Multi-permanent magnet synchronous motor speed consistency control method and device and electronic equipment

Abstract

The present application relates to the technical field of permanent magnet synchronous motor cooperative control, and particularly relates to a multi-permanent magnet synchronous motor speed consistency control method and device and electronic equipment. The method comprises: establishing a compound fault mathematical model; defining a speed consistency error of the compound fault mathematical model, combining a preset time-varying funnel boundary function to construct a preset time funnel error function; designing a virtual control law, introducing a preset time nonlinear filter, the filter taking the virtual control law as input, outputting a filtered signal, defining a current subsystem error and a filter error according to the current state in the compound fault mathematical model and the filtered signal; and deriving a distributed preset time consistency control law according to the compound fault mathematical model, the preset time funnel error function, the current subsystem error and the filter error. The present application can solve the problems of the prior art, such as inability to independently set the convergence time, lack of explicit transient constraint, high computational complexity and insufficient ability to handle compound faults.

Description

Technical Field

[0001] This invention relates to the field of permanent magnet synchronous motor cooperative control technology, and in particular to a method, device and electronic equipment for controlling the speed consistency of multiple permanent magnet synchronous motors. Background Technology

[0002] Permanent magnet synchronous motors (PMSMs) are widely used in high-end equipment such as new energy vehicles, high-speed trains, CNC machine tools, and robots due to their high efficiency, high power density, and dynamic response. A coordinated drive system consisting of multiple PMSMs must strictly maintain speed synchronization and load balance; its control performance directly affects the equipment's machining accuracy, operational quality, and service life.

[0003] The consensus theory of multi-agent systems (MAS) provides a distributed technical path for multi-motor cooperative control. Existing research has developed various consensus protocols and cooperative strategies, combining backstepping, finite / fixed-time control, and funnel constraints, achieving certain results under ideal operating conditions or single-fault conditions. These methods mainly describe the communication topology through graph theory, transforming the synchronization problem into a consensus convergence problem, and using Lyapunov theory to ensure system stability.

[0004] However, existing technologies still have common limitations: the upper bound of convergence time usually depends on the initial state or controller parameters, making it difficult to preset independently according to the process cycle; most methods only focus on steady-state error, lacking explicit settable boundaries for transient overshoot, oscillation amplitude, and other aspects of the entire process; the backstepping method's repeated differentiation of the virtual control law easily leads to an increase in computational load, making it difficult to implement in real time in embedded controllers; and most solutions only address single types of faults, with significantly insufficient fault tolerance for multiplicative / additive concurrent compound faults in actuators. Summary of the Invention

[0005] This invention provides a method, device, and electronic equipment for controlling the speed consistency of multiple permanent magnet synchronous motors, in order to solve the problems in the prior art such as the inability to independently set the convergence time, lack of explicit transient constraints, high computational complexity, and insufficient ability to handle complex faults.

[0006] In a first aspect, embodiments of the present invention provide a method for speed consistency control of multiple permanent magnet synchronous motors, comprising: Establish a mathematical model for composite faults in a multi-PMSM system; Define the rotational speed consistency error of the composite fault mathematical model, and construct a preset time funnel error function based on the rotational speed consistency error and the preset time-varying funnel boundary function to apply full-process performance constraints to the multi-PMSM system error. A virtual control law is designed, and a preset time nonlinear filter is introduced. The filter takes the virtual control law as input and outputs a filtered signal. Based on the current state in the composite fault mathematical model and the filtered signal, the current subsystem error and the filter error are defined. Based on the mathematical model of the composite fault, the preset time funnel error function, the current subsystem error, and the filter error, a distributed preset time consistency control law is derived to counteract the effects of composite faults and unknown external disturbances.

[0007] Secondly, embodiments of the present invention provide a speed consistency control device for multiple permanent magnet synchronous motors, comprising: The composite fault mathematical model construction module is used to build composite fault mathematical models for multi-PMSM systems; The preset time funnel error function construction module is used to define the rotational speed consistency error of the composite fault mathematical model, and construct the preset time funnel error function based on the rotational speed consistency error and the preset time-varying funnel boundary function, so as to apply full-process performance constraints to the multi-PMSM system error. An error construction module is used to design a virtual control law and introduce a preset time nonlinear filter. The filter takes the virtual control law as input and outputs a filtered signal. Based on the current state in the composite fault mathematical model and the filtered signal, the current subsystem error and the filter error are defined. The consistency control law construction module is used to derive a distributed preset time consistency control law based on the composite fault mathematical model, the preset time funnel error function, the current subsystem error, and the filter error, so as to counteract the effects of composite faults and unknown external disturbances.

[0008] Thirdly, embodiments of the present invention provide an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method described in the first aspect or any possible implementation thereof.

[0009] This invention provides a method, device, and electronic equipment for speed consistency control of multiple permanent magnet synchronous motors (PMSMs). By establishing a composite fault mathematical model of the multiple PMSM system, effective fault-tolerant control is achieved under the concurrent action of multiplicative and additive faults of actuators in the multi-PMSM cooperative drive system, significantly improving the robustness of the system under complex composite fault conditions. Then, the speed consistency error of the composite fault mathematical model is defined, and based on the speed consistency error and a preset time-varying funnel boundary function, a preset time funnel error function is constructed to impose full-process performance constraints on the errors of the multiple PMSM system. This ensures that the transient overshoot, convergence rate, and steady-state accuracy of the error are always limited within the user-preset safety boundaries, effectively avoiding mechanical shocks caused by excessive transient overshoot. The method amplifies the impact and synchronization deviation; it designs a virtual control law and introduces a preset time nonlinear filter. The filter takes the virtual control law as input and outputs a filtered signal. Based on the current state and the filtered signal in the composite fault mathematical model, the current subsystem error and filter error are defined. The preset time nonlinear filter replaces the analytical differentiation of the virtual control law with a smoothed filtered signal, fundamentally eliminating the "computational complexity explosion" problem in the backstepping method design, significantly reducing the computational complexity of the control algorithm, and enabling real-time implementation in resource-constrained embedded controllers. Based on the composite fault mathematical model, the preset time funnel error function, the current subsystem error, and the filter error, a distributed preset time consistency control law is derived to counteract the effects of composite faults and unknown external disturbances. The convergence time of this method can be independently preset by the user and is not affected by the coupling of the system initial state and controller parameters. It can meet the rigid requirements of "timed convergence" for high-end equipment and achieve deep coupling between preset time performance and full-process performance constraints. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a flowchart illustrating the implementation of the speed consistency control method for multiple permanent magnet synchronous motors provided in this embodiment of the invention. Figure 2 (a) is a schematic diagram comparing the rotational speeds of four PMSMs within a preset time according to an embodiment of the present invention; Figure 2 (b) is a schematic diagram comparing the consistency error of four PMSMs within a preset time according to an embodiment of the present invention. Figure 2(c) is a schematic diagram comparing the tracking errors of four PMSMs within a preset time according to an embodiment of the present invention; Figure 2 (d) is a schematic diagram comparing the root mean square error (RMSE) of four PMSMs tracking speeds at different time periods provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of the multi-permanent magnet synchronous motor speed consistency control device provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0012] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0013] See Figure 1 The diagram illustrates a flowchart of a speed consistency control method for multiple permanent magnet synchronous motors provided by an embodiment of the present invention, which is described in detail below: Step 101: Establish a mathematical model of composite faults in a multi-PMSM system.

[0014] This step aims to establish a mathematical model that can accurately characterize the concurrent failure characteristics of actuators in a multi-permanent magnet synchronous motor cooperative drive system under complex operating conditions, providing a foundation for subsequent controller design.

[0015] First, the collaborative drive system composed of multiple PMSMs is modeled as a multi-agent system. Each PMSM and its local controller are abstracted as an agent node with autonomous decision-making and communication capabilities. The information interaction topology between agents is described using a directed or undirected graph. Let the set of agents be... , The total number of PMSMs; the communication topology of a multi-PMSM system is determined by an adjacency matrix. Characterization, in which >0 indicates the first Taiwan PMSM can receive the first Information from Taiwan PMSM, otherwise =0. To achieve tracking of the desired speed command, a virtual leader node is introduced, with a reference speed of... The communication weight between each follower and the leader is denoted as . If the first If the Taiwan PMSM can receive reference instructions from the leader, then >0, otherwise =0.

[0016] Secondly, regarding the first Taiwan PMSM, in A second-order dynamic model is established under the =0 control strategy. This model comprehensively considers the electrical and mechanical dynamics of the motor, and its expression is: (1) in, For the first The electric angular velocity of the PMSM in Taiwan, for The rate of change; For the first The q-axis current of the PMSM, for The rate of change; The torque coefficient, For rotational inertia, The coefficient of viscous friction is... This is the load torque; This is the q-axis voltage. For PMSM pole log number, It is a permanent magnet flux linkage. For stator inductance, Stator resistance; For the first The actual measurement output of the PMSM, here is the electric angular velocity. .

[0017] Further consideration should be given to the common complex faults found in actuators in practical engineering, including the concurrent effects of multiplicative and additive faults. Multiplicative faults... Characterizes the degree of efficiency decay of the actuator, satisfying The physical causes of this include a decrease in voltage gain due to aging of inverter power devices and an increase in the on-state voltage drop of the Insulated Gate Bipolar Transistor (IGBT) module. This type of fault causes the actual output amplitude of the actuator to be less than the ideal command amplitude of the controller. Additive faults. The additional bias superimposed on the actuator output represents the physical source of factors such as inverter dead-time effect, zero drift, and voltage vector distortion caused by switching delay. These faults result in non-ideal DC bias or harmonic components in the actuator output. Under the concurrent action of these two faults, the actual faulty output of the actuator can be expressed as: ;in, For the first The ideal control signal for the PMSM controller design is the desired q-axis voltage command.

[0018] Integrating the above composite fault model into the PMSM second-order dynamic model, the composite fault mathematical model is a model established using graph theory topology that includes the concurrent effects of actuator additive faults and multiplicative faults. After simplification, the composite fault mathematical model is as follows: (2) in, Indicates the first The rotational speed of the PMSM, that is, the first Electric angular velocity of Taiwan PMSM , Indicates the first The rate of change of the rotational speed of the PMSM Indicates the first q-axis current of a PMSM , Indicates the first The rate of change of the q-axis current of the PMSM. Indicates the speed loop gain. Indicates the current loop gain. This indicates the nonlinear term of the speed loop. This represents the nonlinear term of the current loop. Indicates the first The actuator of the PMSM actually has a faulty output. Indicates the first Actual measurement output of the PMSM unit. Indicates the first A multiplicative fault in the actuator of the PMSM unit. Indicates the first The ideal control signal for the controller design of the PMSM is... Indicates the first The actuator of the PMSM has an additive fault.

[0019] in, , ; Including factors such as viscous friction and load torque, It includes influencing factors such as back electromotive force and resistance voltage drop.

[0020] This mathematical model accurately characterizes the composite fault characteristics of the concurrent multiplicative and additive faults of actuators in a multi-PMSM cooperative drive system, as well as the inherent nonlinear dynamics within the system. This provides an accurate model foundation for the design of distributed preset time consistency control laws under preset time funnel constraints.

[0021] Step 102: Define the speed consistency error of the composite fault mathematical model, and construct a preset time funnel error function based on the speed consistency error and the preset time-varying funnel boundary function to apply full-process performance constraints to the multi-PMSM system error.

[0022] This step aims to define the rotational speed consistency error of the multi-agent system and construct a preset time funnel error function to achieve full-process performance constraints on the rotational speed synchronization error of the multi-PMSM cooperative drive system, ensuring that both the transient and steady-state behavior of the error are limited within the preset safety boundaries.

[0023] First, define the first Rotational speed consistency error of PMSM This error takes into account the first... The relative speed deviation between the PMSM and its communication neighbors, and the absolute speed deviation between the PMSM and the virtual leader, constitute the speed consistency error: (3) in, Indicates the first Rotational speed consistency error of the PMSM unit The weights of the adjacency matrix represent the communication topology. Indicates the first Taiwan PMSM can receive the first Information from Taiwan PMSM, otherwise ; Indicates the first Actual measurement output of the PMSM unit. , This indicates the total number of PMSMs in a multi-PMSM system. This indicates the communication weight between followers and leaders. Indicates the first The Taiwan PMSM can receive reference instructions from the leader; otherwise... , This indicates the reference rotation speed of the virtual leader.

[0024] The first term of the formula Characterizing the first The sum of the relative speed deviations between the PMSM and all its communication neighbors; this term prompts the follower PMSMs to achieve speed coordination and consistency through information exchange; the second term... Characterizing the first The absolute speed deviation between the PMSM and the virtual leader ensures that the system can track externally desired speed commands. When the speed approaches zero, the rotational speeds of each PMSM reach a consensus and track the reference command.

[0025] In obtaining speed consistency error Based on this, a preset time funnel error function is further constructed. In order to Apply time-varying boundary constraints. The preset time funnel error function is: (4) in, This indicates the preset time funnel error. This is represented as a time-varying funnel boundary function. The time-varying function pre-designed for the user, with its initial value And within the user-preset convergence time Internal monotonic decay to steady-state accuracy That is, satisfying Through this boundary function, the system's forced rotational speed consistency error satisfies... This ensures that the entire error process, including transient overshoot, convergence rate, and steady-state accuracy, is strictly limited within the preset funnel boundary.

[0026] It should be noted that the above funnel error transformation has the following important mathematical properties: First, the transformation has the following properties in its domain: The internal property is strictly monotonically increasing and continuously differentiable, ensuring the smoothness of the controller design; secondly, when hour, ,when hour, This means that as long as it is guaranteed Boundedness ensures Thirdly, by introducing a square root fractional structure, this transformation avoids the non-differentiability singularity problem that may occur when the error crosses the zero point in the traditional funnel transformation, and achieves continuous differentiability in the entire domain, which provides convenience for the subsequent design of controllers based on the Lyapunov method.

[0027] In summary, by defining the speed consistency error and constructing a preset time funnel error function, this step establishes an error transformation framework with explicit transient performance constraints for the multi-PMSM cooperative drive system. This ensures that the speed synchronization error is always limited by the user-preset funnel boundary throughout the entire process, and also lays the foundation for the subsequent design of nonlinear filters and distributed consistency control laws.

[0028] Step 103: Design a virtual control law and introduce a preset time nonlinear filter. The filter takes the virtual control law as input and outputs a filtered signal. Based on the current state and the filtered signal in the composite fault mathematical model, define the current subsystem error and the filter error.

[0029] This step aims to design a virtual control law and introduce a pre-set time nonlinear filter to obtain a smooth filtered signal, replacing the direct analytical differentiation of the virtual control law in the backstepping method, thereby eliminating the problem of computational complexity inflation. Simultaneously, based on the current state and filtered signal in the composite fault mathematical model, the current subsystem error and filter error are defined, laying the foundation for the derivation of the final control law.

[0030] First, based on the preset time funnel error defined in step 102. Design No. Virtual control law of Taiwan PMSM This virtual control law is the first step in the backstepping design, used to generate the desired q-axis current command to drive the speed consistency error to converge to zero. The virtual control law is: (5) in, This represents the virtual control law, i.e., the desired q-axis current command. , , , These represent positive design parameters, Denotes the fractional power parameter, satisfying and The ratio of even numbers to odd numbers. This indicates the user-preset convergence time, which can be set independently according to process requirements. This represents the first gain coefficient of the preset time convergence term. It includes a hyperbolic tangent function to ensure the smoothness of the control signal. Indicates a positive design parameter. This represents the second gain coefficient of the preset time convergence term. Indicates the first The communication neighbor set of Taiwan PMSM, Indicates the first The q-axis current of the motor. express The derivative of Indicates adaptive parameters, This represents the first adaptive estimate of the neural network weight norm. This represents the square of the norm of the basis function vector. The non-linear part is generated using neural network fitting. , Represents nonlinear terms, This represents the input vector. It is the ideal weight vector of the neural network, processed using an adaptive law. These are radial basis functions (RBFs), such as the Gaussian function. To approximate the error and satisfy the strictly bounded condition , This is an unknown constant. The virtual control law, through a combination of a preset time convergence term and an adaptive compensation term, ensures that the speed consistency error is within a certain range. It converges to the zero neighborhood and simultaneously compensates for the nonlinear dynamics and uncertainties within the system online.

[0031] For the weight vector, use The system is then processed, and then adaptive processing is used. In one embodiment, the dynamic nonlinear terms within the multi-PMSM system are approximated online using a neural network. The adaptive law expression of the neural network is as follows: (6) in, , , These are positive design parameters.

[0032] In practical engineering, directly performing analytical differentiation on the aforementioned virtual control law introduces exponentially increasing high-order nonlinear terms and cross-coupling terms, leading to a severe computational complexity explosion problem. This makes the control algorithm difficult to implement in resource-constrained embedded digital signal processors (DSPs) or field-programmable gate arrays (FPGAs). To solve this technical challenge, this step introduces a preset time nonlinear filter. Using the virtual control law as input, the filter dynamically outputs a smoothed signal, thus avoiding direct analytical differentiation of the virtual control law. The preset time nonlinear filter is as follows: (7) in, This represents the output of the preset time nonlinear filter, i.e., the smoothed filtered signal. This represents the filter error, indicating the deviation between the filter output and the virtual control law input. , , These represent positive design parameters, used to adjust the filter's convergence speed and filtering accuracy. This filter ensures filter error accuracy by introducing a hyperbolic tangent function and fractional power terms. The generated filtered signal can converge to a bounded neighborhood within a preset time. Smooth and continuous, effectively avoiding high-frequency jitter.

[0033] After introducing a preset time nonlinear filter, based on the current state and filtered signal in the composite fault mathematical model, the current subsystem error and filter error are defined as follows: (8) (9) in, Represents the error of the current subsystem, characterizing the first Actual q-axis current of the PMSM With filter output The deviation between them. Through the above definition, the complex calculations that originally required analytical differentiation of the virtual control law in the backstepping method are transformed into processing the filtered signal and its derivative, thus effectively avoiding the problem of computational complexity inflation. Meanwhile, the filter error and the current subsystem error will be uniformly analyzed in subsequent steps using the Lyapunov method to ensure the stability of the entire closed-loop system.

[0034] This step, by designing a virtual control law and introducing a pre-defined time-nonlinear filter, replaces the direct analytical differentiation of the virtual control law with a smoothed filtered signal. This fundamentally eliminates the computational complexity explosion problem in the backstepping method design, significantly reducing the computational complexity of the control algorithm and enabling advanced nonlinear control algorithms to be implemented in real time in resource-constrained embedded controllers. Simultaneously, the defined current subsystem error and filter error provide the necessary error variable foundation for the subsequent derivation of the distributed consensus control law.

[0035] Step 104: Based on the mathematical model of the composite fault, the preset time funnel error function, the current subsystem error, and the filter error, derive the distributed preset time consistency control law to counteract the effects of the composite fault and unknown external disturbances.

[0036] This step aims to derive the final distributed preset time consistency control law based on the composite fault mathematical model established in step 101, the preset time funnel error function constructed in step 102, and the current subsystem error and filter error defined in step 103. It also incorporates an adaptive fault compensation mechanism to ensure that all signals in the closed-loop system are bounded and that the speed consistency error converges to the zero neighborhood within the user-preset time.

[0037] The distributed time consistency control law is as follows: (10) in, Indicates the first The actual control input of the PMSM (Power Management System) This represents the adaptive compensation parameters for adaptive compensation of additive faults in the actuator. This represents the second adaptive estimate of the neural network weight norm. This represents the square of the norm of the basis function vector of the neural network.

[0038] The nonlinear part is fitted using a neural network, and its expression is as follows: ;in, Represents nonlinear terms, This represents the input vector. It is the ideal weight vector of the neural network, processed using an adaptive law. These are RBF basis functions, such as the Gaussian function. To approximate the error and satisfy the strictly bounded condition , It is an unknown constant.

[0039] For the weight vector, use The process is then performed, followed by adaptive processing, expressed as follows: (11) in, , , These are positive design parameters.

[0040] The above control law consists of five functional parts: Part 1 Compensating actuator additive faults and unknown disturbances; Part Two Provide damping to suppress error oscillations; Part Three Constructing nonlinear damping terms to compensate for system uncertainties; Part Four The nonlinear dynamics of the current loop are approximated and compensated online using a neural network. Part Five As a preset time convergence term, the current error is guaranteed to be within the preset time by using a combination of fractional powers. It converges inward to the zero neighborhood. (Pre-multiplication factor) It not only compensated for multiplicative faults, but also standardized the dimensions of the control quantities.

[0041] Optionally, considering the impact of actuator additive faults and external unknown disturbances on the system, this step designs an adaptive fault compensation law, which estimates and adjusts the adaptive compensation parameters for actuator additive faults online. This enables proactive suppression of additive faults and unknown disturbances. The corresponding adaptive law expression is: (12) in, , , These represent the positive design parameters. In this adaptive law... The item is updated based on the current error driving parameters. and The terms guarantee the boundedness of parameter estimation and the pre-defined time convergence characteristics.

[0042] Finally, to prove the stability of the entire closed-loop system, the following global Lyapunov function is constructed: ; in, , , These represent the parameter estimation errors. Through analysis of... By taking the derivative and substituting the aforementioned virtual control law, preset time nonlinear filter, control law, and adaptive law into the equation, a rigorous mathematical derivation yields the following: ; in, , , , , , All are positive numbers. This represents the bounded residual term. Based on the preset time stability criterion, the above inequality guarantees that all signals in the closed-loop system are uniformly bounded, and that the speed consistency error converges to an arbitrarily small neighborhood near the origin within a user-preset time.

[0043] In summary, this step, by combining a composite fault model, a preset time funnel error function, and a preset time nonlinear filter, derives a distributed preset time consistency control law. Furthermore, it incorporates a neural network adaptive approximation and fault compensation adaptive mechanism, achieving effective fault-tolerant control under additive actuator faults. Rigorous Lyapunov stability analysis proves that all signals in the closed-loop system are bounded, and the speed consistency error converges to the zero neighborhood within a preset time, ensuring the high-precision speed synchronization control performance of the multi-PMSM cooperative drive system under complex composite fault conditions. The experimental results are as follows: Figure 2 As shown, where, Figure 2 (a) is a diagram showing the speed comparison of four PMSMs within a preset time period, where Ref represents the reference speed value; Figure 2 (b) is a schematic diagram comparing the consistency error of the four PMSMs within a preset time. Figure 2 (c) is a schematic diagram comparing the tracking errors of four PMSMs within a preset time. Figure 2Figure (d) shows a comparison of the root mean square error (RMSE) of the speed tracking of four PMSMs at different time periods. The figure illustrates the experimental results of the multi-permanent magnet synchronous motor speed consistency control method provided in this application, tracking a trapezoidal reference speed trajectory for 150 seconds. Figure 2 As shown in (a) and (c), all four PMSMs achieved smooth dynamic tracking throughout the acceleration, steady-state operation, and deceleration processes, with the response curves of each PMSM almost overlapping. Although unavoidable adverse conditions such as friction, sensor noise, and communication delays exist during system operation, the designed algorithm can effectively suppress these non-ideal disturbances, controlling the steady-state tracking error within ±5 r / min; at the same time, it can also effectively suppress dynamic overshoot at moments of abrupt changes in trajectory slope, such as 60s, 90s, and 120s.

[0044] Depend on Figure 2 As shown in (b), during the entire 150s operation phase, the consistency error was consistently limited within the preset funnel-shaped constraint boundary, eliminating the divergence problem in multi-motor cooperative operation caused by physical disturbances. Combined with... Figure 2 The quantitative analysis results of (d) show that in the three key dynamic transition intervals of 58-66s, 89-92s, and 119-123s, the root mean square error (RMSE) of all motor speed tracking is less than 2.0 r / min, with a peak value of only 1.99 r / min.

[0045] The above-mentioned physical experiment results show that the algorithm can not only guarantee the excellent speed tracking accuracy of a single motor, but also strictly constrain the global consistency error of multiple motors, and has excellent anti-disturbance robustness and engineering practical value.

[0046] The following describes the speed consistency control method for multiple permanent magnet synchronous motors provided by the present invention through three specific embodiments.

[0047] Example 1: Coordinated speed control of electric vehicles driven independently by four hub motors under extreme operating conditions.

[0048] In high-end distributed drive electric vehicles, the four-wheel independent hub motor or wheel-side motor layout eliminates the traditional differential and driveshaft, achieving ultimate flexibility in torque vector distribution. However, during vehicle operation, these issues can easily lead to problems such as split-road surfaces (e.g., dry asphalt on one side and compacted ice and snow on the other), signal drift of wheel speed sensors due to chassis vibration and electromagnetic interference, and output voltage distortion and torque efficiency degradation caused by aging of the inverter's IGBT modules. If these concurrent combined actuator failures are not quickly suppressed, they can cause the four wheels to lose synchronization, generating undesirable yaw torque, leading to vehicle skidding or even loss of control.

[0049] A mathematical model of composite faults in a multi-PMSM system is established. The four hub motors, their local inverters, and controllers are modeled as agents 1 to 4 in a multi-agent system. The vehicle controller, acting as a virtual leader, broadcasts the desired vehicle speed command. The four motors communicate via an onboard Ethernet / CAN-FD topology, and the adjacency matrix weights are determined by the inter-wheel mechanical coupling and communication delay characteristics.

[0050] Regarding the first A hub motor is used. A dynamic model is established based on formula (1), and a composite fault is added, simplifying it to formula (2). Specifically, it characterizes the decrease in actuator multiplicative efficiency caused by the aging of the inverter's IGBTs, such as This indicates that the actual output torque of the wheel is only 80% of the commanded value; Specifically, the model characterizes the additive faults caused by the dead zone effect and zero drift of power devices; it accurately describes the mechanism by which the dual coupling of information distortion and torque attenuation in the vehicle environment affects the consistency of four-wheel speed.

[0051] To achieve the stringent requirements of vehicle stability for speed synchronization, the local neighborhood synchronization error of the four wheels is defined according to formula (3), and the preset time funnel error conversion function is constructed according to formula (4).

[0052] Funnel boundary function Based on the vehicle's active safety requirements: In straight-line driving mode... Designed to be within a preset time Internal deviation from initial allowable deviation Decay to steady-state accuracy In steering mode, to suppress yaw, the preset time is tightened to... This transformation forces the actual wheel speed difference. This completely avoids drive wheel slippage and vehicle instability caused by excessive transient overshoot.

[0053] According to formula (5), the virtual control input is taken. However, the computing resources of the vehicle motor controller (such as the Infineon AURIX series DSP) are limited. Directly differentiating the virtual control law will occupy a large number of CPU cycles, resulting in control cycle jitter. Therefore, a preset time nonlinear filter is introduced, as shown in formula (7). This filter smooths the continuous dynamic surface signal. Replacing the direct analytical differentiation of virtual control laws reduces the floating-point operation burden of the underlying chip by approximately 60%, and its internal parameters... It ensures that the filtering error converges strictly within a preset time, ensuring the real-time and smoothness of torque command issuance, and eliminating the impact of high-frequency torque chattering on the motor bearings caused by traditional sliding mode control.

[0054] Final derivation The distributed preset time consistency control law of the wheel motor, as shown in formula (10), includes a hyperbolic tangent function robust term and an algebraic power preset time convergence term.

[0055] Combining the composite fault model with the adaptive compensation mechanism, an adaptive law for fault compensation is designed based on formula (12), and a related adaptive law for processing nonlinear terms using neural networks is designed based on formulas (6) and (12), which approximates the upper bound of the neural network weights of nonlinear terms online.

[0056] This means that even under extreme conditions where all four wheels experience varying degrees of inverter efficiency degradation, the signals of the entire closed-loop network remain bounded, and the speed consistency error will necessarily be within the user-set time. It converges to a small neighborhood. When the vehicle is driving on a split road, the adaptive compensation term of the faulty wheel automatically increases the torque command of that wheel to offset the performance degradation. The output of the other three wheels is adjusted in real time through topology association to ensure that the yaw rate of the whole vehicle returns to the steady state value within a preset 0.3 seconds, avoiding panic correction by the driver and significantly improving the active safety margin and fault-tolerant driving capability of distributed drive electric vehicles.

[0057] Example 2: High-speed train multi-power unit cooperative traction control system.

[0058] In an 8-car high-speed trainset, four power units (each with two traction motors) are typically configured. Each power unit forms an independent electric transmission system with the wheelsets via a pantograph, traction converter, and wheelset. When the train travels at high speed through curves or through sections with sudden weather changes (such as tunnel entrances or rain / snow), the wheel-rail adhesion coefficient between different wheelsets may vary drastically. Simultaneously, the IGBT / IGCT power modules in the traction converter are prone to gate aging under long-term thermal cycling, leading to output voltage distortion. Speed ​​sensors may also experience measurement deviations due to gear backlash and ice / snow adhesion. Improper handling of these concurrent faults can cause torque imbalances between power units, resulting in wheelset slippage / skidding, excessive train impulse, and even triggering traction closures leading to delays.

[0059] Each power unit's traction motor and its converter are modeled as multiple agents, with the vehicle traction control unit (TCU) acting as the virtual leader. The power units form a chain communication topology via train-level MVB or Ethernet. A multiplicative fault model considering the traction converter is also established. With additive faults The multi-dynamic unit cooperative dynamic equations. Among them, This reflects a decrease in torque output efficiency due to the aging of the power module. This model reflects the DC bias torque caused by dead-zone compensation failure. It clearly demonstrates the impact mechanism of the dual coupling of information loss and torque attenuation on the adhesion utilization and operational stability of high-speed trains.

[0060] Define the wheelset speed consistency error of each power unit Construct a preset time funnel conversion function In view of the characteristics of high-speed train operation, the funnel boundary Based on a comprehensive consideration of the route operation plan and comfort indicators, the following settings are implemented during the startup phase: [Preset time] The time limit is set to 5 seconds to ensure that all power units accelerate synchronously when the train starts in the station, avoiding excessive longitudinal impulse. During the constant speed operation phase, The time limit is set to 2 seconds to ensure rapid convergence of the wheel speed difference during adhesion recovery in rainy or snowy weather. This transformation forces the absolute value of the actual speed synchronization error to be minimized at any given moment. This completely avoids wheel spin and rail abrasion caused by excessive transient overshoot in traditional control systems.

[0061] Train TCUs typically employ industrial-grade DSPs or PowerPC architecture processors with limited floating-point computing power. To avoid lengthy analytical terms arising from direct differentiation of the virtual control law in the backstepping method, a preset time nonlinear filter is introduced to obtain a smooth signal that approximates the virtual control signal. The filter parameters are determined based on the TCU's control cycle (usually 1200°C). The algorithm is tuned to the storage capacity, so that it can be implemented on the existing hardware platform without the need to add an additional FPGA coprocessor, providing an efficient computing foundation for the engineering transformation of the train traction control system.

[0062] The parameterized Young's inequality is used to decouple the coupled disturbances between multiple dynamic units, and an adaptive law is designed. Online approximation of actuator failure or efficiency loss. Derivation of distributed preset time consistency control laws for each power unit. By constructing the total Lyapunov function and applying the robust lemma of the hyperbolic tangent function, the torque command that eliminates high-frequency chattering of the control input is finally obtained.

[0063] Rigorous theoretical proof demonstrates that the derivative of the Lyapunov function satisfies the preset time criterion structure. This means that even under conditions where motor parameters are uncertain due to temperature drift, sudden changes in wheel-rail adhesion, and failure of combined sensors and actuators, the entire closed-loop network signal remains bounded, and the speed consistency error will necessarily be within the user-set time. Converging to a small neighborhood. When a single power unit experiences a sudden drop in output torque due to a converter failure, the remaining healthy units automatically share the traction force through a distributed topology, completing torque redistribution within a preset time to ensure that the train does not lose its timetable while maintaining the longitudinal acceleration change rate between carriages in accordance with the EN 12299 comfort standard.

[0064] Example 3: Multi-pod collaborative control system for large all-electric propulsion ships.

[0065] In large cruise ships, LNG carriers, or military vessels with integrated electric propulsion, two to four podded electric thrusters (PODs) are commonly used to achieve full-radius thrust vector control. Each pod is directly driven by a high-power permanent magnet synchronous motor. When the ship maneuvers in narrow waterways, navigates in rough seas, or performs tactical evasive maneuvers, the thrusters must maintain strict speed synchronization to balance the ship's thrust and avoid shaft torsional vibration and hull resonance. However, high salt spray and strong electromagnetic interference in the marine environment can easily cause model disturbances; the propulsion inverters experience power device aging under long-term high-load operation, leading to multiplicative / additive actuator failures; and propeller cavitation and sudden changes in wind, waves, and current constitute strong unknown external disturbances. These concurrent compound failures can severely disrupt the coordination of multiple thrusters, leading to increased hull rolling, course deviation, and even fatigue fracture of the propulsion shaft.

[0066] Each podded propulsion unit's PMSM and its inverter are modeled as a multi-agent system, with the Vessel Navigation Management System (VMS) acting as the virtual leader. The propulsion units are connected via a fully connected or ring-shaped communication topology using marine fiber optic Ethernet. A system is established that simultaneously considers the multiplicative faults of the inverters. (Power module aging leading to output voltage amplitude attenuation) and additive faults The dynamic equations for multi-thruster coordinated drive (inverter zero drift and dead zone effect) are presented. Wind, wave, and current disturbances, as well as propeller cavitation torque fluctuations, are incorporated into the system's internal dynamic nonlinear terms. It accurately depicts the impact mechanism of the dual coupling of information distortion and thrust attenuation on ship course maintenance and propulsion shaft safety in marine environments.

[0067] Define the consistency error of the rotational speed of each pod's thruster. Construct a preset time funnel conversion function Regarding ship maneuvering characteristics, the funnel boundary Based on a comprehensive setting considering collision avoidance maneuver requirements and shaft torsional vibration limits: during low-speed maneuvers in and out of port, a preset time... The time limit is set to 10 seconds to ensure that all thrusters respond synchronously to rudder commands under low-speed, high-torque conditions, preventing lateral slippage of the hull due to uneven thrust; during high-speed cruising in open waters, The setting of 5 seconds ensures rapid convergence of speed differences when encountering sudden changes in wind, waves, and currents, preventing excessive alternating stress in the shafting system. Through this transformation, the system forces the actual speed synchronization error to always be limited within the safe boundaries allowed by the classification society's specifications, completely avoiding propeller bearing impact and hull structural fatigue caused by excessive transient overshoot.

[0068] Marine propulsion control cabinets typically use industrial frequency converters such as Siemens S120 or ABB ACS880. Their internal control loop computational resources must simultaneously handle motor control, harmonic suppression, and grid feedback functions. A preset time-nonlinear filter is introduced to smooth the filtered signal instead of directly calculating the derivative of the virtual control law, avoiding the computational complexity explosion of the backstepping method. The filter parameters are based on the frequency converter control cycle (typically...). The load rate of the floating-point arithmetic unit is tuned to enable advanced nonlinear control algorithms to be computed in real time on existing marine embedded platforms without the need for additional high-performance industrial control computers, thus reducing system modification costs and classification society certification risks.

[0069] Based on a composite fault model, an adaptive law is designed to approximate the upper bound of the inverter's efficiency loss online. The distributed preset time consistency control law for each pod thruster is derived. The robust lemma of hyperbolic tangent function is used to eliminate high-frequency chattering of control input, thus avoiding structural noise and cavitation erosion of propulsion motor due to torque pulsation.

[0070] The total Lyapunov function is constructed, and the algebraic power inequality is used to prove that its derivative satisfies the preset time stability criterion. This means that even under extremely harsh sea conditions where motor parameters are uncertain due to changes in seawater temperature, sudden changes in wind, waves, and currents, and where some thrusters suffer from the failure of composite sensors and actuators, the signals of the entire closed-loop network are still bounded, and the speed consistency error will necessarily be within the user-set time. It converges to a small neighborhood. When a single podded thruster experiences a sudden drop in thrust due to a frequency converter failure, the remaining thrusters automatically compensate for the thrust loss through distributed topology association, restoring speed synchronization within a preset time. This ensures that the ship can maintain more than 70% of its design speed even in a "battle-damaged" or "faulty" state where some thrusters have failed. It provides highly reliable fault-tolerant control for ship dynamic positioning (DP) and emergency avoidance, significantly improving the survivability and seaworthiness of all-electric propulsion ships.

[0071] This invention provides a method for controlling the speed consistency of multiple permanent magnet synchronous motors (PMSMs). It establishes a mathematical model of a composite fault in the multiple PMSM system; defines the speed consistency error of the composite fault mathematical model; and constructs a preset time funnel error function based on the speed consistency error and a preset time-varying funnel boundary function to impose full-process performance constraints on the errors of the multiple PMSM system. A virtual control law is designed, and a preset time nonlinear filter is introduced. The filter takes the virtual control law as input and outputs a filtered signal. Based on the current state and the filtered signal in the composite fault mathematical model, the current subsystem error and the filter error are defined. Based on the composite fault mathematical model, the preset time funnel error function, the current subsystem error, and the filter error, a distributed preset time consistency control law is derived to counteract the effects of composite faults and unknown external disturbances. This invention achieves the following technical effects by establishing a composite fault mathematical model, defining the speed consistency error and constructing a preset time funnel error function, introducing a preset time nonlinear filter, and deriving a distributed preset time consistency control law: First, it achieves effective fault-tolerant control under the concurrent action of multiplicative and additive faults of actuators in a multi-PMSM cooperative drive system, significantly improving the robustness of the system under complex compound fault conditions. Second, by applying a time-varying constraint to the rotational speed consistency error through a preset time funnel error function, the transient overshoot, convergence rate and steady-state accuracy of the error are always limited within the user-preset safety boundary, effectively avoiding mechanical shock and synchronization deviation amplification caused by excessive transient overshoot. Third, by using a preset time nonlinear filter to smooth the filtered signal instead of analytically differentiating the virtual control law, the "computational complexity explosion" problem in the backstepping method design is fundamentally eliminated, and the computational complexity of the control algorithm is greatly reduced, enabling it to be implemented in real time in resource-constrained embedded controllers. Fourth, the convergence time of this method can be preset independently by the user and is not affected by the coupling of the initial state of the system and the controller parameters. It can meet the rigid requirements of high-end equipment for "timed convergence" and realize the deep coupling of preset time performance and full-process performance constraints.

[0072] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0073] The following are device embodiments of the present invention. For details not described in detail, please refer to the corresponding method embodiments described above.

[0074] Figure 3The diagram shows a schematic of a speed consistency control device for multiple permanent magnet synchronous motors according to an embodiment of the present invention. For ease of explanation, only the parts related to the embodiment of the present invention are shown, and are described in detail below: like Figure 3 As shown, the speed consistency control device for multiple permanent magnet synchronous motors includes: a composite fault mathematical model construction module 31, a preset time funnel error function construction module 33, an error construction module 33, and a consistency control law construction module 34.

[0075] The composite fault mathematical model construction module 31 is used to establish a composite fault mathematical model for a multi-PMSM system. The preset time funnel error function construction module 33 is used to define the speed consistency error of the composite fault mathematical model, and construct the preset time funnel error function based on the speed consistency error and the preset time-varying funnel boundary function to apply full-process performance constraints to the multi-PMSM system error. Error construction module 33 is used to design virtual control law, introduce a preset time nonlinear filter, the filter takes the virtual control law as input and outputs a filtered signal, and defines the current subsystem error and filter error according to the current state and the filtered signal in the composite fault mathematical model. The consistency control law construction module 34 is used to derive a distributed preset time consistency control law based on the composite fault mathematical model, the preset time funnel error function, the current subsystem error and the filter error, so as to counteract the effects of composite faults and unknown external disturbances.

[0076] In one possible implementation, the mathematical model of composite faults is a concurrent model of additive and multiplicative faults in actuators, established using graph theory topology. The mathematical model of composite faults is as follows: ; in, Indicates the first The rotational speed of the PMSM, Indicates the first The rate of change of the rotational speed of the PMSM Indicates the first The q-axis current of the PMSM, Indicates the first The rate of change of the q-axis current of the PMSM. Indicates the speed loop gain. Indicates the current loop gain. This indicates the nonlinear term of the speed loop. This represents the nonlinear term of the current loop. Indicates the first The actuator of the PMSM actually has a faulty output. Indicates the first Actual measurement output of the PMSM unit. Indicates the first A multiplicative fault in the actuator of the PMSM unit. Indicates the first The ideal control signal for the controller design of the PMSM is... Indicates the first The actuator of the PMSM has an additive fault.

[0077] In one possible implementation, the rotational speed consistency error is: ; in, Indicates the first Rotational speed consistency error of the PMSM unit The weights of the adjacency matrix represent the communication topology. Indicates the first Actual measurement output of the PMSM unit. , This indicates the total number of PMSMs in a multi-PMSM system. This indicates the communication weight between followers and leaders. This indicates the reference rotation speed of the virtual leader.

[0078] In one possible implementation, the preset time funnel error function is: ; in, This indicates the preset time funnel error. It is represented as a time-varying funnel boundary function.

[0079] In one possible implementation, the virtual control law is: ; in, Represents a virtual control law. , , , These represent positive design parameters, Represents the fractional power parameter. This indicates the user-preset convergence time. This represents the first gain coefficient of the preset time convergence term. This represents the second gain coefficient of the preset time convergence term. Indicates the first The communication neighbor set of Taiwan PMSM, Indicates the first The q-axis current of the motor. express The derivative of Indicates adaptive parameters, This represents the first adaptive estimate of the neural network weight norm. This represents the square of the norm of the basis function vector; The preset time nonlinear filter is: ; in, This represents the output of the preset time nonlinear filter. Indicates filter error. , , These represent positive design parameters; The current subsystem error and the filter error are respectively: ; ; in, This indicates the error in the current subsystem.

[0080] In one possible implementation, the distributed pre-defined time consistency control law is: ; in, Indicates the first The actual control input of the PMSM (Power Management System) This represents the adaptive compensation parameters for adaptive compensation of additive faults in the actuator. This represents the second adaptive estimate of the neural network weight norm. This represents the square of the norm of the basis function vector of the neural network.

[0081] In one possible implementation, The corresponding adaptive law expression is: ; in, , , These represent the positive design parameters.

[0082] In one possible implementation, the dynamic nonlinear terms within the multiple PMSM system are approximated online using a neural network. The adaptive law expression of the neural network is as follows: ; ; in, , , , , , These represent the positive design parameters.

[0083] The above embodiments provide a speed consistency control device for multiple permanent magnet synchronous motors (PMSMs). A composite fault mathematical model construction module establishes a composite fault mathematical model for the multiple PMSM system. A preset time funnel error function construction module defines the speed consistency error of the composite fault mathematical model and, based on the speed consistency error and a preset time-varying funnel boundary function, constructs a preset time funnel error function to impose full-process performance constraints on the errors of the multiple PMSM system. An error construction module designs a virtual control law and introduces a preset time nonlinear filter. The filter takes the virtual control law as input and outputs a filtered signal. Based on the current state and the filtered signal in the composite fault mathematical model, the current subsystem error and the filter error are defined. A consistency control law construction module derives a distributed preset time consistency control law based on the composite fault mathematical model, the preset time funnel error function, the current subsystem error, and the filter error, thus offsetting the influence of composite faults and unknown external disturbances. This invention achieves the following technical effects by establishing a composite fault mathematical model, defining the speed consistency error and constructing a preset time funnel error function, introducing a preset time nonlinear filter, and deriving a distributed preset time consistency control law: First, it achieves effective fault-tolerant control under the concurrent action of multiplicative and additive faults of actuators in a multi-PMSM cooperative drive system, significantly improving the robustness of the system under complex compound fault conditions. Second, by applying a time-varying constraint to the rotational speed consistency error through a preset time funnel error function, the transient overshoot, convergence rate and steady-state accuracy of the error are always limited within the user-preset safety boundary, effectively avoiding mechanical shock and synchronization deviation amplification caused by excessive transient overshoot. Third, by using a preset time nonlinear filter to smooth the filtered signal instead of analytically differentiating the virtual control law, the "computational complexity explosion" problem in the backstepping method design is fundamentally eliminated, and the computational complexity of the control algorithm is greatly reduced, enabling it to be implemented in real time in resource-constrained embedded controllers. Fourth, the convergence time of this method can be preset independently by the user and is not affected by the coupling of the initial state of the system and the controller parameters. It can meet the rigid requirements of high-end equipment for "timed convergence" and realize the deep coupling of preset time performance and full-process performance constraints.

[0084] Figure 4 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. For example... Figure 4 As shown, the electronic device 4 in this embodiment includes a processor 40 and a memory 41. The memory 41 stores a computer program 42. When the processor 40 executes the computer program 42, it implements the steps in the various method embodiments described above. Alternatively, when the processor 40 executes the computer program 42, it implements the functions of each module / unit in the various device embodiments described above.

[0085] For example, computer program 42 may be divided into one or more modules / units, which are stored in memory 41 and executed by processor 40 to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of computer program 42 in electronic device 4.

[0086] Electronic device 4 may include, but is not limited to, processor 40 and memory 41. Those skilled in the art will understand that... Figure 4 This is merely an example of electronic device 4 and does not constitute a limitation on electronic device 4. It may include more or fewer components than shown, or combine certain components, or different components. For example, electronic device 4 may also include input / output devices, network access devices, buses, etc.

[0087] The processor 40 can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0088] The memory 41 can be an internal storage unit of the electronic device 4, such as a hard disk or RAM. The memory 41 can also be an external storage device of the electronic device 4, such as a plug-in hard disk, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card. Furthermore, the memory 41 can include both internal and external storage units of the electronic device 4. The memory 41 is used to store the computer program 42 and other programs and data required by the electronic device 4. The memory 41 can also be used to temporarily store data that has been output or will be output.

[0089] For the sake of simplicity and clarity, only the above-described functional modules / units are used as examples. In practical applications, the functions described above can be assigned to different functional modules / units as needed. These modules / units can be implemented in hardware, software, or a combination of both.

[0090] This invention also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the methods described in the above-described method embodiments.

[0091] This invention also provides a computer program product, including a computer program. When the computer program is executed by a processor, it implements the methods described in the above-described method embodiments.

[0092] Computer programs include computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. Computer-readable media can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0093] In the above embodiments, the descriptions of each embodiment have their own emphasis. Parts not detailed or described in a particular embodiment can be referred to in the relevant descriptions of other embodiments. Unless otherwise specified or in conflict with logic, the terminology and / or descriptions between different embodiments are consistent and can be referenced interchangeably. Technical features in different embodiments can be combined to form new embodiments based on their inherent logical relationships.

[0094] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for controlling the speed consistency of multiple permanent magnet synchronous motors, characterized in that, include: Establish a mathematical model for composite faults in a multi-PMSM system; Define the rotational speed consistency error of the composite fault mathematical model, and construct a preset time funnel error function based on the rotational speed consistency error and the preset time-varying funnel boundary function to apply full-process performance constraints to the multi-PMSM system error. A virtual control law is designed, and a preset time nonlinear filter is introduced. The filter takes the virtual control law as input and outputs a filtered signal. Based on the current state in the composite fault mathematical model and the filtered signal, the current subsystem error and the filter error are defined. Based on the mathematical model of the composite fault, the preset time funnel error function, the current subsystem error, and the filter error, a distributed preset time consistency control law is derived to counteract the effects of composite faults and unknown external disturbances.

2. The method for controlling the speed consistency of multiple permanent magnet synchronous motors according to claim 1, characterized in that, The composite fault mathematical model is a concurrent model of actuator additive faults and multiplicative faults established using graph theory topology. The composite fault mathematical model is as follows: ； in, Indicates the first The rotational speed of the PMSM, Indicates the first The rate of change of the rotational speed of the PMSM Indicates the first The q-axis current of the PMSM, Indicates the first The rate of change of the q-axis current of the PMSM. Indicates the speed loop gain. Indicates the current loop gain. This indicates the nonlinear term of the speed loop. This represents the nonlinear term of the current loop. Indicates the first The actuator of the PMSM actually has a faulty output. Indicates the first Actual measurement output of the PMSM unit. Indicates the first A multiplicative fault in the actuator of the PMSM. Indicates the first The ideal control signal for the controller design of the PMSM is... Indicates the first The actuator of the PMSM has an additive fault.

3. The method for controlling the speed consistency of multiple permanent magnet synchronous motors according to claim 2, characterized in that, The rotational speed consistency error is: ； in, Indicates the first Rotational speed consistency error of the PMSM unit The weights of the adjacency matrix represent the communication topology. Indicates the first Actual measurement output of the PMSM unit. , This indicates the total number of PMSMs in a multi-PMSM system. This indicates the communication weight between followers and leaders. This indicates the reference rotation speed of the virtual leader.

4. The method for controlling the speed consistency of multiple permanent magnet synchronous motors according to claim 3, characterized in that, The preset time funnel error function is: ； in, This indicates the preset time funnel error. It is represented as a time-varying funnel boundary function.

5. The method for controlling the speed consistency of multiple permanent magnet synchronous motors according to claim 4, characterized in that, The virtual control law is: ； in, This represents the virtual control law. , , , These represent positive design parameters, Represents the fractional power parameter. This indicates the user-preset convergence time. This represents the first gain coefficient of the preset time convergence term. This represents the second gain coefficient of the preset time convergence term. Indicates the first The communication neighbor set of Taiwan PMSM, Indicates the first The q-axis current of the motor. express The derivative of Indicates adaptive parameters, This represents the first adaptive estimate of the neural network weight norm. This represents the square of the norm of the basis function vector; The preset time nonlinear filter is: ； in, This represents the output of the preset time nonlinear filter. Indicates filter error. , , These represent positive design parameters; The current subsystem error and the filter error are respectively: ； ； in, This indicates the error in the current subsystem.

6. The method for controlling the speed consistency of multiple permanent magnet synchronous motors according to claim 5, characterized in that, The distributed preset time consistency control law is as follows: ； in, Indicates the first The actual control input of the PMSM (Power Management System) This represents the adaptive compensation parameters for adaptive compensation of additive faults in the actuator. This represents the second adaptive estimate of the neural network weight norm. This represents the square of the norm of the basis function vector of the neural network.

7. The method for controlling the speed consistency of multiple permanent magnet synchronous motors according to claim 6, characterized in that, The The corresponding adaptive law expression is: ； in, , , These represent the positive design parameters.

8. The method for controlling the speed consistency of multiple permanent magnet synchronous motors according to claim 7, characterized in that, The internal dynamic nonlinear terms of the multi-PMSM system are approximated online using a neural network. The adaptive law expression of the neural network is as follows: ； ； in, , , , , , These represent the positive design parameters.

9. A speed consistency control device for multiple permanent magnet synchronous motors, characterized in that, include: The composite fault mathematical model construction module is used to build composite fault mathematical models for multi-PMSM systems; The preset time funnel error function construction module is used to define the rotational speed consistency error of the composite fault mathematical model, and construct the preset time funnel error function based on the rotational speed consistency error and the preset time-varying funnel boundary function, so as to apply full-process performance constraints to the multi-PMSM system error. An error construction module is used to design a virtual control law and introduce a preset time nonlinear filter. The filter takes the virtual control law as input and outputs a filtered signal. Based on the current state in the composite fault mathematical model and the filtered signal, the current subsystem error and the filter error are defined. The consistency control law construction module is used to derive a distributed preset time consistency control law based on the composite fault mathematical model, the preset time funnel error function, the current subsystem error, and the filter error, so as to counteract the effects of composite faults and unknown external disturbances.

10. An electronic device, characterized in that, It includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method as described in any one of claims 1 to 8.

Images