.2 Traction force correction based on adhesion constraints.

[0084] The system receives the calculated wheel-rail adhesion values ​​of the two driving axles within the same traction bogie from the onboard sensing module. and ,in Let be the calculated value of the wheel-rail adhesion force of the first moving axle in the b-th traction bogie of the j-th motor. Let be the calculated wheel-rail adhesion force for the second driving axle within the b-th traction bogie of the j-th train, where the subscript 'a' indicates adhesion force. The corresponding calculated wheel-rail adhesion coefficients are as follows: and .

[0085] Since the two traction motors within the same traction bogie are controlled by the same traction inverter unit, when the adhesion of either axle decreases or a tendency to idle occurs, the corresponding traction inverter unit of that traction bogie needs to be torque-limited as a whole. Therefore, the most unfavorable axle constraint principle is adopted, using the minimum value of the calculated adhesion of the two axles as the constraint boundary to correct the equivalent target traction force of a single axle. The corrected equivalent target traction force of a single axle is: .

[0086] Accordingly, the corrected target traction force for the traction bogie is: .

[0087] Therefore, when the adhesion state of the two driving axles in the traction bogie meets the traction requirements, the traction bogie can output the original target traction force; when the adhesion state of either driving axle decreases, the system limits the equivalent target traction force of a single driving axle based on the calculated adhesion force of the driving axle with the poorer adhesion state, and further limits the overall torque of the traction inverter unit corresponding to the entire traction bogie.

[0088] 6.3 Calculation of target output torque of motor.

[0089] Based on the conversion relationship between traction force and wheelset-side driving torque, the corrected bogie target traction force is converted into an equivalent wheelset-side target driving torque: , In the formula, R w The preset wheelset radius parameter is d, where the subscript d represents the driving torque.

[0090] For a bogie control method where one traction inverter unit controls two traction motors within the same traction bogie, combined with a preset gear ratio R... g and transmission efficiency The equivalent target driving torque on the wheelset side is converted into the equivalent target output torque of a single traction motor within the bogie: , In the formula, the subscript m represents the traction motor side. This equivalent target output torque is used to characterize the torque control target, correction value, or upper limit constraint allowed by the traction inverter unit for the traction bogie under the current adhesion state, and does not indicate that the two traction motors in the same traction bogie have independent torque control commands.

[0091] 6.4 Final Torque Command Output and Closed-Loop Control.

[0092] The traction control unit uses the equivalent target output torque of the aforementioned single traction motor as the torque command limit value for the traction inverter unit corresponding to the traction bogie; when the traction control unit already has the original equivalent motor torque requirement T corresponding to the traction bogie... m,j,b,eq,req When (t), the final equivalent motor torque command is limited to: , In the formula, T m,j,b,eq,cmd (t) represents the equivalent motor torque command that is finally output to the traction inverter unit corresponding to the b-th traction bogie in the \(j\)-th EMU, and the subscript cmd indicates the final command value.

[0093] The traction inverter unit drives the traction motor and transmission system according to the final torque command. The actual torque output by the traction motor, the angular velocity of the wheelset rotation, and other signals are fed back to the operation information acquisition module to form a complete closed-loop control circuit.

[0094] Through the above control logic, when the wheel-rail adhesion is good, the online calculated wheel-rail adhesion force is high, allowing the system to tolerate higher bogie-level target traction force and equivalent target output torque of the traction motor. When the wheel-rail adhesion of any axle within the same traction bogie decreases, the system uses the axle with poor adhesion as the basis for bogie-level traction output constraint, correspondingly reducing or limiting the target traction force of that traction bogie, and further reducing or limiting the traction output of the corresponding traction inverter unit, so that the traction output matches the current longitudinal force transmission state between the wheel and rail. The remaining traction bogies in the entire train that have not experienced adhesion limitation or slippage can continue to output according to their corresponding original target traction force, thereby realizing the active constraint of the traction output of each traction bogie level in traction anti-slip control.

[0095] Although this application describes the embodiment using the b-th traction bogie in the j-th motor car as the basic control object, each motor car and each traction bogie in the entire train can perform online calculation of wheel-rail adhesion state and traction anti-slip control in the same way, thereby actively constraining the traction output before the wheels obviously slip.

[0096] Unlike adhesion sensing methods used only for status display or warning, this embodiment further compares the adhesion force between the wheel and rail of two moving axles within the same traction bogie, calculated online, with the equivalent target traction force of a single moving axle of the traction bogie. The moving axle with poorer adhesion is used as the constraint basis for the overall traction output of the traction bogie, generating a correction value for the equivalent target traction force of a single moving axle. This yields the corrected target traction force at the traction bogie level and its equivalent wheelset-side target driving torque. Subsequently, the system converts this into the equivalent target output torque of a single traction motor through the traction transmission relationship, serving as the torque control target, correction value, or upper limit constraint for the corresponding traction inverter unit of the traction bogie. Therefore, this embodiment can correct the traction output of the corresponding traction bogie based on the wheel-rail adhesion state of each moving car and each traction bogie in the entire train. This allows the traction control system to dynamically adjust the traction output according to the current longitudinal force transmission state between the wheel and rail, thereby reducing the risk of wheelset slippage.

[0097] VII. Vehicle-mounted display and vehicle-to-ground communication.

[0098] The system outputs the real-time calculated wheel-rail adhesion coefficient, wheel-rail adhesion force, adhesion state level, and early warning information to the driver's console display interface, enabling the driver to intuitively grasp the current wheel-rail adhesion level. When low adhesion, extremely low adhesion, or ultra-extremely low adhesion states are detected, the system reminds the driver to take appropriate driving measures through the display interface, audible and visual prompts, or existing vehicle warning methods.

[0099] Meanwhile, the system uploads adhesion status data to the ground dispatch center through vehicle-to-ground communication networks (4G / 5G / Wi-Fi, etc.), which can be used for low adhesion section identification, adhesion status monitoring, operation organization optimization and safety decision support, providing data support for vehicle operation safety monitoring, line operation and maintenance and operation dispatch decisions.

[0100] In addition, it should be emphasized that the role of this application is mainly reflected in two aspects.

[0101] On the one hand, this application enables online sensing of wheel-rail adhesion status during the traction process of rail vehicles. By acquiring real-time operating parameters such as wheelset angular velocity, wheelset vertical load, and motor output torque from the traction system, and converting the motor output torque into driving torque acting on the wheelset through gear transmission, the wheel-rail adhesion force is calculated online based on the wheelset rotational dynamics, further obtaining the wheel-rail adhesion coefficient. This process requires no additional sensors and does not rely on braking actuator parameters such as brake cylinder pressure, brake shoe or brake pad friction characteristics, and can continuously acquire wheel-rail adhesion status information during traction operation. The calculated wheel-rail adhesion coefficient can be used for onboard display, adhesion status classification, low-adhesion section identification and early warning, and can be uploaded to the ground dispatch or operation and maintenance management platform through the vehicle-to-ground communication system, providing data support for vehicle operation safety monitoring, low-adhesion section identification, and operation scheduling decisions.

[0102] On the other hand, this application can also apply the wheel-rail adhesion state information calculated online during traction to traction anti-slip control. The key to traction anti-slip control is to match the traction output of each motor car and each traction bogie with the current longitudinal force transmission state between the wheel and rail of its corresponding axle, so as to avoid a certain axle from spinning due to excessive traction output. Existing anti-slip control usually mainly identifies the slip trend based on signals such as wheelset speed difference, slip ratio, wheelset angular acceleration or impact rate. However, this application can further provide the wheel-rail adhesion force and wheel-rail adhesion coefficient corresponding to two axles in the same traction bogie, so that the traction control system can form traction output constraints at the traction bogie level based on real-time calculation results.

[0103] During the entire train traction process, the train control and management system or the vehicle-level traction control system can allocate traction requirements to each motor car based on the driver's control level, ATO (Automatic Train Operation) requirements, vehicle operating status, traction unit configuration, traction capacity of each motor car, load status, traction system availability status, and fault clearance status. For ease of explanation, this application uses the j-th motor car as an example. Let the original target traction force obtained by the j-th motor car be F. tr,car,j,req (t), the original target traction force is issued by the upper-level traction control system or the vehicle-level traction control system, and this application does not limit its specific distribution method within the entire train.

[0104] For the j-th motor, the traction control unit of this vehicle generates the original target traction force corresponding to each traction bogie of the motor based on the existing traction force distribution strategy, the status of the traction inverter unit, the status of the traction motor, the load status, and the fault torque limiting status.

[0105] During the bogie-level traction output constraint process, the system converts the original target traction force of the traction bogie into an equivalent target traction force corresponding to a single moving axle, and compares it with the calculated wheel-rail adhesion force corresponding to two moving axles within the same traction bogie. When the adhesion state of any moving axle within a traction bogie decreases or a tendency to spin occurs, the system uses the moving axle with the poorer adhesion state within that traction bogie as the basis for traction output constraint, generates a correction value for the equivalent target traction force of a single moving axle, and further obtains the corrected target traction force of the traction bogie.

[0106] Since the two driving axles within the same traction bogie are driven by the traction motor unit corresponding to the same traction inverter unit, the wheel-rail adhesion state of the two driving axles is calculated separately, but the traction output is constrained as a whole at the traction bogie level. When the adhesion state of either driving axle decreases, the system restricts the traction output of the traction inverter unit corresponding to that traction bogie, while the other traction bogies in the entire train that have not experienced adhesion limitation or slippage can continue to output according to their corresponding original target traction force. Therefore, this application can apply the online calculation results of the wheel-rail adhesion state during the traction process to the traction anti-slip control and achieve traction output constraints at the traction bogie level at the execution level.

[0107] The corrected target traction force at the bogie level is further converted into an equivalent target driving torque on the wheelset side, and then, combined with the gear ratio and transmission efficiency, converted into an equivalent target output torque of a single traction motor within the bogie. This serves as the torque control target, correction value, or upper limit constraint for the corresponding traction inverter unit. When wheel-rail adhesion conditions are good, the system can allow higher target traction forces at the bogie level and equivalent target output torques of the traction motors to fully utilize traction capacity. When the adhesion conditions of any driving axle within a traction bogie deteriorate, the system reduces or limits the target traction force of that bogie and the traction output of the corresponding traction inverter unit, adapting the traction output to the current longitudinal force transmission state between the wheel and rail. This reduces the risk of wheel slippage and improves adhesion utilization and vehicle running stability under low adhesion conditions.

[0108] This application utilizes the vehicle's existing speed measurement system, traction control system, and load acquisition device to acquire in real time the wheelset angular velocity, motor output torque, vertical load, vehicle speed, and vehicle parameters during the traction process. Combining gear transmission relationships and wheelset rotation dynamics models, it constructs an online calculation link for wheel-rail adhesion state under traction conditions suitable for real-time on-board applications, enabling real-time acquisition, display, early warning, and uploading of wheel-rail adhesion force, wheel-rail adhesion coefficient, and adhesion state level during the traction process.

[0109] Furthermore, this application also uses the wheel-rail adhesion force and wheel-rail adhesion coefficient calculated online for traction anti-slip control. During the entire train traction process, the train control and management system or the vehicle-level traction control system can allocate traction requirements to each motor car according to the driver's control level, ATO operation requirements, vehicle operating status, traction unit configuration, and traction system availability. For ease of explanation, this application uses the j-th motor car as an example. Let the original target traction force obtained by the j-th motor car be F. tr,car,j,req (t), the original target traction force is issued by the upper-level traction control system or the vehicle-level traction control system, and this application does not limit its specific distribution method within the entire train.

[0110] For the j-th motor car, the traction control unit generates the original target traction force for each traction bogie of the motor car based on the existing traction force distribution strategy, the status of the traction inverter unit, the status of the traction motor, the load status, and the fault torque limiting status. Let F be the original target traction force of the b-th traction bogie in the j-th motor car. tr,j,b,req (t). For trains employing traction bogie-level control, the system acquires the calculated wheel-rail adhesion values ​​for the two driving axles within the same traction bogie, and uses the driving axle with poorer adhesion as the basis for the traction output constraint of that traction bogie, generating the target traction force and its equivalent wheelset-side target driving torque for that traction bogie. Subsequently, the system, combining gear ratio and transmission efficiency, converts the equivalent wheelset-side target driving torque into the equivalent target output torque of a single traction motor within that traction bogie, and uses this as the torque control target, correction value, or upper limit constraint for the corresponding traction inverter unit of that traction bogie, ensuring that the traction output of the traction bogie matches the current longitudinal force transmission state between the wheel and rail, thereby reducing the risk of wheelset slippage. The remaining trains and traction bogies in the entire train can be controlled for traction slippage prevention in the same manner.

Claims

1. A method for online sensing of wheel-rail adhesion state during the traction process of a rail vehicle, characterized in that, The process is executed by an onboard sensing system configured on the rail vehicle. This system communicates with the rail vehicle's traction control unit and is connected to a speed measuring device and a load acquisition device. The rail vehicle has multiple axles, each driven by a corresponding traction motor. Traction motors within the same traction bogie are driven by the same traction inverter unit. The traction control unit generates and distributes the original target traction force to each traction bogie. The method includes: The traction control unit acquires the motor output torque signal from the traction motor in real time, the speed measuring device acquires the wheelset angular velocity signal of the corresponding wheelset of the moving axle, and the load acquisition device acquires the vertical load signal. The on-board sensing system calculates the wheel-rail adhesion value of each moving axle based on the motor output torque signal and the wheel-pair angular velocity signal. The onboard sensing system provides the traction control unit with the calculated wheel-rail adhesion values ​​of each moving axle within the same traction bogie; The traction control unit determines the equivalent target traction force corresponding to a single moving axle in the traction bogie based on the allocated original target traction force, and uses the minimum value among the calculated wheel-rail adhesion values ​​of each moving axle in the traction bogie as a constraint to correct the equivalent target traction force, thus obtaining the corrected equivalent target traction force for a single moving axle. The traction control unit generates the corrected target traction force of the traction bogie based on the corrected equivalent target traction force of a single driving axle, and converts the corrected target traction force into the equivalent target driving torque of the wheelset side, and then into the equivalent target output torque of a single traction motor. This torque command limit value is used as the torque command limit value of the traction inverter unit corresponding to the traction bogie, so as to perform anti-idle torque limiting control on the traction motor of the traction bogie.

2. The method according to claim 1, characterized in that, The rail vehicle includes at least one motor car, the motor car includes at least one traction bogie, the traction bogie includes multiple driving axles, and each driving axle is equipped with wheelsets; the original target traction force is generated by the traction control unit according to the driver control level, ATO operation requirements, vehicle operating status, traction unit configuration, traction capacity of each motor car, load status, traction system availability status and fault clearing status and is distributed to each traction bogie.

3. The method according to claim 1, characterized in that, The onboard sensing system calculates the wheel-rail adhesion values ​​for each axle based on the motor output torque signal and the wheelset angular velocity signal. Specifically, this includes: The onboard sensing system converts the motor output torque signal into a physical quantity of driving torque acting on the wheelset based on the preset gear ratio and preset transmission efficiency. The wheelset rotation dynamics equation is established by the vehicle-mounted sensing system to characterize the relationship between the physical quantity of driving torque, the wheelset angular velocity signal and the physical quantity of wheel-rail adhesion to be calculated; After the vehicle-mounted sensing system performs a Laplace transform on the dynamic equations, the equations are filtered and reconstructed in the complex frequency domain using a first-order low-pass filter. The reconstructed expression is then subjected to an inverse Laplace transform to calculate the wheel-rail adhesion value, thus avoiding the introduction of noise by directly differentiating the wheel-pair angular velocity signal.

4. The method according to claim 3, characterized in that, The onboard sensing system converts the motor output torque signal into a driving torque physical quantity acting on the wheelset based on a preset gear ratio and preset transmission efficiency, specifically according to the following formula: ， Where T(t) is the physical quantity of the driving torque acting on the wheelset. For transmission efficiency, R g T is the gear ratio. m (t) represents the motor output torque signal.

5. The method according to claim 4, characterized in that, The wheelset rotation dynamics equations are established using the onboard sensing system, and Laplace transform, filtering reconstruction, and inverse Laplace transform are performed to calculate the wheel-rail adhesion value. Specifically, this includes: The rotational dynamics equations of the wheelset are established as follows: ， Where J is a preset equivalent rotational inertia parameter for the wheelset. This is the wheel-to-wheel angular velocity signal. For the wheel's angular acceleration, F a (t) represents the physical quantity of wheel-rail adhesion to be calculated, R w Here are the preset wheelset radius parameters, and T(t) is the physical quantity of driving torque; Taking the Laplace transform of this dynamic equation, we get: ， Where Fa(p) is the Laplace transform of Fa(t), and T(p) is the Laplace transform of T(t). for Laplace transform, p represents the initial wheel-pair angular velocity signal at the moment the system begins calculation, where p is the complex frequency; A first-order low-pass filter is introduced into the vehicle-mounted perception system, and its transfer function is: , This is the cutoff angular frequency of a first-order low-pass filter; The viscosity expression is filtered using a first-order low-pass filter to obtain the filtered expression: , Laplace transform of the calculated wheel-rail adhesion force; Reconstruct the filtered expression and define the complex frequency domain auxiliary variable Z(p) to satisfy: ； Make ; Taking the inverse Laplace transform of Z(p), we obtain the time-domain auxiliary state variable z(t) which satisfies the first-order differential equation: , Let z(t) be the rate of change of the time-domain auxiliary state variable with respect to time. The first-order differential equation is solved online to obtain z(t), and the calculated value of the wheel-rail adhesion force is then determined. 。 6. The method according to claim 5, characterized in that, Cutoff angular frequency of a first-order low-pass filter The value range is 5 s −1 up to 100 s −1 Furthermore, the vehicle-mounted sensing system obtains the current vehicle speed from the speed measuring device, and based on at least one of the following: the current vehicle speed, the fluctuation intensity of the wheel-pair angular velocity signal, and the rate of change of the motor output torque signal, the vehicle-mounted sensing system... Perform dynamic segmentation adjustments.

7. The method according to claim 5, characterized in that, Initial wheel angular velocity signal Take the first set of valid wheel-pair angular velocity signal samples at the start of system calculation or the average value over several sampling periods after startup; the initial value z(0) of the time-domain auxiliary state variable is determined according to the following formula: ,in, Take zero at the initial stage of vehicle traction establishment, or take T(0) / R based on the initial driving torque. w T(0) is the physical quantity of the driving torque at the initial moment of calculation; the vehicle-mounted sensing system performs smooth transition processing on the calculated wheel-rail adhesion value within a preset time after startup.

8. The method according to claim 1, characterized in that, The method also includes: The wheel-rail adhesion coefficient is calculated by the onboard sensing system based on the calculated wheel-rail adhesion force and vertical load signal, specifically according to the following formula: ,in, The wheel-rail adhesion coefficient is denoted as . Q(t) is the calculated value of the adhesion force between the wheel and rail, and Q(t) is the vertical load signal. The on-board sensing system determines the current wheel-rail adhesion state level based on the wheel-rail adhesion coefficient and drives the on-board audible and visual warning device to output prompt information.

9. The method according to claim 1, characterized in that, The operations performed by the traction control unit are specifically for the j-th motor car, including: Obtain the original target traction force F of the b-th traction bogie in the j-th train. tr,j,b,req (t), and determine the equivalent target traction force corresponding to a single moving axle within the traction bogie: , where subscript j is the train index, subscript b is the traction bogie index, subscript req represents the original target value, subscript eq represents the equivalent value, and subscript tr represents the traction force; Receive the calculated wheel-rail adhesion values ​​of the two driving axles within the same traction bogie sent by the onboard sensing system. and ,in Let be the calculated value of the wheel-rail adhesion force of the first moving axle in the b-th traction bogie of the j-th motor. The value of wheel-rail adhesion for the second moving axle in the b-th traction bogie of the j-th EMU is given by the subscript a, where a indicates adhesion. Using the minimum value of the calculated wheel-rail adhesion force of each driving axle in the traction bogie as a constraint, the corrected equivalent target traction force of a single driving axle is determined: ； Generate the corrected target traction force for the traction bogie: ； The corrected bogie target traction force is converted into an equivalent wheelset side target driving torque: , where R w The preset wheelset radius parameter, with the subscript d indicating the driving torque; According to the preset gear ratio R g and transmission efficiency The equivalent target driving torque on the opposite wheels is converted into the equivalent target output torque of a single traction motor: The subscript m indicates the traction motor side; The equivalent target output torque of a single traction motor is used as the torque command limit value for the corresponding traction inverter unit of the traction bogie, specifically including: When the traction control unit already has the original equivalent motor torque requirement T corresponding to the traction bogie m,j,b,eq,req When (t), the final equivalent motor torque command is limited to: , among which, T m,j,b,eq,cmd (t) represents the equivalent motor torque command ultimately output to the traction inverter unit corresponding to the b-th traction bogie in the j-th EMU, T m,j,b,eq,req (t) represents the original equivalent motor torque demand generated by the traction control unit, and the subscript cmd indicates the final command value; this is used to perform anti-idle torque limiting control on the traction motor of the traction bogie.

10. An online sensing system for wheel-rail adhesion state during the traction process of a rail vehicle, characterized in that, For implementing the method as described in any one of claims 1 to 9, the online sensing system includes an operation information acquisition module, an online adhesion calculation module, a target value conversion module, a motor target output torque calculation module, and a traction anti-idle slip control module; The operation information acquisition module is used to acquire the motor output torque signal, the wheel pair angular velocity signal, the vertical load signal, the vehicle speed signal, and vehicle parameters, including the wheelset equivalent moment of inertia parameter, wheelset radius parameter, gear ratio, and transmission efficiency. The online adhesion calculation module receives the motor output torque signal, the angular velocity signal of each driving wheel pair, the vertical load signal, and vehicle parameters sent by the operation information acquisition module. Based on the gear ratio and transmission efficiency, it converts the motor output torque signal into a physical quantity of driving torque acting on the wheelset. Based on the rotational dynamics of the wheelset, it calculates the adhesion value between each driving axle and the rail and the wheel-rail adhesion coefficient, and outputs the adhesion value between each driving axle and the rail and the wheel-rail adhesion coefficient to the target value conversion module. The target value conversion module receives the original target traction force of each traction bogie and the calculated wheel-rail adhesion force of each moving axle within the same traction bogie. It uses the minimum of the calculated wheel-rail adhesion force of each moving axle within the same traction bogie as the constraint of the equivalent target traction force of a single moving axle. Based on this constraint, it corrects the equivalent target traction force of a single moving axle determined by the original target traction force, generates the corrected equivalent target traction force of a single moving axle, and further generates the corrected bogie target traction force. It then converts the bogie target traction force into the equivalent wheelset side target driving torque and outputs it to the motor target output torque calculation module. The motor target output torque calculation module is used to receive the equivalent wheel-side target driving torque, convert the equivalent wheel-side target driving torque into the equivalent target output torque of a single traction motor according to the gear ratio and transmission efficiency, and output it to the traction anti-idle control module. The traction anti-slip control module is used to receive the equivalent target output torque of a single traction motor, use it as the torque command limit value of the corresponding traction inverter unit, and output anti-slip control commands to the traction motor and transmission system to perform anti-slip torque limiting control on the traction motor of the traction bogie. The traction motor and transmission system receive anti-idle control commands to execute traction output, and feed back the motor output torque signal, the wheel-to-wheel angular velocity signal, the vertical load signal and the vehicle speed signal to the operation information acquisition module to form a closed-loop control circuit.

Images