A motor operating state simulation test system

Abstract

The present application relates to motor drive control and hardware ring simulation test technical field, specifically to a kind of motor operating state simulation test system.It includes: signal acquisition module, for collecting the control instruction issued by measured controller, and obtains the environmental parameter and load parameter of test setting;Manifold space construction module, for reflecting the high-dimensional riemann manifold space of motor energy state, and change into the metric tensor of manifold space;Dynamic curvature evolution unit, for mapping external load disturbance and multi-physical field coupling effect into the dynamic change of metric tensor, geodesic trajectory represents the evolution path of motor operating state, to solve the operating state of motor next time;Physical quantity reconstruction module, for projecting the geodesic trajectory coordinates on manifold space back to Euclidean space, restore as electric signal physical quantity, and feedback to measured controller, form closed loop test.The present application significantly improves the physical authenticity and confidence of motor running simulation under high temperature or magnetic saturation state.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of motor drive control and hardware loop simulation testing, in particular to a motor operating state simulation testing system. BACKGROUND

[0002] With the increasing performance requirements of motor drive systems, the influence of multi-physical field coupling on the operating state of the motor is increasingly significant. The complex nonlinear characteristics make it extremely difficult to accurately evaluate the performance of the controller.

[0003] Currently, the traditional simulation methods of lumped parameter model or look-up table are generally used for hardware-in-the-loop testing. The existing technology usually simulates the motor behavior by simplifying the circuit equation or discrete data interpolation, and relies on microprocessor serial calculation to solve the physical state. However, the traditional lumped parameter model has the problem of precision loss under the condition of multi-physical field coupling, and cannot truly reflect the dynamic physical processes such as cross saturation and thermal demagnetization. At the same time, the traditional look-up table method is prone to numerical jump and derivative discontinuity, and is limited by computing resources. The existing testing system is difficult to maintain high model complexity while ensuring microsecond-level real-time, often leading to over-idealized simulation signals, which masks the design defects of the controller under extreme conditions.

[0004] Therefore, how to realize high-precision and high-fidelity closed-loop simulation of multi-physical field coupling effects under the premise of ensuring the real-time operation of the system, and break the restrictive relationship between calculation accuracy and speed, has become a problem to be solved in the field.

[0005] The above information disclosed in the above background section is only used to strengthen the understanding of the background of the present disclosure, and therefore it can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0006] To solve the above technical problems, the present application discloses a motor operating state simulation testing system. Specifically, the technical scheme of the present application is as follows:

[0007] The signal acquisition module is used to acquire the control instructions issued by the tested controller, and to obtain the environmental parameters and load parameters set for testing.

[0008] The manifold space construction module is used to establish a high-dimensional Riemann manifold space reflecting the energy state of the motor based on the physical parameters of the motor, and to convert the environmental parameters and load parameters into metric tensors of the manifold space. The metric tensors are used to define the geometric structure and curvature of the manifold space.

[0009] The dynamic curvature evolution unit is used to map external load disturbances and multi-physics coupling effects into dynamic changes of the metric tensor. It uses a symplectic geometric integrator to calculate the geodesic trajectory of the state point on the dynamically changing manifold space. The geodesic trajectory represents the evolution path of the motor's operating state, thereby solving the motor's operating state at the next moment.

[0010] The physical quantity reconstruction module is used to project the geodesic trajectory coordinates on the manifold space back to Euclidean space, restore them to electrical signal physical quantities, and feed the electrical signal physical quantities back to the controller under test through the hardware interface to form a closed-loop test.

[0011] Preferably, the manifold space construction module includes:

[0012] The offline training unit is used to perform multi-physics coupling training in advance through the finite element model, establish the mapping relationship between physical field parameters and manifold curvature, and generate a physical field-curvature lookup table.

[0013] The state-geometry mapping unit is used to retrieve the physics field-curvature lookup table based on the current environmental and load parameters during real-time simulation and construct the metric tensor of the real-time manifold space. When the ambient temperature increases or the magnetic saturation increases, the values ​​of the metric components and curvature of specific dimensions of the manifold space are adjusted accordingly to simulate nonlinear physical characteristics.

[0014] Preferably, the dynamic curvature evolution unit includes:

[0015] The tensor flow computation subunit is used to receive the pulse width modulation signal from the controller under test, and modify the distribution of the metric tensor in real time according to the duty cycle and frequency of the pulse width modulation signal, forming a dynamically changing Riemannian metric field on the manifold space.

[0016] The symplectic integral operation subunit is used to perform exponential mapping operations along the tangent vector on a curved geometric surface to calculate the manifold coordinates of the state point at the next time step. The symplectic integral operation subunit maintains the energy conservation property of the system during the calculation process to suppress numerical divergence.

[0017] Preferably, the dynamic curvature evolution unit further includes:

[0018] The fault simulation subunit is used to apply geometric distortion on a preset path in the manifold space. By modifying the metric tensor of the local region, a high curvature geometric distortion region is created, which forces the motion trajectory of the state point to deviate from the normal geodesic, so as to simulate the short circuit or open circuit fault between motor turns.

[0019] Preferably, the system further includes:

[0020] The stability assessment module is used to generate a manifold curvature distribution map and monitor the rate of change of geometric curvature in the region where the state point is located in real time.

[0021] The stability evaluation module is preset with a curvature change threshold, configured as follows: when the rate of change of geometric curvature of the region entered by the state point is less than or equal to the curvature change threshold, the system is determined to be in the linear stable region; when the rate of change of geometric curvature of the region entered by the state point is higher than the curvature change threshold, the system is determined to be in a critical unstable state, and the trajectory smoothness index is output as the evaluation basis for the robustness of the controller.

[0022] Preferably, the physical quantity reconstruction module includes:

[0023] The coordinate projection unit is used to map the coordinates of state points in a high-dimensional manifold space into three-phase current values ​​and rotor position angles.

[0024] The digital-to-analog converter interface unit is used to convert the three-phase current value and rotor position angle into an analog voltage signal and output it to the controller under test at a microsecond update rate. The output frequency of the digital-to-analog converter interface unit is configured to simulate high-frequency harmonic components.

[0025] Preferably, the system operates on a heterogeneous computing platform based on a field-programmable gate array (FPGA);

[0026] The manifold space construction module utilizes digital signal processing slices of field-programmable gate arrays to perform parallel interpolation operations on metric tensors, thereby enabling real-time simulation of multi-physics coupling.

[0027] Preferably, the following steps are also included:

[0028] Collect control commands issued by the controller under test, and obtain the environmental and load parameters set for the test.

[0029] A high-dimensional Riemannian manifold space reflecting the energy state of the motor is established based on the motor's physical parameters, and environmental parameters and load parameters are converted into metric tensors of the manifold space. The metric tensors are used to define the geometry and curvature of the manifold space.

[0030] External load disturbances and multiphysics coupling effects are mapped to the dynamic changes of the metric tensor. The geodesic trajectory of the state point on the dynamically changing manifold space is calculated using a symplectic geometric integrator. The geodesic trajectory represents the evolution path of the motor's operating state, thereby solving the motor's operating state at the next moment.

[0031] The coordinates of the geodesic trajectory on the manifold space are projected back into Euclidean space, restored to electrical signal physical quantities, and fed back to the controller under test through a hardware interface.

[0032] Preferably, the calculation of the geodesic trajectory of the state point in the dynamically changing manifold space using the symplectic geometric integrator includes:

[0033] The metric tensor field of the manifold space is updated in real time based on the current control commands and environmental parameters.

[0034] In the updated metric tensor field, starting from the current state point, an exponential mapping is performed along the tangent vector direction to solve for the coordinates of the state point at the next moment. The solution process follows the energy conservation constraint of symplectic geometry.

[0035] Preferably, the system further includes:

[0036] Real-time calculation of the geometric curvature of the trajectory of state points in manifold space;

[0037] Set a safe curvature threshold. When the calculated geometric curvature is less than or equal to the safe curvature threshold, it is recorded as normal operating data. When the calculated geometric curvature is greater than the safe curvature threshold, a potential fault warning signal is generated, and the trajectory smoothness data at this time is recorded.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] 1. This invention abandons the traditional linear parameter model and constructs a high-dimensional Riemannian manifold space that reflects the energy state of the motor, transforming complex nonlinear physical characteristics such as ambient temperature and magnetic saturation into metric tensors of the manifold space. By utilizing a physics-curvature lookup table generated through offline training, the system can adjust the curvature of the spatial geometry in real time, thereby accurately mapping the multiphysics coupling effect into geometric topological changes. This method effectively solves the problems of parameter distortion and low coupling accuracy in traditional simulations under extreme conditions, and significantly improves the physical realism and confidence of motor operation simulations under high temperature or magnetic saturation conditions.

[0040] 2. This invention employs a symplectic geometric integrator to calculate the geodesic trajectory of the state point on the dynamic manifold space, thereby solving for the motor's operating state. Unlike conventional numerical integration methods, the symplectic integral operation strictly adheres to the energy conservation constraints of the Hamiltonian system during the calculation process, effectively suppressing the accumulation and divergence of numerical errors. This means that during long-term or high-frequency dynamic response testing, the system can not only maintain extremely high calculation accuracy but also ensure that the energy characteristics of the simulation model do not drift over time, thus providing the controller under test with a stable, continuous feedback signal that conforms to the physical conservation laws, enhancing the reliability of closed-loop testing.

[0041] 3. By applying geometric distortion on a preset path in the manifold space, i.e., by modifying the metric tensor of a local region to create a high curvature region, a novel simulation of inter-turn short-circuit or open-circuit faults in a motor is achieved. This method is no longer limited to simple circuit parameter mutations, but forces the trajectory of the state point to deviate from the normal geodesic from the energy topology level, thereby enabling a deeper reproduction of the complex electromagnetic dynamic evolution process inside the motor when a fault occurs. This allows the test system to simulate early, minor faults with more concealed and dynamic characteristics, greatly enriching the test dimensions for controller fault diagnosis and fault tolerance.

[0042] 4. The system introduces manifold curvature distribution map and geometric curvature change rate as criteria for system stability, breaking through the limitations of traditional time-domain error analysis. By setting a curvature change threshold, the system can monitor in real time whether the state point enters the critical unstable state, and use trajectory smoothness as an evaluation index. This evaluation method based on differential geometry can intuitively quantify the robustness boundary of the motor controller in dealing with nonlinear disturbances, and can keenly capture potential instability trends that are difficult to detect by traditional methods. It provides more rigorous and scientific data support for optimizing the parameter tuning and safety boundary setting of the control algorithm. Attached Figure Description

[0043] The present invention will be further explained below with reference to the accompanying drawings and embodiments:

[0044] Figure 1 This is a system structure diagram of the present invention. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0046] Example 1:

[0047] Please see Figure 1 A motor operating state simulation test system, comprising:

[0048] The signal acquisition module is used to acquire control commands issued by the controller under test and to obtain the environmental and load parameters set for the test.

[0049] The manifold space construction module is used to build a high-dimensional Riemannian manifold space that reflects the energy state of the motor based on the motor's physical parameters, and to convert environmental parameters and load parameters into metric tensors of the manifold space. The metric tensors are used to define the geometry and curvature of the manifold space.

[0050] The dynamic curvature evolution unit is used to map external load disturbances and multi-physics coupling effects into dynamic changes of the metric tensor. It uses a symplectic geometric integrator to calculate the geodesic trajectory of the state point on the dynamically changing manifold space. The geodesic trajectory represents the evolution path of the motor's operating state, thereby solving the motor's operating state at the next moment.

[0051] The physical quantity reconstruction module is used to project the geodesic trajectory coordinates on the manifold space back to Euclidean space, restore them to electrical signal physical quantities, and feed the electrical signal physical quantities back to the controller under test through the hardware interface to form a closed-loop test.

[0052] This embodiment details the execution logic of a motor operation state simulation test system based on manifold dynamics. This system aims to solve the accuracy loss problem of traditional lumped parameter models under multi-physics coupling. The signal acquisition module, as the sensing front end of the system, connects the controller under test to the host computer and acquires six PWM control commands and environmental load settings at a set frequency.

[0053] Building upon this, the manifold space construction module does not directly solve the circuit equations. Instead, it constructs a high-dimensional Riemannian space based on the motor's physical parameters, converting the scalar values ​​of environmental and load parameters into metric tensors within this space. The source is the manifold space construction module, which is calculated based on environmental and load parameters. Its physical meaning is the core operator that defines the geometry and curvature of the manifold. The unit is a dimensionless tensor field.

[0054] The dynamic curvature evolution unit serves as a simulation engine, mapping external load disturbances and multi-physics coupling effects such as magnetic saturation and thermal demagnetization into a dynamic change field of the metric tensor. It also uses a symplectic geometric integrator to calculate the geodesic trajectory of the state point on the dynamic manifold. The geodesic trajectory originates from the geometric solution of the principle of least action, and its physical meaning is the natural path of the system's energy evolution. Its unit is a generalized coordinate sequence.

[0055] The physical quantity reconstruction module projects the abstract geodesic trajectory coordinates on the manifold space back into Euclidean space, restoring them into physical quantities such as three-phase current and rotor position, and feeds them back to the controller under test through a hardware interface. This embodiment transforms the complex nonlinear electromagnetic-thermal coupling behavior into a geometric evolution process on a high-dimensional Riemannian manifold, and naturally handles the cross saturation and parameter drift problems that are difficult to describe by traditional methods by utilizing the intrinsic properties of manifold geometry. While ensuring microsecond-level real-time performance, it achieves high-fidelity closed-loop simulation of multi-physics coupling effects, effectively breaking the constraints between computational accuracy and speed in HIL testing.

[0056] Example 2:

[0057] The manifold space construction module includes: an offline training unit, used to perform multi-physics coupling training in advance through a finite element model, establish the mapping relationship between physical field parameters and manifold curvature, and generate a physical field-curvature lookup table; and a state-geometry mapping unit, used to retrieve the physical field-curvature lookup table and construct the metric tensor of the real-time manifold space based on the current environmental parameters and load parameters during real-time simulation. When the ambient temperature increases or the magnetic saturation level increases, the values ​​of the metric components and curvature of specific dimensions of the manifold space are adjusted accordingly to simulate nonlinear physical characteristics.

[0058] This embodiment further refines the hybrid architecture implementation path of the manifold space construction module; the offline training unit is started before the system runs, and uses high-precision finite element analysis software to perform a full-condition scan of the motor under test, inputting the current vector, rotor position and temperature variables, and training to generate a physical field-curvature lookup table, quantifying the nonlinear characteristics of the physical field into the cross-sectional curvature of the Riemannian manifold; wherein, the cross-sectional curvature is generated by the finite element model training, and its physical meaning is a geometric quantity characterizing the resistance of a specific region to the motion of the state point, with the unit being the reciprocal square of the length;

[0059] During real-time simulation, the state-geometry mapping unit uses the real-time collected environmental and load parameters as index keys to quickly retrieve the above lookup table; in response to the increase in ambient temperature or magnetic saturation, the unit dynamically adjusts the metric component values ​​and curvature of the corresponding dissipation dimension in the manifold space.

[0060] Specifically, the current tensor Build a A positive definite matrix; where the diagonal elements Corresponding to the self-inductance or mechanical inertia of each phase structure, off-diagonal elements This corresponds to the cross-connection effect between multiple physics fields, such as phase-to-phase mutual inductance or magnetocaloric connection terms;

[0061] The state-geometry mapping unit performs the following tensor update formula to simulate nonlinear physical characteristics:

[0062] For diagonal elements :

[0063] ;

[0064] in, Real-time ambient temperature; For reference temperature; This is the default temperature influence coefficient, in units of: This is used to characterize the correction rate of temperature change to the corresponding geometric path length; for off-diagonal elements... Its value is not simply zero, but is based on the fundamental electric angle. Dynamic modulation is performed to simulate the dq-axis cross-penetration effect:

[0065] ;

[0066] in, The mutual inductance coupling coefficient, For rotor electrical angle, The initial phase offset angle is given; this full tensor modeling method ensures that the manifold space has a non-trivial geometric structure, rather than a simple weighted Euclidean space.

[0067] in, This is the reference value for the linear region at room temperature. For real-time ambient temperature, The preset reference ambient temperature, for example, 25 degrees Celsius. Based on current amplitude The magnetic saturation coefficient is obtained by retrieving the physical field-curvature lookup table. This is a preset temperature-weighted correction factor. This is a preset magnetic saturation weighting factor; for example, when the temperature... When an increase in resistance leads to an increase in the metric component of the corresponding dimension. The increase in numerical value geometrically lengthens the path of charge flow, increases the length of geodesics, and physically manifests as increased dissipation.

[0068] This mechanism, which combines offline training with online mapping, cleverly utilizes the high precision of the finite element model and the high speed of the lookup table method. At the same time, due to the natural geometric continuity of the manifold curvature, this scheme can ensure a smooth transition of simulation results when dealing with the lookup table interpolation interval, avoiding the numerical jumps and derivative discontinuities common in traditional parameter lookup table methods, and significantly improving the numerical stability of the system under transient conditions.

[0069] Example 3:

[0070] The dynamic curvature evolution unit includes: a tensor flow calculation subunit, which receives the pulse width modulation signal from the controller under test and modulates the distribution of the metric tensor in real time according to the duty cycle and frequency of the pulse width modulation signal to form a dynamically changing Riemannian metric field on the manifold space; and a symplectic integral operation subunit, which performs exponential mapping operations along the tangent vector on the curved geometric surface to calculate the manifold coordinates of the state point at the next time step. The symplectic integral operation subunit maintains the energy conservation characteristics of the system during the calculation process to suppress numerical divergence.

[0071] This embodiment details the core computational flow of the dynamic curvature evolution unit; the tensor flow calculation subunit receives and parses the duty cycle of the PWM signal. The switching frequency is converted into a source term that alters the manifold geometry; specifically, the PWM signal is modeled as a time-varying perturbation term of the metric tensor, modifying the distribution of the metric tensor in real time.

[0072] ;

[0073] in, For only generalized coordinates The static background metric tensor, This is the tensor for measuring the perturbation amplitude caused by pulse modulation. The carrier frequency of the PWM signal is Hz; thus, a dynamically changing Riemannian metric field is formed on the manifold space; the symplectic integral operation subunit abandons the traditional Euler method and defines the Hamiltonian function of the system:

[0074] ;

[0075] in, For generalized coordinate support, they correspond to the accumulated charge and bottom mechanical angle of each phase of the stator, respectively; For generalized momentum effects, these correspond to stator flux linkage and rotor angular momentum, respectively; The physical meaning is the generalized mass moment that includes the inductance tensor and the moment of inertia; The potential energy function is given; this sub-unit performs an exponential mapping operation based on the symplectic structure along the tangent vector direction; where, The source is the definition of the system energy function, the physical meaning is the Hamiltonian, and the unit is joule; The source is the metric tensor and the physical conservation law; the physical meaning is the Lie derivative operator of the Hamiltonian vector field; the unit is per second.

[0076] To maintain energy conservation while accurately simulating the modified consumption dissipation effect, the symplectic integral unit employs a generalized symplectic scheme based on operator splitting for state updates:

[0077] ;

[0078] in, The dissipative operator corresponds to the phase space volume shrinkage caused by resistive heat dissipation; its physical meaning is the dissipative step. ,in The time interval for the sub-steps of the segmentation algorithm is specifically defined by the momentum decay update formula:

[0079] ;

[0080] The stator resistance matrix is ​​used; by separating the dissipative terms from the Hamiltonian conservation terms, a symplectic algorithm is used that preserves the geometric stability of the non-dissipative part and accurately restores the energy decay characteristics of the physical system.

[0081] in, For the integration time step, Superscript here This represents the matrix transpose; it is used to solve problems related to tensors. This embodiment addresses the numerical simulation problem of indivisible Hamiltonian systems dependent on position coordinates, employing a symplectic algorithm based on implicit kinetic energy partitioning. Specifically, the Hamiltonian function... Split into kinetic energy subsystems With potential energy subsystem ;

[0082] Corresponding operator The potential energy subsystem is set explicitly, and its update formula is:

[0083] ;

[0084] Corresponding operator The kinetic energy subsystem involves implicit midpoint simulation. This embodiment uses a fixed-point iterative method for approximate simulation within the FPGA. To ensure the time determinism of the hard real-time system, the iteration process is strictly configured to a fixed number of iterations, for example, a fixed 3 iterations, rather than dynamic termination based on an error threshold; the implicit midpoint format is as follows:

[0085] ;

[0086] Operator and These correspond to the Lie derivative operators for the kinetic and potential energy terms in the Hamiltonian system, respectively; specifically, for the splitting of the Hamiltonian function... Operator Indicated by For the Hamiltonian subsystem in time The evolutionary unmapping within the subunit; through the above exponential mapping operation, the subunit precisely advances the state point to the manifold coordinates of the next moment on the curved geometric surface;

[0087] By employing a symplectic geometric integrator instead of traditional numerical solutions, the system strictly adheres to the symplectic structural characteristics of the physical system at the mathematical level. This forces the conservation of energy or the correct dissipation rate to be maintained during the calculation process. This geometric conformity fundamentally suppresses numerical divergence, especially under extreme conditions such as high-speed field weakening or high-frequency injection of the simulated motor, which can maintain the physical realism of the simulated trajectory for a long time without causing collapse.

[0088] Example 4:

[0089] The dynamic curvature evolution unit also includes a fault simulation subunit, which is used to apply geometric distortion on a preset path in the manifold space. By modifying the metric tensor of the local region, a high curvature geometric distortion region is created, which forces the motion trajectory of the state point to deviate from the normal geodesic, so as to simulate the short circuit or open circuit fault between motor turns.

[0090] This embodiment further elaborates on the fault simulation mechanism based on geometric distortion; the fault simulation subunit determines a specific region in the manifold space representing the stator winding or power device as the fault location; in response to a fault command triggered by the operator, the subunit immediately modifies the metric tensor of that local region; specifically, the fault simulation subunit modifies the reference tensor by adding a Gaussian function. Generate fault state tensors :

[0091] ;

[0092] in, For faults that occur by default in the coordinate system, such as those occurring in the region corresponding to a specific stator tooth, the coordinates... Use the defined fault-affected geometry width range. It is the identity matrix; For fault intensity coefficients, such as for inter-turn short-circuit faults, Positive values ​​are used to increase the viscosity of the local space; nearby values ​​increase, such as when geodesics converge or bend locally; this indicates an open circuit fault. Take negative values ​​and ensure Positive definiteness is used to maintain system stability;

[0093] When the state point moves to the vicinity of this region, its trajectory is forced to deflect sharply due to the geometric constraints of the geodesic equation, deviating from the normal evolution path. This geometric deviation accurately reproduces the current surge caused by a short circuit or the torque pulsation caused by an open circuit at the physical level.

[0094] This scheme abstracts various complex circuit faults into geometric pits or obstacles on a manifold space, eliminating the need to re-derive differential equations for each fault. This improves the versatility and development efficiency of the fault model and can realistically simulate the nonlinear transient process and chaotic characteristics that accompany the fault at the moment of occurrence.

[0095] Example 5:

[0096] The system also includes a stability assessment module, used to generate a manifold curvature distribution map and monitor the rate of change of geometric curvature in the region where the state point is located in real time. The stability assessment module is preset with a curvature change threshold, configured as follows: when the rate of change of geometric curvature in the region entered by the state point is less than or equal to the curvature change threshold, the system is determined to be in the linear stable region; when the rate of change of geometric curvature in the region entered by the state point is higher than the curvature change threshold, the system is determined to be in a critical unstable state, and the trajectory smoothness index is output as an evaluation basis for the robustness of the controller.

[0097] This embodiment introduces stability evaluation logic based on geometric properties; the stability evaluation module generates a manifold curvature distribution map in real time and calculates the rate of change of the geometric curvature of the current state point; wherein, The source is real-time differential calculation of manifold geometry, and its physical meaning is the rate of change of geometric curvature, with the unit being per second; The source is offline test data statistics preset, and the physical meaning is the critical threshold between the linear region and the unstable region, with the unit being per second;

[0098] The system executes logical judgments: if the rate of change of geometric curvature in the state point region is less than or equal to the curvature change threshold, the system determines that it is currently in the linear stable region; conversely, if the rate of change is higher than the threshold, the system determines that it has entered a critical unstable state, which means that the motor is about to enter a deep saturation or out-of-step oscillation region; in this state, the system calculates and outputs the trajectory smoothness index:

[0099] ;

[0100] in, The source is a sliding window integral based on the rate of change of curvature, and its physical meaning is a trajectory smoothness index, used to quantify the degree of geometric oscillation of the system in the critical region. This is a preset integration time window, in dimensionless units;

[0101] This module provides a predictive evaluation dimension that goes beyond traditional physical quantities. It can reveal the degradation of system stability in advance through abnormal fluctuations in manifold geometric curvature before the motor actually experiences overcurrent or shutdown faults. This provides deep geometric insights for controller robustness optimization and enhances the system's safety design capabilities.

[0102] Example 6:

[0103] The physical quantity reconstruction module includes: a coordinate projection unit, which maps the coordinates of state points in the high-dimensional manifold space to three-phase current values ​​and rotor position angles; and a digital-to-analog conversion interface unit, which converts the three-phase current values ​​and rotor position angles into analog voltage signals and outputs them to the controller under test at a microsecond update rate. The output frequency of the digital-to-analog conversion interface unit is configured to simulate high-frequency harmonic components.

[0104] This embodiment details the signal restoration mechanism of the physical quantity reconstruction module; the coordinate projection unit utilizes the canonical transformation relationship based on Hamiltonian mechanics to map the generalized coordinates containing multiple electromagnetic and thermodynamic states in the high-dimensional manifold space back to Euclidean space; specifically, the current vector The solution does not rely on black-box projection, but is calculated directly using the inverse matrix of the current tensor:

[0105] ;

[0106] The platform's position angle is directly supported by a wide range of coordinates. Number of mechanical angles Extraction; the digital-to-analog converter interface unit receives these digital physical quantities and converts them into analog voltage signals through a high-speed DAC; in this process, the output frequency of the unit is configured to be more than a hundred times the motor switching frequency, for example, reaching the megahertz level, to ensure that it can carry and output high-frequency harmonic components;

[0107] Through this microsecond-level high-frequency update and harmonic reconstruction, the system can provide the controller under test with an extremely realistic electromagnetic fingerprint signal containing the 5th and 7th harmonics and stator slot effect ripple. This effectively avoids the risk of traditional simulations masking the controller's filtering design defects due to overly idealized signals, and ensures a high degree of consistency between the test environment and the real physical environment.

[0108] Example 7:

[0109] The system runs on a heterogeneous computing platform based on field-programmable gate arrays (FPGAs); the manifold space construction module uses digital signal processing slices of the FPGA to perform parallel interpolation operations on metric tensors to achieve real-time simulation of multi-physics coupling.

[0110] This embodiment defines the hardware acceleration implementation of the system. The system is deployed on a heterogeneous computing platform combining ARM and FPGA, where the ARM core is responsible for low-speed tasks and the FPGA is responsible for manifold dynamics calculations. Specifically, the manifold space construction module calls the digital signal processing slices (DSPSlices) inside the FPGA to decompose the massive matrix operation task in the metric tensor construction process into multiple parallel multiply-accumulate pipelines. Here, DSPSlice comes from FPGA hardware resources, and its physical meaning is digital signal processing slice, with the unit being a slice.

[0111] By leveraging the locality of manifold computation, curvature interpolation in multiple dimensions can be performed simultaneously at the hardware level without interference, thus enabling real-time simulation of multi-physics coupling. This hardware parallelism-based implementation eliminates the timing jitter caused by traditional sequential software execution, ensuring that the single-step simulation step size remains stably locked within the microsecond range when handling extremely complex metric tensor operations, meeting the stringent requirements of hardware-in-the-loop testing for hard real-time performance.

[0112] Example 8:

[0113] The process also includes the following steps: acquiring control commands issued by the controller under test and obtaining the environmental and load parameters set for the test; establishing a high-dimensional Riemannian manifold space reflecting the energy state of the motor based on the motor's physical parameters, and converting the environmental and load parameters into a metric tensor of the manifold space, which is used to define the geometry and curvature of the manifold space; mapping external load disturbances and multi-physics coupling effects into dynamic changes in the metric tensor, using a symplectic geometric integrator to calculate the geodesic trajectory of the state point on the dynamically changing manifold space, which represents the evolution path of the motor's operating state, thereby solving the motor's operating state at the next moment; projecting the coordinates of the geodesic trajectory on the manifold space back into Euclidean space, restoring it to electrical signal physical quantities, and feeding back the electrical signal physical quantities to the controller under test through a hardware interface;

[0114] This embodiment provides specific execution steps for a motor operation state simulation test system based on manifold geometry; the system executes instructions and acquires parameters by acquiring PWM control instructions in real time through a high-frequency interface and synchronously reading the ambient temperature sequence and load torque fluctuation parameters set by the host computer; it performs manifold space initialization and mapping, constructs a high-dimensional Riemannian manifold space based on the motor physical parameters, and converts the real-time acquired environmental and load parameters into a metric tensor that defines the geometric structure of the space;

[0115] The dynamic evolution calculation is performed to map the physical field coupling effect into the dynamic change of the metric tensor. Using the symplectic geometric integrator as the starting point, the geodesic trajectory along the tangential motion is calculated to solve the running state at the next moment. The physical regression and closed-loop feedback are performed to restore the geodesic trajectory coordinates to physical quantities such as current and rotation speed, and convert them into analog voltages to be fed back to the controller under test.

[0116] The system establishes a complete physics-geometry-physics simulation closed loop. By transforming physical evolution into path tracing of geometric topology, it eliminates the excessive reliance on the serial computing power of microprocessors and achieves real-time, high-precision reproduction of extremely complex physical phenomena under limited computing resources.

[0117] Example 9:

[0118] The calculation of the geodesic trajectory of the state point on the dynamically changing manifold space using the symplectic geometric integrator includes: updating the metric tensor field of the manifold space in real time based on the current control command and environmental parameters; in the updated metric tensor field, starting from the current state point, performing an exponential mapping along the tangent vector direction to solve for the coordinates of the state point at the next moment, and the solution process follows the energy conservation constraints of symplectic geometry.

[0119] This embodiment is a further specification of the symplectic geometric integral calculation steps based on Embodiment 8; according to the current control commands and boundary conditions, the system updates the metric tensor field of the manifold space in real time, which describes the local geometric properties of each point on the manifold; wherein, The source is real-time updated calculation, the physical meaning is a metric tensor field, and the unit is dimensionless;

[0120] In the updated metric tensor field, the system performs a solution process that follows the symplectic geometric energy conservation constraint; specifically, starting from the current state point, it uses the partial derivatives of the Hamiltonian with respect to the generalized momentum and coordinates to perform an exponential mapping along the tangent vector direction, precisely solving for the coordinates of the state point at the next moment; where, The source is the Lie group exponential mapping, the physical meaning is the exponential mapping operation, and the unit is the dimensionless operator;

[0121] This step, by strictly adhering to symplectic geometric constraints, mathematically eliminates the possibility of artificial energy increases or losses during numerical integration, ensuring that even in long-term virtual durability tests, the simulation data will not exhibit non-physical numerical drift, thus guaranteeing the reliability of the test results.

[0122] Example 10:

[0123] The system also includes: real-time calculation of the geometric curvature of the trajectory of state points in the manifold space; setting a safe curvature threshold, recording normal operation data when the calculated geometric curvature is less than or equal to the safe curvature threshold; generating a potential fault warning signal when the calculated geometric curvature is greater than the safe curvature threshold, and recording the trajectory smoothness data at this time;

[0124] This embodiment further specifies the safety monitoring steps based on Embodiment 8. In each step of the simulation, the system calculates the geometric curvature of the coordinate point position modified in real time. Considering the real-time requirements of the FPGA, this embodiment does not calculate all Riemann tensors, but instead calculates the tensors along the current vector. With width vector The pseudo-curvature of the spanned plane The calculation formula is simplified to:

[0125] ;

[0126] Wherein, the denominator represents the fraction... and The square of the area of ​​Zhang Cheng's parallelogram, numerator The fully contracted form of the Riemann curvature tensor is specifically expanded and calculated as follows:

[0127] ;

[0128] above It calculates at most the principal contraction terms related to the current direction of motion, thereby completing the curvature calculation in microseconds.

[0129] The system incorporates a safety curvature threshold for determination; among which, The source is the system safety specification preset, the physical meaning is the safety curvature threshold, and the unit is the reciprocal square of the length;

[0130] When the calculated geometric curvature is less than or equal to the threshold, the system records it as normal operating data; conversely, when the geometric curvature is greater than the threshold, indicating that the state point has entered the dangerous area of ​​high curvature, the system immediately generates a potential fault warning signal and simultaneously records the trajectory smoothness data at this time.

[0131] This step endows the simulation system with self-diagnostic capabilities. By recording the degree of danger of the operating conditions during the process, i.e. curvature data, it provides engineers with a brand-new perspective on fault analysis, enabling them to quickly locate the specific operating conditions in which the motor approaches its physical limits, thereby assisting in the design of safer control strategies.

[0132] It should be noted that the above 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 preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A motor operating condition simulation test system, characterized in that: The system comprises: a signal acquisition module, configured to acquire a control instruction sent by a measured controller and to obtain an environment parameter and a load parameter set for testing; a manifold space construction module, configured to establish a high-dimensional Riemann manifold space reflecting an energy state of the motor based on motor physical parameters, and to convert the environment parameter and the load parameter into a metric tensor of the manifold space, the metric tensor being used to define a geometric structure and a bending degree of the manifold space; a dynamic curvature evolution unit, configured to map external load disturbance and multi-physical field coupling effects into dynamic changes of the metric tensor, to calculate a geodesic trajectory of a state point on the dynamically changed manifold space by using a symplectic geometric integrator, and to solve a running state of the motor at a next time point, the geodesic trajectory representing an evolution path of the running state of the motor; a physical quantity reconstruction module, configured to project a coordinate of the geodesic trajectory on the manifold space back to an Euclidean space, to restore the coordinate into an electrical signal physical quantity, and to feed back the electrical signal physical quantity to the measured controller through a hardware interface to form a closed loop test.

2. The motor operating condition simulation test system according to claim 1, characterized in that: The manifold space construction module comprises: an offline training unit, configured to perform multi-physical field coupling training in advance through a finite element model, to establish a mapping relationship between physical field parameters and manifold curvatures, and to generate a physical field-curvature lookup table; a state-geometry mapping unit, configured to retrieve the physical field-curvature lookup table according to the current environment parameter and the load parameter in a real-time simulation process, and to construct a metric tensor of a real-time manifold space, wherein when the environment temperature rises or the magnetic saturation degree increases, the metric component values and the curvatures of specific dimensions of the manifold space are adjusted to simulate nonlinear physical characteristics.

3. The motor operating condition simulation test system according to claim 1, characterized in that: The dynamic curvature evolution unit comprises: a tensor flow calculation subunit, configured to receive a pulse width modulation signal of the measured controller, to modify a distribution of the metric tensor in real time according to a duty ratio and a frequency of the pulse width modulation signal, and to form a dynamically changed Riemann metric field on the manifold space; a symplectic integral operator unit, configured to perform an exponential mapping operation along a tangent vector on a curved geometric surface to calculate manifold coordinates of the state point at a next time point, the symplectic integral operator unit maintaining energy conservation characteristics of the system in the calculation process to suppress numerical divergence.

4. The motor operating condition simulation test system according to claim 3, characterized in that: The dynamic curvature evolution unit further comprises: a fault simulation subunit, configured to apply a geometric distortion on a preset path of the manifold space, to manufacture a high-curvature geometric distortion area by modifying the metric tensor of a local area, and to force a motion trajectory of the state point to deviate from a normal geodesic to simulate a motor turn-to-turn short circuit or open circuit fault.

5. The motor operating condition simulation test system according to claim 1, characterized in that: The system further comprises: a stability evaluation module, configured to generate a manifold curvature distribution diagram and to monitor a geometric curvature change rate of an area where the state point is located in real time; wherein the stability evaluation module is preset with a curvature change threshold, and is configured to determine that the system is in a linear stable region when the geometric curvature change rate of the area where the state point is located is less than or equal to the curvature change threshold, and to determine that the system enters a critical unstable state when the geometric curvature change rate of the area where the state point is located is higher than the curvature change threshold, and to output a trajectory smoothness index as an evaluation basis for robustness of the controller.

6. The motor operating condition simulation test system according to claim 1, characterized in that: The physical quantity reconstruction module comprises: a coordinate projection unit, configured to map a coordinate of the state point in the high-dimensional manifold space into three-phase current values and a rotor position angle. A digital-to-analog conversion interface unit is configured to convert the three-phase current values and the rotor position angle into analog voltage signals and output the analog voltage signals to the controller under test at a microsecond-level update rate, and the output frequency of the digital-to-analog conversion interface unit is configured to be able to simulate high-frequency harmonic components.

7. The motor operating condition simulation test system according to claim 2, characterized in that: The system runs on a heterogeneous computing platform based on field programmable gate arrays. The manifold space construction module uses the digital signal processing slice of the field programmable gate array to perform parallel interpolation operations on the metric tensor, so as to realize real-time simulation of multi-physical field coupling.

8. A motor operating condition simulation test system, characterized by: The method further comprises the following steps: Control instructions issued by the controller under test are collected, and environmental parameters and load parameters of a test setting are obtained; A high-dimensional Riemannian manifold space reflecting the energy state of the motor is established based on motor physical parameters, and the environmental parameters and load parameters are converted into a metric tensor of the manifold space, the metric tensor being used to define the geometric structure and bending degree of the manifold space; External load disturbances and multi-physical field coupling effects are mapped into dynamic changes of the metric tensor, and a symplectic geometry integrator is used to calculate a geodesic trajectory of a state point on the dynamically changing manifold space, the geodesic trajectory representing an evolution path of the motor operating state, so as to solve the operating state of the motor at the next moment; Geodesic trajectory coordinates on the manifold space are projected back to the Euclidean space to restore the electrical signal physical quantities, and the electrical signal physical quantities are fed back to the controller under test through a hardware interface.

9. The motor operating condition simulation test system according to claim 8, characterized in that: The calculation of the geodesic trajectory of the state point on the dynamically changing manifold space by the symplectic geometry integrator comprises: The metric tensor field of the manifold space is updated in real time according to the current control instructions and environmental parameters; In the updated metric tensor field, exponential mapping is performed along the tangent vector direction from the current state point as the starting point to solve the state point coordinates at the next moment, and the solving process follows the energy conservation constraint of symplectic geometry.

10. The motor operating condition simulation test system of claim 8, wherein: The system further comprises: The geometric curvature of the state point trajectory in the manifold space is calculated in real time; A safety curvature threshold is set, when the calculated geometric curvature is less than or equal to the safety curvature threshold, the normal operating data is recorded, and when the calculated geometric curvature is greater than the safety curvature threshold, a potential fault warning signal is generated, and the trajectory smoothness data at this time is recorded.

Images