Multi-motor dynamic synchronous regulation and control method based on virtual line shaft state feedback

Abstract

The invention discloses a multi-motor dynamic synchronous regulation and control method based on virtual line shaft state feedback, and the method comprises the steps: constructing a multi-motor virtual line shaft control model, setting an expected traction torque value through the output of a virtual main shaft, and setting each physical motor as a following motor; according to the collected state data of each motor, performing joint probability modeling on the state and disturbance of each motor by adopting a preset modeling method, and generating an estimated mean value and a covariance of each motor state; designing a nonsingular terminal sliding mode surface and calculating a control signal; setting a soft saturation buffer area range, constructing a soft switching function, carrying out buffer correction on the control signal, and outputting a corrected control signal; and synchronous control of multi-motor traction output is realized according to the correction control signal. According to the multi-motor dynamic synchronous regulation and control method, the synchronous stability, the anti-interference capability and the output smoothness in a complex dynamic environment can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of motor control, and particularly to a multi-motor dynamic synchronization control method and system based on virtual main shaft state feedback. Background Art

[0002] High-speed trains are an important economic pillar in China today and the preferred means of transportation for most people. Their safety issues cannot be ignored. Traction system control technology is the key to the safe operation control technology of high-speed trains. Referring to Figure 1 as shown, the control hierarchy of the high-speed train traction system can be mainly divided into three levels: train-level control, vehicle-level control, and drive-level control. Train-level control refers to the overall control of the entire train. The instructions issued by the driver are transmitted to each power unit through train-level control to achieve the control of the train's traction, braking, acceleration, deceleration and other operating conditions, and at the same time, the safety of the train is protected and diagnosed. Vehicle-level control mainly performs optimized adhesion control, braking force distribution, and diagnosis and protection, mainly processing the traction force and braking force and sending them to the four-quadrant control and inverter control devices. Drive-level control is the control, diagnosis and monitoring of the four-quadrant control and inverter control devices.

[0003] For a traction locomotive powered by multiple traction motors, maintaining dynamic stability of the total traction power during operation is a prerequisite for ensuring the stable operation of the locomotive. In order to solve the train safety problems caused by sudden wheel slip and other faults of the locomotive wheelset, it is necessary to seek a control strategy so that when a certain wheelset of the locomotive fails, the remaining normal wheelsets can bear the traction power within their tolerable range to avoid the continuous occurrence of faults and ensure that the train can still operate safely. This requires researching a multi-motor traction power total coordination control strategy based on a virtual main shaft.

[0004] Currently, the key to the multi-motor traction power total coordination control strategy based on a virtual main shaft is to utilize the feedback mechanism of the virtual main shaft synchronization control strategy. When the train is running stably, ensure that the total traction power of each motor is consistent with the system given value; when faults such as wheel slip occur, the virtual main shaft can promptly sense the fault through its feedback mechanism, and then adjust the total traction power of the motor, ensuring dynamic balance of the total traction power while avoiding overloading of each motor, and minimizing the adverse effects on the train safety performance caused by faults such as wheel slip.

[0005] However, the existing multi-motor coordination control strategies mainly focus on the consistency of states such as speed and position. Due to the complex system working environment, frequent load disturbances, and inconsistent parameters between motors, it is easy to cause problems such as fluctuations in the multi-motor state synchronization error affecting the coordination consistency of the total traction power, actuator limitations leading to non-linear saturation behavior, and traditional sliding mode control being prone to chattering in the saturation region, reducing stability and response quality.

[0006] Therefore, there is an urgent need for a multi-motor dynamic synchronization control method that can improve synchronization stability, anti-interference ability and output smoothness in complex dynamic environments. Summary of the invention

[0007] Purpose of the invention: In order to overcome the above-mentioned deficiencies, the purpose of the present invention is to provide a multi-motor dynamic synchronous control method and system based on virtual main shaft state feedback, which can improve the output consistency and synchronous control accuracy of multiple motors and enhance the anti-interference ability of the system, and can also realize soft buffer passivation processing of control signals near physical boundaries, effectively suppress system oscillations and weaken nonlinear surge effects.

[0008] In order to solve the above technical problems, the present invention provides a multi-motor dynamic synchronization control method based on virtual main shaft state feedback, comprising: S1: Construct a multi-motor virtual main shaft control model, set the expected traction torque value through the virtual main shaft output and set each physical motor as a follower motor, so as to achieve coordinated traction of the total amount; S2: Based on the collected state data of each motor, a preset modeling method is used to perform joint probability modeling on the state and disturbance of each motor, and an estimated mean and covariance of each motor state is generated; S3: designing a non-singular terminal sliding surface and calculating a control signal according to the multi-motor virtual main shaft control model and the feedback control law; S4: setting a soft saturation buffer range and constructing a soft switching function, performing buffer correction on the control signal and outputting a corrected control signal; S5: Implement synchronous control of multi-motor traction output according to the modified control signal.

[0009] As a preferred embodiment of the present invention, in S1, the virtual main axis in the multi-motor virtual main axis control framework is not an actual physical motor. It sets the traction reference torque trajectory through control logic and feeds back the torque value of each follower motor in real time as virtual main axis load feedback.

[0010] As a preferred embodiment of the present invention, in S1, the method comprises: S11: Set the dynamics of the virtual spindle to: , and then design the control law according to the target trajectory so that the output of the virtual spindle reaches: ,in, is the equivalent moment of inertia of the virtual spindle, is the virtual spindle angular velocity, is the controller output driving torque, is the sum of the traction torques of all follower motors, is the expected traction torque value set; S12: Make the sum of the outputs of all follower motors equal to the set desired traction torque value: , and introduce the total error term: , where is the traction torque of the j-th motor; S13: Set the ability weights of each follower motor: , then the output of each motor is: , and introduce the control error: , is the error between the desired traction torque and the output torque of the j-th motor.

[0011] As a preferred embodiment of the present invention, in S13, using the control error combined with the current estimated state of the motor, dynamically update the ability weights of each follower motor: , , where is the estimated value of the torque error of the j-th motor, is the uncertainty of the estimated motor output torque, is the weight decay factor, is the immediate available ability index of the j-th motor.

[0012] As a preferred embodiment of the present invention, in S2, the method includes: S21: In each control cycle, collect the actual speed, output torque, and control error of each motor, and construct a joint state vector: , where, is the speed of the j-th motor; S22: Use the Bayesian method or the Gaussian process regression method to model the joint state vector: , where is the state estimation mean, , is the state covariance matrix, .

[0013] As a preferred embodiment of the present invention, in S3, the method includes: S31: Define the sliding mode surface: , where, is the error change rate, is the sliding mode surface gain for adjusting the surface morphology, is a positive integer to prevent singularity; S32: Calculate the control signal using the feedback control law: , where is the equivalent control term, , To control the input gain, is a known non-linear model term, is the derivative of the reference trajectory for feedforward compensation; is a compensation term for resisting modeling errors and disturbances, , is the basic fixed gain, is an adaptive adjustment coefficient for weighing the uncertainty strength and control effort, is the total uncertainty of all state variables.

[0014] As a preferred embodiment of the present invention, in S4, the method includes: S41: Define the soft saturation buffer width for judging the vicinity of the boundary transition zone: , and then set the soft saturation buffer range: ; S42: Define the saturation limit value for judging the current vicinity of the boundary: S43: Construct a buffer weight function using a Sigmoid-form function: , where α is the buffer slope control coefficient; S44: Buffer-correct the control signal and output the corrected control signal: , where when is close to the boundary, approaches 1, and the output of the control signal gradually approaches the boundary without overshooting. When is far from the boundary, tends to 0, and the control signal is not affected.

[0015] A multi-motor dynamic synchronization control method based on virtual total axis state feedback, the soft saturation buffer width is set to 5% - 20% of the maximum control signal amplitude, and the saturation limit value is one of the saturation boundaries that the control signal is about to reach. When , When , .

[0016] As a preferred embodiment of the present invention, in S41, the method includes: S411: Extract the total uncertainty of all state variables in the compensation term ; S412: Using the total uncertainty of all state variables as a dynamic index, automatically expand the buffer width , where is the initial buffer width, is the dynamic expansion upper limit, is the adjustment slope factor.

[0017] This application also provides a multi-motor dynamic synchronization control system based on virtual total axis state feedback using the above method, including: An overall framework module for constructing a multi-motor virtual total axis control model, outputting a set expected traction torque value through the virtual main axis and setting each physical motor as a follower motor, so as to achieve consistent coordination of the total traction; A joint distribution module for jointly probabilistically modeling the states and disturbances of each motor according to the collected state data of each motor by using a preset modeling method, and generating the estimated mean and covariance of the state of each motor; A non-singular control module for designing a non-singular terminal sliding mode surface and calculating a control signal according to the multi-motor virtual total axis control model and the feedback control law; A soft saturation buffer module for setting the soft saturation buffer range and constructing a soft switching function, buffering and correcting the control signal and outputting a corrected control signal; A control execution module for realizing synchronous control of multi-motor traction output according to the corrected control signal.

[0018] The above technical solutions of this application have the following advantages compared with the prior art: 1. By constructing a virtual total axis control framework, setting the expected traction output uniformly, and modeling each actual motor as a follower unit, the multi-motor coordinated drive has the ability of global consistent scheduling; and by introducing an ability weight adaptive reallocation mechanism, the motor target output ratio is automatically adjusted in real time according to the motor health state and output stability, improving the dynamic balance ability and fault tolerance performance of the system under the condition of individual motor performance degradation or limitation.

[0019] 2. Adopting non-singular terminal sliding mode design, combining the uncertainty information of system state estimation, and dynamically adjusting the controller gain, so that the system can maintain stable and fast response performance in the case of external disturbance or model deviation aggravation, and avoid the singular point problem of sliding mode control.

[0020] 3. Constructed a soft saturation buffer mechanism based on the Sigmoid function to achieve continuous and smooth signal passivation when the control signal approaches the boundary, effectively suppressing system oscillation, weakening the nonlinear surge effect, prolonging the actuator life, and improving system stability; and taking the joint state covariance matrix as the core index, driving the sliding mode gain adjustment and soft saturation buffer adjustment simultaneously to achieve dynamic adaptation based on uncertainty. Description of the Drawings

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on the provided drawings.

[0022] Figure 1 It is a schematic diagram of the hierarchical control framework of the high-speed train traction system provided in the prior art.

[0023] Figure 2 It is a schematic diagram of the basic working principle of the virtual main shaft provided in the embodiment of the present invention.

[0024] Figure 3 It is a schematic diagram of the control framework of the virtual main shaft provided in the embodiment of the present invention.

[0025] Figure 4 It is a schematic diagram of the phase plane of the sliding mode variable structure control provided in the embodiment of the present invention.

[0026] Figure 5 It is a schematic diagram of the flow of the multi-motor dynamic synchronization regulation method provided in the embodiment of the present invention.

[0027] Figure 6 It is a schematic diagram of the module connection of the multi-motor dynamic synchronization regulation system provided in the embodiment of the present invention.

[0028] Figure 7 It is a schematic diagram of the traction total amount coordination and consistency framework of the virtual main shaft provided in the embodiment of the present invention.

[0029] Figure 8 It is a schematic diagram of the MATLAB simulation model provided in the embodiment of the present invention.

[0030] Figure 9 It is a schematic diagram of the pulse interference signal provided in the embodiment of the present invention.

[0031] Figure 10 It is a schematic diagram of the high-frequency interference signal provided in the embodiment of the present invention.

[0032] Figure 11 It is a schematic diagram of the tracking effect and tracking error of the virtual controller provided in the embodiment of the present invention.

[0033] Figure 12 It is a schematic diagram of the controller effect and tracking error of the total amount coordination and consistency provided in the embodiment of the present invention.

[0034] Figure 13 It is a schematic diagram of the system output torque and tracking error provided in the embodiment of the present invention.

[0035] Figure 14It is a schematic diagram of the system given torque, virtual controller output, and output torque of each motor provided by the embodiments of the present invention.

[0036] Figure 15 It is a schematic diagram of the total torque tracking and tracking error of the system when a sudden load is applied provided by the embodiments of the present invention.

[0037] Figure 16 It is a schematic diagram of the RT-LAB experimental platform provided by the embodiments of the present invention.

[0038] Figure 17 It is a schematic diagram of the tracking effect of the virtual controller provided by the embodiments of the present invention.

[0039] Figure 18 It is a schematic diagram of the effect of the total amount collaborative consistent controller provided by the embodiments of the present invention.

[0040] Figure 19 It is a schematic diagram of the system tracking effect provided by the embodiments of the present invention. Detailed implementation manners

[0041] The embodiments of the present invention will be described in detail below. The examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions from beginning to end. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to explain the present invention, but should not be construed as limiting the present invention.

[0042] The multi-motor synchronous control strategies mainly include parallel control strategy, master-slave control strategy, cross-coupling control strategy, adjacent cross-coupling control strategy, and virtual main shaft control strategy. The parallel control strategy is the earliest proposed control strategy when studying multi-motor synchronous control. All motors in this control strategy run in parallel, ensuring that each motor can receive the command signal and track the signal to run stably. The master-slave control strategy takes one motor as the master motor and the rest as slave motors. The cross-coupling control strategy is to subtract and differentiate the output values of two motors, and then feedback the differential signals to their respective feed-forward channels respectively, and act on the two motors together with the input signals. The adjacent cross-coupling control strategy improves the cross-coupling control strategy, making this control method applicable to multi-motor synchronous control systems. This control strategy differentiates the output values of adjacent motors in the multi-motor system and feedbacks them to their respective feed-forward channels, and acts on the two adjacent motors together with the input signals. The virtual main shaft control strategy ELS (Electronic Line Shafting) introduces an overall closed-loop feedback control on the basis of the master-slave control structure, and replaces the real main shaft by simulating a mechanical main shaft. While feedbacking the speeds or positions of each motor, the torques of each motor are used as the total feedback to act on the virtual main shaft. If a certain motor is disturbed, the virtual main shaft can timely sense this change through torque feedback, and the virtual controller adjusts the output value of the virtual main shaft to adjust the output values of the rest of the motors, so as to achieve the purpose of synchronous control.

[0043] In an actual multi-motor control system, the controller controls the controlled object through the actuator. However, due to its own physical limitations and application environment, etc., the actuator will exhibit a saturation phenomenon, and its output will also be restricted due to saturation. When the actuator saturates, the control signal cannot fully act on the controlled object, and the controlled object cannot give normal feedback, which will lead to a decline in the performance of the controller. At the same time, because the state of the controller changes before and after the actuator saturation, there will be a partial degree of out-of-control inside the system, resulting in a decline in the performance of the control system, too long system adjustment time, and further affecting the accuracy of the entire multi-motor system control. If the actuator is in a saturated state for a long time, it will even cause serious consequences such as system collapse. When the system is running normally, , the output of the controller can fully act on the controlled object; while when the actuator saturates, , at this time the output of the controller cannot fully act on the controlled object, which will lead to phenomena such as system disorder and too long adjustment time.

[0044] The basic working principle of the virtual main shaft control method is as Figure 2As shown, after applying a control signal to the virtual main axis simulated from the slave system, the virtual main axis starts to move and drives n moving axes to move together through the driving torque. At the same time, the load torques borne by the n moving axes are also fed back to the virtual main axis to balance the driving torque of the virtual main axis.

[0045] The core of the virtual main axis control strategy is its feedback mechanism. As Figure 3 can be seen, this closed-loop system consists of the system given signal θ d , the virtual main axis, the controller j (j = 1, 2,..., n), and the multi-motor j (j = 1, 2,..., n). θ ref is the output angular displacement of the virtual main axis, T is the driving torque of the virtual controller, is the load torque of the virtual main axis, θ j (j = 1, 2,..., n) is the output angular displacement of the j-th motor, and T Lj (j = 1, 2,..., n) is the load torque of the j-th motor. While the virtual main axis control structure feeds back the speeds or positions of each slave motor, it can use the total torque of each slave motor as the total feedback to act on the virtual main axis. When the multi-motor system is in a steady state, each slave axis follows the virtual main axis to move, and a good synchronous control effect can be achieved. When one or more slave axes deviate from the reference value output by the virtual main axis due to external interference, the virtual main axis can promptly sense this change through torque feedback, and then the virtual controller adjusts the output value of the virtual main axis to further adjust the output values of other slave axes, thereby achieving the purpose of synchronous control. When the multi-motor control system is stable, the dynamic equilibrium equation of the system is: , where is the moment of inertia of the virtual main axis. From the formula, it can be seen that the virtual main axis control strategy promptly senses and adjusts the output angular displacement of the virtual main axis through total torque feedback, and then adjusts the angular displacements of each slave axis to achieve the effect of rapid synchronization and tracking of the given angular displacement among each slave axis, ensuring the stability and robustness of this control system.

[0046] The non-singular terminal sliding mode variable structure control method uses a variable structure controller to make the system state reach a hyperplane determined by the switching function from any initial position within a finite time, and maintain asymptotic motion on this plane, and this hyperplane is called the sliding mode surface. As Figure 4As shown in the figure, in ordinary sliding mode variable structure control, a linear sliding surface is usually selected so that the system state can reach the sliding surface, i.e., s(x)=0, from any position within a finite time. After that, the state tracking error can asymptotically converge to zero, and its convergence speed can be adjusted by selecting different sliding surface parameters. When the system moves on the sliding surface, it has the invariance of strong anti-interference ability. However, since the state tracking error of the system cannot converge to zero within a finite time when moving on the sliding surface, this leads to the problem of uncertain adjustment time of the system, which will have an adverse impact on the dynamic performance of the control system. To solve the problem of uncertain movement time on the sliding surface, a non-linear function is usually introduced on the sliding plane to construct a terminal sliding surface to ensure that when the system moves on the sliding surface, the state tracking error of the system can converge to zero within a finite time, thereby shortening the system adjustment time and improving the dynamic performance of the control system.

[0047] Thus, referring to Figure 5 As shown in the figure, in some embodiments, the multi-motor dynamic synchronization control method based on virtual main axis state feedback involves the following steps: S1: Construct a multi-motor virtual main axis control model. Set the desired traction torque value through the virtual main axis output and set each physical motor as a follower motor to achieve coordinated consistency of the total traction.

[0048] Among them, the virtual main axis in the multi-motor virtual main axis control framework is not an actual physical motor. It sets the traction reference torque trajectory through the control logic and real-time feedback of the torque values of each follower motor as the virtual main axis load feedback.

[0049] Specifically, in S1, the method includes: S11: Set the dynamics of the virtual main axis as: , and then design the control law according to the target trajectory to make the output of the virtual main axis reach: , where is the equivalent moment of inertia of the virtual main axis, set by the designer; is the angular velocity of the virtual main axis, is the driving torque output by the controller, is the sum of the traction torques of all follower motors, is the set desired traction torque value.

[0050] S12: Make the sum of the outputs of all follower motors equal to the set desired traction torque value , that is: , and introduce the total error term: , where is the traction torque of the jth motor; among them, the task of total coordinated consistency control is to make: .

[0051] S13: To ensure the total quantity is consistent, it is also necessary to design the distribution mechanism of each motor and set the ability weight of each follower motor: Then, the expected reference output of each motor is: And introduce the control error: , is the error between the expected traction torque and the output torque of the j-th motor; among them, the design goal of the controller is to make each →0, so as to achieve local tracking plus global total quantity consistency.

[0052] In this embodiment of the present application, the weight can be set according to the motor specifications, current limit, thermal load capacity, etc.; thus, the above overall process is: set the target traction torque value , construct a virtual motor state model and feedback the sum of the outputs of each motor, calculate the reference output of each motor and introduce the control error .

[0053] S2: According to the collected state data of each motor, use a preset modeling method to jointly probabilistically model the state and disturbance of each motor, and generate the estimated mean and covariance of each motor state.

[0054] Specifically, in S2, the method includes: S21: In each control cycle, collect the actual speed, output torque, and control error of each motor, and construct a joint state vector: , where is the speed of the j-th motor; S22: Use the Bayesian method or the Gaussian process regression method to model the joint state vector: , where is the state estimation mean, , is the state covariance matrix, .

[0055] Among them, in a small-scale system with low-dimensional states, the Bayesian filtering method, such as the extended Kalman filter (EKF) or the unscented Kalman filter (UKF) method, can be used to recursively estimate the system state to generate the system state equation: , , where is the motor state transition matrix, which can be empirically modeled or linearly modeled; is the control input matrix, is the measurement output matrix, is the system process noise, which satisfies the Gaussian distribution; To measure noise, satisfying ∼N(0, R j ). EKF / UKF estimates the joint state mean in real time through a two-step loop from prediction to update and covariance .

[0056] Among them, when used for strongly nonlinear and discrete sampling systems, Gaussian process regression (GPR) is adopted, which is a non-parametric Bayesian model. The state variables are used as multi-dimensional inputs to regress the mapping relationship between the internal states of the system: Among them, is the mean function, usually set to 0 or a linear function; is the kernel function, and the commonly used RBF (radial basis function) kernel is used; is the Gaussian process distribution, which is used to generate the state prediction distribution; Using GPR does not require an accurate system model, and directly predicts the state distribution based on historical observations.

[0057] Thus, whether using EKF, UKF or GPR, the joint state distribution is finally obtained: .

[0058] Thus, the overall process described above is as follows: A state estimation sub-module is set for each motor in the control system; The state is collected within the control period: , using the EKF or GPR model, to obtain .

[0059] S3: Design a non-singular terminal sliding mode surface and calculate the control signal according to the multi-motor virtual main shaft control model and the feedback control law.

[0060] Specifically, in S3, the method includes: S31: Define the sliding mode surface: , where is the error change rate, is the sliding mode surface gain used to adjust the surface morphology, is a positive integer to prevent singularity, satisfying ; Control the nonlinear attenuation rate of the error, and the sliding mode surface is non-singular, that is, there will be no problem of dividing by zero or infinity near , which can ensure that the error converges to zero within a finite time.

[0061] S32: Calculate the control signal with the feedback control law: , where is the equivalent control term, which is used to cancel the known dynamic part of the system, , is the control input gain, is a known non - linear model term (such as back - electromotive force, load), is the derivative of the reference trajectory for feed - forward compensation.

[0062] is a compensation term for resisting modeling errors and disturbances, and the gain is: , is the basic fixed gain, is an adaptive adjustment coefficient for weighing the intensity of uncertainty and the control effort, is the total amount of uncertainty of all state variables; the gain size is automatically adjusted according to the change of state uncertainty, so as to achieve adaptive robust control.

[0063] Thus, the above overall process is as follows: within the control period, the state estimator outputs the joint distribution, extracts , calculates the state uncertainty index , calculates the error , its difference change rate , and constructs the sliding mode surface , constructs the equivalent control term , adjusts the robust compensation term according to to form the control signal .

[0064] S4: Set the soft - saturation buffer range and construct a soft - switching function to buffer and correct the control signal and output the corrected control signal.

[0065] Among them, in a multi - motor or actuator system, the control signal (such as voltage, current, PWM duty cycle) has physical output limitations. When the controller output approaches or exceeds these limits, it is easy to generate: actuator saturation, where the output cannot continue to increase; drastic system non - linear changes, exciting high - frequency oscillations; sudden changes in the control signal, resulting in a decline in the dynamic response quality.

[0066] Specifically, in S4, the method includes: S41: Define the soft - saturation buffer width for judging the boundary transition region: , and then set the soft - saturation buffer range: ; where the soft - saturation buffer width is set to 5% - 20% of the maximum control signal amplitude.

[0067] S42: Define the saturation limit value for judging the current proximity to the boundary: where the saturation limit value is one of the saturation boundaries that the control signal output is about to reach. When When When When then

[0068] S43: Construct the buffer weight function using the Sigmoid form function: where α is the buffer slope control coefficient; when i.e., when approaching the boundary, ; when i.e., when approaching the boundary, then

[0069] S44: Buffer-correct the control signal and output the corrected control signal: where when is close to the boundary, tends to 1, and the output of the control signal gradually approaches the boundary without overshooting. When is far from the boundary, tends to 0, and the control signal is not affected.

[0070] Thus, the overall process described above is: The controller generates the original control signal, determines whether the output is close to the saturation range, calculates , generates the transition weight, outputs the corrected control signal , and drives the actuator to execute the corrected control signal then

[0071] S5: Implement synchronous control of multi-motor traction output according to the corrected control signal.

[0072] Specifically, in this application, a multi-motor virtual main shaft control model is used as the overall framework, and distributed state estimation, nonsingular terminal sliding mode, and soft saturation buffer regulation design are introduced therein. Each motor control link is uniformly scheduled through the virtual main shaft, and the soft saturation area dynamically regulates the amplitude of the control signal to avoid drastic changes in control nonlinearity.

[0073] In some embodiments of this application, in S13, the control error is combined with the current estimated state of the motor to dynamically update the ability weight of each follower motor: , , where is the estimated value of the torque error of the j-th motor, is the uncertainty of the estimated motor output torque; is the weight decay factor, which is used to adjust the influence of errors and uncertainties on the ability evaluation; is the immediate available ability index of the j-th motor.

[0074] Specifically, for the j-th motor, the torque error estimation value of the motor is used and the uncertainty estimation value of its output torque to construct its immediately available capacity index . The larger the motor error and the more uncertain the estimation, the smaller the capacity index , indicating that it is not suitable to undertake more output tasks at present; furthermore, after obtaining the capacity indices of all motors , calculate the capacity weight of each motor , and then the traction output target of each motor in the current cycle can be updated to .

[0075] Thus, when the output error of a certain motor is large or the state uncertainty is high, its capacity weight automatically decreases, thereby reducing its load task and preventing the abnormal performance of the motor from affecting the overall system; while for the motor with reliable state estimation and small error, its capacity index is high, and it automatically undertakes more output tasks to maintain the stability of the total traction of the system; realizing the load adaptive reallocation of the multi-motor system under dynamic working conditions.

[0076] In some embodiments of the present application, in S41, the method includes: S411: Extract the total uncertainty of all state variables in the compensation term .

[0077] S412: Using the total uncertainty of all state variables as a dynamic index, automatically expand the buffer width , where is the initial buffer width, usually set to 5% of the maximum control amplitude; is the dynamic expansion upper limit, usually set to 10% - 15% of the maximum control amplitude; is the adjustment slope factor, which controls the sensitivity of uncertainty to expansion, usually set to 5 - 15; thus when the state uncertainty is small ; when the uncertainty increases automatically expands towards , the buffer increases in width, suppressing the risk of signal jumps when the control signal approaches the boundary.

[0078] Specifically, through the joint application of this strategy and the compensation term , a dual-channel adaptive regulation mechanism based on state estimation uncertainty is realized. Thus, when the system faces severe motor state disturbances, increased estimation errors or sensor fluctuations, the buffer will be appropriately expanded to avoid actuator overshoot or jitter caused by the sudden change of the control signal into the saturation region.

[0079] ​In some embodiments, referring to Figure 6 as shown, there is also provided a multi-motor dynamic synchronization control system based on virtual main shaft state feedback using the above method, including: An overall framework module 101, configured to construct a multi-motor virtual main shaft control model, output a set expected traction torque value through the virtual main shaft, and set each physical motor as a follower motor, so as to achieve the consistency of the total traction amount; A joint distribution module 102, configured to perform a joint probability modeling on the state and disturbance of each motor according to the collected state data of each motor by using a preset modeling method, and generate an estimated mean value and covariance of the state of each motor; A non-singular control module 103, configured to design a non-singular terminal sliding mode surface and calculate a control signal according to the multi-motor virtual main shaft control model and a feedback control law; A soft saturation buffer module 104, configured to set a soft saturation buffer range and construct a soft switching function, buffer and correct the control signal, and output a corrected control signal; A control execution module 105, configured to implement synchronous control of multi-motor traction output according to the corrected control signal.

[0080] In some embodiments, there is also provided a computer medium, on which a computer program is stored, and the computer program is executed by a processor to implement the above multi-motor dynamic synchronization control method based on virtual main shaft state feedback.

[0081] In some embodiments, there is also provided a computer, including the above computer medium.

[0082] Thus, referring to Figure 7 and Figure 8 as shown. In order to verify the effectiveness of the synchronous control method described in this application, mathematical simulation and semi-physical simulation experiments were carried out. The motor parameters used in the simulation and experiment are shown in the following table: In the table, R is the motor resistance, L is the inductance, b eq is the equivalent viscous friction coefficient, J eq is the equivalent motor moment of inertia, k m is the motor torque coefficient, k e is the back electromotive force constant, k t is the gearbox transmission ratio. In the simulation and experiment, motor 1 with the smallest moment of inertia was selected as the virtual main shaft.

[0083] To verify the control performance of the multi-motor system under acceleration, constant speed, and deceleration states, a time-varying reference instruction signal is given to the system. When time t < 0.3 s, the multi-motor system starts and is in the acceleration state; when 0.3 s ≤ t ≤ 0.7 s, the multi-motor system is in the constant speed state; when 0.7 s < t ≤ 1 s, the multi-motor system is in the deceleration state. Assume the motor starts with parameter perturbation, that is, the system composite disturbance d 2j ≠ 0. At 0.2 s, a pulse disturbance signal as shown in Figure 9 is applied to motor 2; at 0.5 s, a high-frequency disturbance signal as shown in Figure 10 is applied to motor 3. After the motor is disturbed, the control effect of the system is represented by Figures 11 to 14 , where Figure 11 represents the tracking effect of the virtual controller and its tracking error; Figure 12 represents the tracking effect of the total amount coordinated controller and its tracking error; Figure 13 represents the output torque of each motor in the system and the total traction torque; the figure represents the system tracking error; Figure 14 represents the system given torque, virtual total shaft output torque, and output torque diagram of each motor.

[0084] Analysis Figure 11 shows that the tracking error between the given reference torque and the output torque of the virtual main shaft appears at 0.2 s and 0.5 s. When the system undergoes uncertain disturbances, its error value does not exceed 0.2%, and the adjustment time of the virtual controller does not exceed 0.002 s. Analysis Figure 12 shows that when the system is disturbed at 0.2 s and 0.5 s, the total amount coordinated tracking error value does not exceed 0.02%, and the adjustment time of the total amount coordinated controller does not exceed 0.0002 s. Analysis Figure 13 shows that when motor 1 is disturbed by a pulse signal at 0.2 s, its output torque decreases. To ensure the total amount coordination of the traction power, and under the action of the virtual total shaft, the output torques of motor 2 and motor 3 increase within a certain range. When motor 2 is disturbed by a high-frequency signal at 0.5 s, the output torque of motor 2 decreases. Similarly, under the action of the virtual total shaft, the output torques of motor 1 and motor 3 increase within a certain range. Limited by the braking instruction, the output torques of the three motors all show a downward trend after 0.8 s. After 0.97 s, the output torques of motor 1 and motor 2 are zero. At this time, motor 3 adjusts its output torque to reach the given torque value to maintain the total amount coordination of the traction power. Analysis Figure 13 shows that under different signal disturbances, the system tracking error value does not exceed 0.2%, and the tracking time does not exceed 0.002 s. Analysis Figure 14It can be seen that during the operation of the system, the sum of the output torques of the three motors is equal to the output torque value of the virtual controller and the system's given torque value. From the above simulation results, it can be seen that this application can ensure that the multi-motor system still has good robustness and convergence performance under parameter perturbation and unknown disturbances, and the obtained simulation waveforms are consistent with the theoretical analysis.

[0085] To enable the system simulation to simulate the actual working conditions, based on the above simulation, it is assumed that the train is running stably in the constant-speed operation section as Figure 13 shown. At t = 0.5 s, 50% of the original load is added and the duration is 0.2 s. Figure 15 This is the waveform diagram of the system's given torque and the total output torque of the motor, as well as the error value between the two. It can be Figure 15 seen that when the load is suddenly added, the total output torque value of the motor decreases and then quickly rebounds to the system's given torque value. Its error value does not exceed 2%, and the controller's adjustment time does not exceed 0.004 s. The simulation results show that this application can ensure that the system has strong anti-load disturbance ability.

[0086] To make the simulation process of the system as close as possible to the actual engineering environment, a hardware-in-the-loop simulation experiment is carried out. The equipment used in the experiment is Figure 16 the RT-LAB experimental platform shown. This experimental platform consists of a DSP controller, an OP56000 simulator, connecting wires, and a host computer. During the experiment, first, the multi-motor traction power total coordination control system model based on the virtual main shaft built in the Simulink environment is input into the RT-LAB experimental platform and run on the OP56000 simulator. Then, the multi-motor traction power total coordination controller model designed in this chapter is input into the DSP controller, and finally, the output results are obtained by running. The same parameters and conditions as those in the above mathematical simulation process are selected for the experiment.

[0087] Analysis Figure 17 shows that the output torque of the virtual main shaft can well track the system's given torque. Analysis Figure 18 shows that this controller can ensure that the total output torque value of the system's multi-motors is consistent with the output torque value of the virtual main shaft. Analysis Figure 19 shows that under the action of the two controllers, the total output torque value of the system's multi-motors can be consistent with the system's given torque value. Thus, Figures 17 to 19 it demonstrates the consistency between the hardware-in-the-loop simulation experiment and the MATLAB simulation results.

[0088] In the description of this specification, the descriptions referring to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples", etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0089] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A multi-motor dynamic synchronization control method based on virtual main shaft state feedback, characterized in that It includes the following steps: S1: Construct a multi-motor virtual main shaft control model, output the set expected traction torque value through the virtual main shaft and set each physical motor as a follower motor, so as to achieve the coordination and consistency of the total traction; S2: According to the state data of each motor collected, use a preset modeling method to perform joint probability modeling on the state and disturbance of each motor, and generate the estimated mean and covariance of each motor state; S3: Design a non-singular terminal sliding mode surface and calculate the control signal according to the multi-motor virtual main shaft control model and the feedback control law; S4: Set the soft saturation buffer range and construct a soft switching function to buffer and correct the control signal and output the corrected control signal; S5: Realize the synchronous control of multi-motor traction output according to the corrected control signal.

2. The multi-motor dynamic synchronization regulation method based on virtual main shaft state feedback according to claim 1, wherein In S1, the virtual main shaft in the multi-motor virtual main shaft control framework is not an actual physical motor. It sets the traction reference torque trajectory through the control logic and real-time feedbacks the torque values of each follower motor as the virtual main shaft load feedback.

3. A multi-motor dynamic synchronization control method based on virtual main shaft state feedback according to claim 1 or 2, characterized in that In S1, the method includes: S11: Set the dynamics of the virtual spindle to: , and then design a control law according to the target trajectory to make the output of the virtual spindle reach: , where is the equivalent moment of inertia of the virtual spindle, is the angular velocity of the virtual spindle, is the driving torque output by the controller, is the total traction torque of all follower motors, is the set desired traction torque value; S12: Make the sum of the outputs of all the following motors equal to the set desired traction torque value: , and introduce a total error term: , where is the traction torque of the j-th motor; S13: Set the ability weight of each follower motor: , then the output of each motor is: , and introduce the control error: , is the error between the expected traction torque and the output torque of the j-th motor.

4. A multi-motor dynamic synchronization control method based on virtual main shaft state feedback according to claim 3, characterized in that In S13, the ability weights of each follower motor are dynamically updated by combining the control error with the current estimated state of the motor: , , where is the estimated value of the torque error of the j-th motor, is the uncertainty of the estimated motor output torque, is the weight decay factor, is the immediate available ability index of the j-th motor.

5. A multi-motor dynamic synchronization control method based on virtual main shaft state feedback according to claim 3, characterized in that, In S2, the method includes: S21: In each control period, collect the actual speed, output torque, and control error of each motor, and construct a joint state vector: , where is the speed of the j-th motor; S22: Use the Bayesian method or the Gaussian process regression method to model the joint state vector; , where is the state estimation mean, , is the state covariance matrix, .

6. A multi-motor dynamic synchronization control method based on virtual main shaft state feedback according to claim 5, characterized in that In S3, the method includes: S31: Define the sliding mode surface: , where is the error change rate, is the sliding mode surface gain used to adjust the surface morphology, is a positive integer to prevent singularity; S32: Calculate the control signal with the feedback control law: , where is the equivalent control term, , To control the input gain, is a known non-linear model term, is the derivative of the reference trajectory for feedforward compensation; is the compensation term for resisting modeling errors and disturbances, , is the basic fixed gain, is the adaptive adjustment coefficient for weighing the uncertainty intensity and control effort, is the total amount of uncertainty of all state variables.

7. A multi-motor dynamic synchronization control method based on virtual main shaft state feedback according to claim 6, characterized in that In S4, the method includes: S41: Define the width of the soft saturation buffer for determining the proximity to the boundary transition region: , and then set the soft saturation buffer range: ; S42: Define the saturation limit value used to judge the current proximity to the boundary; S43: Construct a buffer weight function using a Sigmoid-form function: , where α is a buffer slope control coefficient; S44: Buffer and correct the control signal and output the corrected control signal; , where when is close to the boundary, tends to 1, and the output of the control signal gradually approaches the boundary without overshoot. When is far from the boundary, tends to 0, and the control signal is not affected.

8. A multi-motor dynamic synchronization control method based on virtual main shaft state feedback according to claim 1 or 7, characterized in that The width of the soft saturation buffer is set to 5% - 20% of the maximum control signal amplitude, and the saturation limit value is one of the saturation boundaries that the control signal is about to reach. When time, When time, .

9. A multi-motor dynamic synchronization control method based on virtual main shaft state feedback according to claim 7, characterized in that, In S41, the method includes: S411: Extract compensation terms The total uncertainty of all state variables ; S412: Automatically expand the buffer width using the total uncertainty of all state variables as the dynamic index, where is the initial buffer width, is the dynamic expansion upper limit, and is the adjustment slope factor. ​ 10. A multi-motor dynamic synchronization control system based on virtual main shaft state feedback using the method according to any one of claims 1-9, characterized in that It includes: The overall framework module is used to construct a multi-motor virtual main shaft control model, output the set expected traction torque value through the virtual main shaft and set each physical motor as a follower motor, so as to achieve the coordination and consistency of the total traction; The joint distribution module is used to perform joint probability modeling on the state and disturbance of each motor according to the state data of each motor collected, and generate the estimated mean and covariance of each motor state; The non-singular control module is used to design a non-singular terminal sliding mode surface and calculate the control signal according to the multi-motor virtual main shaft control model and the feedback control law; The soft saturation buffer module is used to set the soft saturation buffer range and construct a soft switching function to buffer and correct the control signal and output the corrected control signal; The control execution module is used to realize the synchronous control of multi-motor traction output according to the corrected control signal.

Images