A method for identifying time-varying reactance parameters of synchronous motors based on the principle of superconducting circuits
By using a time-varying reactance parameter identification method for synchronous motors based on the superconducting circuit principle, and combining the finite element model and the surrogate model, the problem of low identification efficiency of reactance parameters of synchronous motors in dynamic processes is solved, and efficient and accurate reactance parameter identification and simulation are achieved.
Patent Information
- Application Number
- CN202611121534.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-27
- Publication Date
- 2026-11-10
AI Technical Summary
Existing methods for identifying reactance parameters of synchronous motors are insufficient to reflect changes in magnetic saturation state during dynamic processes, resulting in low efficiency in reactance parameter identification and inadequate simulation accuracy, especially in large power systems where the computational load is large and the efficiency is low.
Based on the principle of superconducting circuits, a transient electromagnetic finite element model of a synchronous motor is established to obtain the magnetic field distribution during the dynamic process. A surrogate model is constructed using convolutional neural networks and extreme learning machines to reflect the relationship between the motor's magnetic field and reactance parameters, thus forming a time-varying characteristic curve.
It improves the efficiency of synchronous motor reactance parameter identification and dynamic simulation accuracy, reduces the computational load, and is suitable for synchronous motor dynamic performance analysis and power system simulation.
Smart Images

Figure CN123069736A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of synchronous motor parameter identification, and specifically relates to a method for identifying time-varying reactance parameters of synchronous motors based on the principle of superconducting circuits. Background Technology
[0002] Synchronous motors are key power generation equipment in power systems, and their operating status directly affects the transient stability, voltage stability, and effectiveness of protection and control strategies of the power system. When a synchronous motor experiences dynamic processes such as a sudden rise or fall in terminal voltage, a sudden change in load, or a sudden three-phase short circuit, the armature winding current, excitation winding current, and damping circuit current change rapidly. This alters the magnetic field distribution in the stator core, rotor core, and air gap region, leading to dynamic changes in the degree of magnetic circuit saturation. Since the magnetic circuit saturation state is closely related to the reactance parameters of the synchronous motor, the synchronous reactance, transient reactance, and ultra-transient reactance parameters of the synchronous motor are not always constant but exhibit certain time-varying characteristics during dynamic processes.
[0003] Existing methods for identifying the reactance parameters of synchronous motors mainly include three-phase sudden short-circuit tests, frequency domain response methods, time domain response methods, and parameter calculation methods based on finite element models. Among them, the three-phase sudden short-circuit test can obtain some reactance parameters of the synchronous motor based on the short-circuit current decay process; the frequency domain response method and the time domain response method can establish a motor parameter model through external excitation and response data; the finite element model can perform relatively precise calculations of the internal electromagnetic field distribution of the motor by combining the motor geometry, winding connection method, material magnetization characteristics, and operating conditions.
[0004] However, traditional experimental identification and response analysis methods typically only yield a set of equivalent fixed parameters under a specific operating condition, making it difficult to continuously reflect the evolution of reactance parameters caused by magnetic saturation changes in synchronous motors during actual dynamic processes. For dynamic processes such as sudden voltage changes at the generator terminals, short-circuit impacts, and excitation regulation, if fixed reactance parameters are still used for transient analysis, it is difficult to accurately characterize the correspondence between the internal magnetic field state and parameter changes of the motor during the super-transient stage, transient stage, and subsequent recovery stage. This affects the assessment of short-circuit current peaks, the formulation of excitation control strategies, the verification of protection settings, and the accuracy of system transient simulation.
[0005] Meanwhile, although refined finite element models can accurately reflect the internal magnetic field distribution and local magnetic saturation state of synchronous motors, performing complete finite element parameter calculations at every time point in the dynamic process, or using refined finite element models to jointly solve multiple synchronous motors in large-scale power system transient simulations, will generate a large computational load and time cost, making it difficult to meet the requirements of computational efficiency and simulation scale in engineering applications. Therefore, how to quickly and accurately obtain the characteristics of synchronous motor reactance parameters changing over time while preserving the dynamic magnetic saturation state information of the motor is a key problem that needs to be solved in synchronous motor parameter identification and power system dynamic simulation.
[0006] Therefore, it is necessary to propose a time-varying reactance parameter identification method for synchronous motors. This method can obtain the d-axis and q-axis reactance parameters of the synchronous motor at different time points based on the distribution of the internal magnetic field of the motor during dynamic processes, combined with techniques such as frozen permeability, superconducting loop parameter calculation, and surrogate model mapping. It can also form a time-varying characteristic curve of the reactance parameters changing with the dynamic process, thereby solving the technical problems in existing technologies, such as the difficulty of fixed parameter identification methods in characterizing dynamic time-varying characteristics, the low efficiency of fine finite element point-by-point calculation, and the difficulty of balancing accuracy and computational efficiency in transient simulation of large power systems. Summary of the Invention
[0007] The technical problem to be solved by the present invention is to provide a method for identifying time-varying reactance parameters of synchronous motors based on the principle of superconducting circuits. The present invention can reflect the influence of changes in magnetic saturation state on reactance parameters during the dynamic process of synchronous motors, and improve the efficiency of time-varying reactance parameter identification and the accuracy of dynamic simulation parameters.
[0008] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A method for identifying time-varying reactance parameters of a synchronous motor based on the principle of superconducting circuits, comprising the following steps: S1. Based on the electromagnetic structure parameters of the synchronous motor, establish the electromagnetic transient finite element solution model of the synchronous motor. Perform time-domain finite element simulation under the condition of sudden change in terminal voltage to obtain the two-dimensional cross-sectional magnetic density cloud map of the motor at different time points during the dynamic process. S2. Select multiple sample time points from the dynamic process, obtain the rotor position angle corresponding to each sample time point, and freeze the permeability of the stator and rotor in the two-dimensional plane of the motor at each sample time point. S3. The frozen permeability at each sample time point is assigned to a new motor model at the same rotor position angle. Based on the superconducting circuit principle, the rotor position, winding connection state and material electromagnetic properties are set. DC excitation is applied to the three-phase stator windings, steady-state current and flux linkage data are collected, and the d-axis and q-axis reactance parameters corresponding to each sample time point are calculated. S4. Construct a neural network composed of a convolutional neural network (CNN) and an extreme learning machine (ELM). Use the two-dimensional cross-sectional magnetic flux density map of the motor corresponding to each sample time point as the input dataset and the corresponding d-axis and q-axis reactance parameters as the output training set to train the neural network and generate a surrogate model to characterize the correspondence between the motor magnetic flux density map and the reactance parameters. S5. Input the two-dimensional cross-sectional magnetic density cloud map of the motor corresponding to other time points in the dynamic process into the proxy model to obtain the d-axis and q-axis reactance parameter values at the corresponding time points, and form the time-varying characteristic curve of the synchronous motor reactance parameter changing with the dynamic process.
[0009] Preferably, in step S1, when obtaining the magnetic flux density cloud map of the two-dimensional cross-section of the motor at different time points, the magnetic flux density cloud map of the two-dimensional cross-section of the motor adopts the same upper limit value, lower limit value and magnetic flux density change step size, so as to ensure that the image features of the input dataset have a uniform scale.
[0010] Preferably, in step S2, the plurality of sample time points include sample time points where the rotor magnetic pole axis is aligned with the A-phase axis, and sample time points where the rotor inter-pole axis is aligned with the A-phase axis; wherein, the sample time point selection density at the beginning of the transient process is greater than the sample time point selection density in the subsequent stages of the transient process.
[0011] Preferably, the number of sample time points where the rotor magnetic pole axis is aligned with the A-phase axis is 30 to 60, and the number of sample time points where the rotor inter-pole axis is aligned with the A-phase axis is 30 to 60.
[0012] Preferably, in step S3, the d-axis and q-axis reactance parameters include at least one of d-axis synchronous reactance, q-axis synchronous reactance, d-axis transient reactance, d-axis ultra-transient reactance, and q-axis ultra-transient reactance.
[0013] Preferably, when calculating the d-axis synchronous reactance, the stator and rotor permeability corresponding to when the rotor magnetic pole axis is aligned with the A-phase axis is used, the motor speed is set to 0, and the excitation winding is set to an open circuit state. If the synchronous motor is a salient-pole synchronous motor, the solid rotor material is set to be magnetically permeable but not electrically conductive, and the slot wedge material is set to air. If the synchronous motor is a salient-pole synchronous motor, the damping winding material is set to air. DC power is applied to the A, B, and C phase windings respectively, so that the three-phase currents in steady state are respectively... , , At this point, the A-phase current is equivalent to the d-axis current, the A-phase flux linkage is equivalent to the d-axis flux linkage, and the per-unit value of the d-axis synchronous reactance is calculated according to the following formula:
[0014] In the formula, This is the per-unit value of the d-axis synchronous reactance. This represents the per-unit value corresponding to the nominal value of the d-axis flux linkage when the armature current is stable. The armature current has a nominal value The corresponding per-unit value.
[0015] Preferably, when calculating the q-axis synchronous reactance, the stator and rotor permeability corresponding to the rotor pole axis being aligned with the A-phase axis are imported, and the motor speed is set to 0. If the synchronous motor is a salient-pole synchronous motor, the solid rotor material is set to be magnetically permeable but non-conductive, and the slot wedge material is set to air. If the synchronous motor is a salient-pole synchronous motor, the damping winding material is set to air. DC power is applied to the A, B, and C phase windings respectively, so that the three-phase currents in steady state are respectively... , , At this point, the A-phase current is equivalent to the q-axis current, and the A-phase flux linkage is equivalent to the q-axis flux linkage. The per-unit value of the q-axis synchronous reactance is calculated according to the following formula: ; In the formula, This is the per-unit value of the q-axis synchronous reactance. This is the per-unit value corresponding to the nominal value of the q-axis flux linkage when the armature current is stable. The armature current has a nominal value The corresponding per-unit value.
[0016] Preferably, when calculating the d-axis transient reactance, the stator and rotor permeability corresponding to when the rotor magnetic pole axis is aligned with the A-phase axis is used. The motor speed is set to 0, the excitation winding is set to a short-circuited state, and the excitation winding resistance is set to 0. If the synchronous motor is a salient-pole synchronous motor, the solid rotor material is set to be magnetically permeable but non-conductive, and the slot wedge material is set to air. If the synchronous motor is a salient-pole synchronous motor, the damping winding material is set to air. DC power is applied to the A, B, and C phase windings respectively, so that the three-phase currents in steady state are respectively... , , At this point, the A-phase current is equivalent to the d-axis current, and the A-phase flux linkage is equivalent to the d-axis transient flux linkage. The per-unit value of the d-axis transient reactance is calculated according to the following formula:
[0017] In the formula, This represents the per-unit value of the d-axis transient reactance. This represents the per-unit value corresponding to the nominal value of the transient flux linkage along the d-axis when the armature current is stable. The armature current has a nominal value The corresponding per-unit value.
[0018] Preferably, when calculating the d-axis transient reactance, the stator and rotor permeability corresponding to when the rotor magnetic pole axis is aligned with the A-phase axis are imported, the motor speed is set to 0, the excitation winding is set to a short-circuited state, and the excitation winding resistance is set to 0; if the synchronous motor is a salient-pole synchronous motor, the solid rotor material resistance is set to 0, and the slot wedge material resistance is set to 0; if the synchronous motor is a salient-pole synchronous motor, the damping winding material resistance is set to 0; DC power is applied to the A, B, and C phase windings respectively, so that the three-phase currents in steady state are respectively , , At this point, the A-phase current is equivalent to the d-axis current, and the A-phase flux linkage is equivalent to the d-axis ultra-transient flux linkage. The per-unit value of the d-axis ultra-transient reactance is calculated according to the following formula:
[0019] In the formula, This represents the per-unit value of the d-axis ultratransient reactance. This represents the per-unit value corresponding to the named value of the d-axis transient flux linkage when the armature current is stable. The armature current has a nominal value The corresponding per-unit value.
[0020] Preferably, when calculating the q-axis transient reactance, the stator and rotor permeability corresponding to the rotor pole axis being aligned with the A-phase axis are imported, and the motor speed is set to 0; if the synchronous motor is a salient-pole synchronous motor, the resistivity of the solid rotor material is set to 0, and the resistivity of the slot wedge material is set to 0; if the synchronous motor is a salient-pole synchronous motor, the resistivity of the damping winding material is set to 0; DC power is applied to the A, B, and C phase windings respectively, so that the three-phase currents in steady state are respectively , , At this point, the A-phase current is equivalent to the q-axis current, and the A-phase flux linkage is equivalent to the q-axis transient flux linkage. The per-unit value of the q-axis transient reactance is calculated according to the following formula:
[0021] In the formula, This represents the per-unit value of the q-axis transient reactance. This represents the per-unit value corresponding to the nominal value of the q-axis transient flux linkage when the armature current is stable. The armature current has a nominal value The corresponding per-unit value.
[0022] Preferably, the three-phase current formed by passing a DC power supply through the A, B, and C three-phase windings is... , , Used to form an equivalent magnetic field direction corresponding to the A-phase axis on the stator side; when the rotor pole axis is aligned with the A-phase axis, the A-phase current and A-phase flux linkage are used as d-axis current and d-axis flux linkage in reactance calculation, respectively; when the rotor pole axis is aligned with the A-phase axis, the A-phase current and A-phase flux linkage are used as q-axis current and q-axis flux linkage in reactance calculation, respectively.
[0023] Preferably, in step S4, the convolutional neural network (CNN) is used to extract the spatial magnetic field distribution features in the magnetic density cloud map of the two-dimensional cross-section of the motor, and the extreme learning machine (ELM) is used to output the corresponding d-axis and q-axis reactance parameters based on the spatial magnetic field distribution features.
[0024] Preferably, in step S5, the time-varying characteristic curves of the synchronous motor reactance parameters as a function of the dynamic process include at least one of the following: d-axis synchronous reactance time-varying curve, q-axis synchronous reactance time-varying curve, d-axis transient reactance time-varying curve, d-axis ultra-transient reactance time-varying curve, and q-axis ultra-transient reactance time-varying curve.
[0025] A system for identifying time-varying reactance parameters of a synchronous motor based on the superconducting loop principle is disclosed. The system employs the aforementioned method for identifying time-varying reactance parameters of a synchronous motor based on the superconducting loop principle. It includes a finite element simulation module, a permeability freezing module, a superconducting loop parameter calculation module, a surrogate model training module, and a time-varying parameter identification module, which are sequentially and interactively linked. Wherein: The finite element simulation module is used to establish a transient electromagnetic finite element solution model of the synchronous motor based on the electromagnetic structure parameters of the synchronous motor. It performs time-domain finite element simulation under the condition of sudden change in terminal voltage and obtains the two-dimensional cross-sectional magnetic density cloud map of the motor at different time points during the dynamic process. The permeability freezing module is used to select multiple sample time points from the dynamic process, obtain the rotor position angle corresponding to each sample time point, and freeze the permeability of the stator and rotor in the two-dimensional plane of the motor at each sample time point. The superconducting loop parameter calculation module is used to assign the frozen permeability at each sample time point to a new motor model at the same rotor position angle. Based on the superconducting loop principle, the rotor position, winding connection state and material electromagnetic properties are set, and DC excitation is applied to the three-phase stator windings. Steady-state current and flux linkage data are collected. The d-axis and q-axis reactance parameters corresponding to each sample time point are calculated according to the ratio of the per-unit reactance value to the per-unit current value of the corresponding axis flux linkage. The surrogate model training module is used to construct a neural network composed of a convolutional neural network (CNN) and an extreme learning machine (ELM). The two-dimensional cross-sectional magnetic flux density map of the motor corresponding to each sample time point is used as the input dataset, and the corresponding d-axis and q-axis reactance parameters are used as the output training set to train the neural network and generate a surrogate model to characterize the correspondence between the motor magnetic flux density map and the reactance parameters. The time-varying parameter identification module is used to input the two-dimensional cross-sectional magnetic density cloud map of the motor corresponding to other time points in the dynamic process into the proxy model, obtain the d-axis and q-axis reactance parameter values at the corresponding time points, and form the time-varying characteristic curve of the synchronous motor reactance parameter changing with the dynamic process.
[0026] Preferably, the superconducting loop parameter calculation module includes a rotor position setting unit, a winding state setting unit, a material property setting unit, a three-phase DC excitation unit, and a flux linkage acquisition unit; wherein: The rotor position setting unit is used to set the rotor magnetic pole axis to align with the A-phase axis or the rotor inter-pole axis to align with the A-phase axis. The winding state setting unit is used to set the excitation winding to an open circuit state or a short circuit state. The material property setting unit is used to set the electromagnetic properties of solid rotor material, slot wedge material, or damping winding material according to the type of synchronous motor. The three-phase DC excitation unit is used to apply DC power to the A, B, and C phase windings respectively, so that the three-phase currents in steady state are respectively , , ; The flux linkage acquisition unit is used to acquire the flux linkage of phase A in steady state and calculate the corresponding per-unit value of reactance parameter in conjunction with the per-unit value of phase A current.
[0027] A computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the method for identifying time-varying reactance parameters of a synchronous motor based on the principle of a superconducting circuit.
[0028] The present invention can achieve the following beneficial effects: 1. This invention establishes an electromagnetic transient finite element solution model of a synchronous motor under the condition of sudden voltage change at the motor terminal, and obtains the two-dimensional cross-sectional magnetic density cloud map of the motor at different time points during the dynamic process. It can use the magnetic field distribution state of the synchronous motor in the transient process as the basic data for identifying the reactance parameters, so that the identification results can reflect the actual situation of the change of the magnetic circuit saturation state inside the synchronous motor with time.
[0029] 2. This invention freezes the permeability of the stator and rotor at the sample time point and assigns the frozen permeability to a new motor model at the same rotor position angle. This allows the magnetic saturation state corresponding to each sample time point to be maintained in the subsequent reactance parameter calculation process, thereby avoiding the problem that traditional fixed parameter identification methods cannot reflect dynamic magnetic saturation changes and improving the correspondence between the synchronous motor reactance parameter identification results and the actual dynamic process.
[0030] 3. This invention sets the rotor position, winding connection state, and material electromagnetic properties based on the superconducting circuit principle. By applying DC excitation to the three-phase stator windings and collecting steady-state current and flux linkage data, it calculates parameters such as d-axis synchronous reactance, q-axis synchronous reactance, d-axis transient reactance, d-axis ultra-transient reactance, and q-axis ultra-transient reactance. It can identify multiple types of reactance parameters under a unified calculation framework, improving the systematicness and applicability of the parameter calculation process.
[0031] 4. This invention uses the sample time point when the rotor magnetic pole axis is aligned with the A-phase axis for calculating the d-axis reactance parameter, and the sample time point when the rotor inter-pole axis is aligned with the A-phase axis for calculating the q-axis reactance parameter. Furthermore, the A-phase current and A-phase flux linkage are used to participate in the corresponding axis reactance calculation, so that the calculation of the d-axis and q-axis reactance parameters has a clear axial correspondence, reducing the impact of rotor position changes on the accuracy of parameter identification during dynamic processes.
[0032] 5. This invention constructs a surrogate model composed of a Convolutional Neural Network (CNN) and an Extreme Learning Machine (ELM). Using the two-dimensional cross-sectional magnetic flux density map of the motor as input and the corresponding d-axis and q-axis reactance parameters as output, it is trained to establish a nonlinear mapping relationship between the spatial distribution characteristics of the magnetic field and the reactance parameters. After training, for other time points in the dynamic process, only the corresponding magnetic flux density map needs to be input to obtain the corresponding reactance parameters, reducing the computational burden of repeatedly calculating the superconducting circuit parameters at each time point.
[0033] 6. This invention utilizes CNN to extract spatial features of the stator region, rotor region, air gap region, and local magnetic saturation region in the two-dimensional cross-sectional magnetic density cloud map of the motor, and uses ELM to complete the fast mapping between features and reactance parameters. This enables the surrogate model to have both magnetic field feature expression capability and parameter output efficiency, and can improve the analysis speed of time-varying parameters of synchronous motor while ensuring identification accuracy.
[0034] 7. This invention can generate time-varying characteristic curves of synchronous motor reactance parameters as they change with dynamic processes, including at least one of the following: d-axis synchronous reactance time-varying curve, q-axis synchronous reactance time-varying curve, d-axis transient reactance time-varying curve, d-axis ultra-transient reactance time-varying curve, and q-axis ultra-transient reactance time-varying curve. This provides parameter basis for synchronous motor dynamic performance analysis, short-circuit current peak assessment, excitation control strategy optimization, protection setting verification, and power system transient simulation parameter correction.
[0035] 8. This invention establishes a connection between fine finite element modeling and rapid mapping of proxy models. It can reflect the changes in the electromagnetic field and magnetic saturation state inside the synchronous motor using the finite element method, while avoiding the need to perform complete fine finite element calculations for every time point and every synchronous motor in large-scale power system simulations. This achieves a good balance between calculation accuracy and calculation efficiency and has high engineering application value. Attached Figure Description
[0036] The present invention will be further described below with reference to the accompanying drawings and embodiments: Figure 1 This is a flowchart of the method for identifying time-varying reactance parameters of a synchronous motor based on the principle of superconducting circuits, according to the present invention. Figure 2 This is a structural block diagram of the synchronous motor time-varying reactance parameter identification system based on the superconducting circuit principle of the present invention; Figure 3 This is a schematic diagram of the electromagnetic transient finite element simulation and magnetic density cloud map acquisition process of the synchronous motor in this invention; Figure 4 This is a schematic diagram of the sample time point selection, rotor position angle acquisition, and permeability freezing process in this invention; Figure 5 This is a schematic diagram of the reactance parameter calculation process based on the superconducting circuit principle in this invention; Figure 6 This is a schematic diagram showing the correspondence between the rotor position when calculating the d-axis reactance parameters and the q-axis reactance parameters in this invention; Figure 7 This is a schematic diagram of the training and reactance parameter mapping process of the CNN-ELM surrogate model in this invention; Figure 8 This is a schematic diagram of the time-varying reactance parameter identification result curve of the synchronous motor in this invention.
[0037] In the diagram: 100 - Synchronous motor time-varying reactance parameter identification system; 110 - Finite element simulation module; 120 - Permeability freezing module; 130 - Superconducting circuit parameter calculation module; 140 - Surrogate model training module; 150 - Time-varying parameter identification module; 131 - Rotor position setting unit; 132 - Winding state setting unit; 133 - Material property setting unit; 134 - Three-phase DC excitation unit; 135 - Magnetic flux acquisition unit. Detailed Implementation
[0038] like Figure 1As shown, a time-varying reactance parameter identification method for synchronous motors based on the superconducting loop principle is applicable to the dynamic process reactance parameter identification of synchronous motors after a sudden change in terminal voltage. This method is based on the electromagnetic transient finite element solution model of the synchronous motor, obtaining two-dimensional cross-sectional magnetic flux density maps of the motor at different times during the dynamic process; by freezing the permeability of the stator and rotor at each sample time point, the magnetic saturation state corresponding to that time point is maintained; then, based on the superconducting loop principle, the reactance parameters under different axial directions, different winding connection states, and different material electromagnetic properties are calculated; finally, using a surrogate model composed of a convolutional neural network (CNN) and an extreme learning machine (ELM), the correspondence between the two-dimensional cross-sectional magnetic flux density map of the motor and the d-axis and q-axis reactance parameters is established, thereby obtaining the time-varying characteristic curves of the synchronous motor reactance parameters as the dynamic process changes.
[0039] The synchronous motor can be either a salient-pole synchronous motor or a non-salient-pole synchronous motor. For salient-pole synchronous motors, the damping winding material properties are set according to corresponding requirements during different reactance calculations; for non-salient-pole synchronous motors, the electromagnetic properties of the solid rotor material and slot wedge material are set according to corresponding requirements during different reactance calculations. The operating conditions where the terminal voltage changes abruptly include sudden drops in terminal voltage, sudden increases in terminal voltage, sudden three-phase short circuits under no-load conditions, and sudden load changes, which can cause dynamic changes in the distribution of the magnetic field inside the motor.
[0040] like Figure 1 As shown, this method includes the following steps.
[0041] S1. Establishment of electromagnetic transient finite element solution model for synchronous motor and acquisition of magnetic density cloud map.
[0042] An electromagnetic transient finite element model of the synchronous motor is established based on its electromagnetic structural parameters. These electromagnetic structural parameters include the stator core geometry, rotor core geometry, air gap length, stator slot structure, rotor pole structure, excitation winding structure, damping winding structure, stator three-phase winding connection method, core material magnetization curve, and conductor material electromagnetic parameters.
[0043] After establishing the electromagnetic transient finite element solution model of the synchronous motor, a dynamic operating condition with a sudden change in the terminal voltage is set, and time-domain finite element simulation is performed. During the simulation, the two-dimensional cross-sectional magnetic flux density distribution data of the motor at different time points are output according to the preset time step, and a two-dimensional cross-sectional magnetic flux density cloud map of the motor at the corresponding time point is generated based on the magnetic flux density distribution data. The two-dimensional cross-sectional magnetic flux density cloud map of the motor is used to characterize the spatial magnetic field distribution state of the stator core, rotor core, air gap, and winding region of the motor at that time point.
[0044] To ensure that the image features of subsequent input datasets have a uniform scale, the magnetic flux density cloud maps of the two-dimensional cross-section of the motor at different time points adopt consistent upper and lower magnetic flux density values, as well as magnetic flux density change step sizes. Specifically, a uniform magnetic flux density display range is determined within the magnetic flux density data range of the entire dynamic process, and the same color mapping rules and magnetic flux density grading rules are applied to the magnetic flux density cloud maps output at each time point. This ensures that the image differences between magnetic flux density cloud maps at different time points are caused by changes in the actual distribution of the magnetic field, rather than by differences in image display scale.
[0045] S2, Sample time point selection, rotor position angle acquisition, and permeability freezing.
[0046] Multiple sample time points are selected from the dynamic process. These sample time points provide the input samples and output labels required for training the surrogate model. For each sample time point, the rotor position angle corresponding to that time point is obtained, and the permeability of the stator and rotor in the two-dimensional plane of the motor at that time point is frozen.
[0047] The permeability freezing refers to extracting the local permeability distribution of the stator and rotor core regions at a specific time point in the electromagnetic transient finite element solution model, and maintaining this permeability distribution in the subsequent static reactance calculation model so that it does not re-iterate with DC excitation. Through this process, the magnetic saturation state corresponding to the sample time point is preserved, enabling subsequent reactance parameter calculations to reflect the actual magnetic circuit saturation level at that time point during the dynamic process.
[0048] The multiple sample time points include two categories: the first category is sample time points where the rotor magnetic pole axis is aligned with the A-phase axis, and the second category is sample time points where the rotor inter-pole axis is aligned with the A-phase axis. The sample time points where the rotor magnetic pole axis is aligned with the A-phase axis are used for d-axis reactance parameter calculation, and the sample time points where the rotor inter-pole axis is aligned with the A-phase axis are used for q-axis reactance parameter calculation.
[0049] When selecting sample time points, the sample time point selection density at the beginning of the transient process is greater than that in the subsequent stages. This is because after a sudden change in terminal voltage, the armature current, excitation winding current, and internal magnetic field of the motor change rapidly in the initial stage. Using a higher density of sample time points improves the surrogate model's ability to represent the rapidly changing parameter range. In the later stages of the dynamic process, the rate of magnetic field change decreases, and using a relatively lower density of sample time points reduces the computational load.
[0050] In one implementation, the number of sample time points aligning the rotor magnetic pole axis with the A-phase axis is 30 to 60, and the number of sample time points aligning the rotor inter-pole axis with the A-phase axis is also 30 to 60. These two types of sample time points correspond to the magnetic circuit states in the d-axis and q-axis directions, respectively, and are used to subsequently obtain training labels for the d-axis and q-axis reactance parameters.
[0051] S3. Calculation of sample time-point reactance parameters based on the principle of superconducting circuits The frozen permeability at each sample time point is assigned to a new motor model at the same rotor position angle. This new motor model is used for static electromagnetic field calculations under the frozen permeability conditions, and its geometry, winding structure, and material region division are consistent with the synchronous motor electromagnetic transient finite element solution model. For each sample time point, the frozen stator and rotor permeability at that time point are assigned to the corresponding stator core and rotor core regions in the new motor model, respectively.
[0052] In the new motor model, the rotor position, winding connection state, and material electromagnetic properties are set based on the superconducting circuit principle. DC excitation is applied to the three-phase windings A, B, and C of the stator, and steady-state current and flux linkage data are collected to calculate the d-axis and q-axis reactance parameters at each sample time point. The d-axis and q-axis reactance parameters include at least one of the following: d-axis synchronous reactance, q-axis synchronous reactance, d-axis transient reactance, d-axis ultra-transient reactance, and q-axis ultra-transient reactance.
[0053] When DC excitation is applied to the three-phase stator windings, DC power is supplied to the A, B, and C phase windings respectively, so that the three-phase currents in steady state are respectively , , The aforementioned three-phase DC current is used to form an equivalent magnetic field direction corresponding to the A-phase axis on the stator side. When the rotor pole axis is aligned with the A-phase axis, the A-phase current and A-phase flux linkage are used as d-axis current and d-axis flux linkage in reactance calculation, respectively; when the rotor pole axis is aligned with the A-phase axis, the A-phase current and A-phase flux linkage are used as q-axis current and q-axis flux linkage in reactance calculation, respectively.
[0054] When calculating reactance parameters, the motor speed is set to 0. After the current reaches steady state under three-phase DC excitation, the A-phase flux linkage is collected, and the A-phase flux linkage is equivalent to the d-axis flux linkage, q-axis flux linkage, d-axis transient flux linkage, d-axis ultra-transient flux linkage, or q-axis ultra-transient flux linkage according to the rotor position correspondence. Then, the per-unit value of the corresponding reactance parameter is calculated based on the ratio of the per-unit value of the corresponding axis flux linkage to the per-unit value of the armature current.
[0055] Specifically, the calculation method for the d-axis synchronous reactance is as follows.
[0056] Import the stator and rotor permeability corresponding to the rotor magnetic pole axis aligned with the A-phase axis, set the motor speed to 0, and set the excitation winding to an open circuit state. If the synchronous motor is a salient-pole synchronous motor, set the solid rotor material to be magnetically permeable but non-conductive, and set the slot wedge material to air; if the synchronous motor is a salient-pole synchronous motor, set the damping winding material to air. Apply DC power to the A, B, and C phase windings respectively, so that the three-phase currents in steady state are respectively... , , At this point, the phase A current is equivalent to the d-axis current, the phase A flux linkage is equivalent to the d-axis flux linkage, and the per-unit value of the d-axis synchronous reactance is calculated according to the following formula:
[0057] In the formula, This is the per-unit value of the d-axis synchronous reactance. This represents the per-unit value corresponding to the nominal value of the d-axis flux linkage when the armature current is stable. The armature current has a named value The corresponding per-unit value.
[0058] The calculation method for the q-axis synchronous reactance is as follows.
[0059] Import the stator and rotor permeability corresponding to the alignment of the rotor pole axes with the A-phase axis, and set the motor speed to 0. If the synchronous motor is a salient-pole synchronous motor, set the solid rotor material to be magnetically permeable but non-conductive, and the slot wedge material to be air; if the synchronous motor is a salient-pole synchronous motor, set the damping winding material to air. Apply DC power to the A, B, and C phase windings respectively, so that the three-phase currents in steady state are respectively... , , At this point, the phase A current is equivalent to the q-axis current, the phase A flux linkage is equivalent to the q-axis flux linkage, and the per-unit value of the q-axis synchronous reactance is calculated according to the following formula:
[0060] In the formula, This is the per-unit value of the q-axis synchronous reactance. This is the per-unit value corresponding to the nominal value of the q-axis flux linkage when the armature current is stable. The armature current has a named value The corresponding per-unit value.
[0061] The calculation method for d-axis transient reactance is as follows.
[0062] To determine the stator and rotor permeability corresponding to aligning the rotor pole axes with the A-phase axis, set the motor speed to 0, the excitation winding to a short-circuited state, and the excitation winding resistance to 0. If the synchronous motor is a salient-pole synchronous motor, set the solid rotor material to be magnetically permeable but non-conductive, and the slot wedge material to air; if the synchronous motor is a salient-pole synchronous motor, set the damping winding material to air. Apply DC power to the A, B, and C phase windings respectively, so that the steady-state three-phase currents are respectively... , , At this point, the A-phase current is equivalent to the d-axis current, and the A-phase flux linkage is equivalent to the d-axis transient flux linkage. The per-unit value of the d-axis transient reactance is calculated according to the following formula:
[0063] In the formula, This represents the per-unit value of the d-axis transient reactance. This represents the per-unit value corresponding to the nominal value of the transient flux linkage along the d-axis when the armature current is stable. The armature current has a named value The corresponding per-unit value.
[0064] The calculation method for the d-axis transient reactance is as follows.
[0065] Import the stator and rotor permeability corresponding to the rotor magnetic pole axis aligned with the A-phase axis. Set the motor speed to 0, the excitation winding to a short-circuited state, and the excitation winding resistance to 0. If the synchronous motor is a salient-pole synchronous motor, set the resistivity of the solid rotor material to 0 and the resistivity of the slot wedge material to 0; if the synchronous motor is a salient-pole synchronous motor, set the resistivity of the damping winding material to 0. Apply DC power to the A, B, and C phase windings respectively, so that the three-phase currents in steady state are respectively... , , At this point, the phase A current is equivalent to the d-axis current, and the phase A flux linkage is equivalent to the d-axis ultra-transient flux linkage. The per-unit value of the d-axis ultra-transient reactance is calculated according to the following formula:
[0066] In the formula, This represents the per-unit value of the d-axis ultra-transient reactance. This represents the per-unit value corresponding to the named value of the d-axis transient flux linkage when the armature current is stable. The armature current has a named value The corresponding per-unit value.
[0067] The calculation method for q-axis transient reactance is as follows.
[0068] Import the stator and rotor permeabilities corresponding to the alignment of the rotor pole axes with the A-phase axis, and set the motor speed to 0. If the synchronous motor is a salient-pole synchronous motor, set the resistivity of the solid rotor material to 0 and the resistivity of the slot wedge material to 0; if the synchronous motor is a salient-pole synchronous motor, set the resistivity of the damping winding material to 0. Apply DC power to the A, B, and C phase windings respectively, so that the three-phase currents in steady state are respectively... , , At this point, the A-phase current is equivalent to the q-axis current, and the A-phase flux linkage is equivalent to the q-axis transient flux linkage. The per-unit value of the q-axis transient reactance is calculated according to the following formula:
[0069] In the formula, This represents the per-unit value of the q-axis transient reactance. This represents the per-unit value corresponding to the nominal value of the q-axis transient flux linkage when the armature current is stable. The armature current has a named value The corresponding per-unit value.
[0070] S4, CNN-ELM surrogate model construction and training A neural network consisting of a Convolutional Neural Network (CNN) and an Extreme Learning Machine (ELM) is constructed. The two-dimensional cross-sectional magnetic flux density map of the motor at each sample time point is used as the input dataset, and the d-axis and q-axis reactance parameters at each sample time point calculated in step S3 are used as the output training set to train the neural network, generating a surrogate model representing the correspondence between the motor's magnetic flux density map and the reactance parameters.
[0071] The convolutional neural network (CNN) is used to extract the spatial magnetic field distribution features from the two-dimensional cross-sectional magnetic flux density map of the motor. Specifically, the CNN performs convolution processing and feature dimensionality reduction processing on the input magnetic flux density map to extract the spatial distribution features of the stator core region, rotor core region, air gap region, and magnetic saturation concentration region, and converts the two-dimensional magnetic flux density map into a feature vector that can characterize the magnetic field distribution state.
[0072] The Extreme Learning Machine (ELM) is used to output corresponding d-axis and q-axis reactance parameters based on the spatial magnetic field distribution characteristics. The feature vector output by the CNN is input into the ELM, which performs a nonlinear mapping from the spatial magnetic field distribution characteristics to per-unit values of the reactance parameters. The output reactance parameters include at least one of the following: d-axis synchronous reactance, q-axis synchronous reactance, d-axis transient reactance, d-axis ultra-transient reactance, and q-axis ultra-transient reactance.
[0073] During training, sample pairs are formed by pairing the magnetic flux density cloud map corresponding to the sample time point with the reactance parameters calculated in step S3. Each sample pair includes an input magnetic flux density cloud map and at least one corresponding reactance parameter label. Through training, the surrogate model obtains the correspondence between different magnetic flux density distribution states and different reactance parameters. After training, the surrogate model can output the reactance parameter values for the corresponding time point based on the input magnetic flux density cloud map without recalculating the superconducting circuit parameters for each time point.
[0074] S5. Identification of time-varying reactance parameters at other time points during the dynamic process.
[0075] Input the two-dimensional cross-sectional magnetic flux density cloud map of the motor corresponding to other time points besides the sample time point into the surrogate model trained in step S4 to obtain the d-axis and q-axis reactance parameter values at the corresponding time points. Arrange the reactance parameter values output by the surrogate model in chronological order to form the time-varying characteristic curve of the synchronous motor reactance parameters changing with the dynamic process.
[0076] The time-varying characteristic curves include at least one of the following: d-axis synchronous reactance time-varying curve, q-axis synchronous reactance time-varying curve, d-axis transient reactance time-varying curve, d-axis ultra-transient reactance time-varying curve, and q-axis ultra-transient reactance time-varying curve. These time-varying characteristic curves are used to characterize the variation of synchronous motor reactance parameters over time caused by changes in magnetic saturation state during dynamic processes involving sudden changes in terminal voltage.
[0077] like Figures 2 to 7 As shown, a synchronous motor time-varying reactance parameter identification system based on the superconducting loop principle adopts the above-mentioned synchronous motor time-varying reactance parameter identification method based on the superconducting loop principle, including a finite element simulation module, a permeability freezing module, a superconducting loop parameter calculation module, a surrogate model training module, and a time-varying parameter identification module that are sequentially and interactively linked.
[0078] The finite element simulation module is used to establish a transient electromagnetic finite element solution model of the synchronous motor based on the electromagnetic structure parameters of the synchronous motor. It performs time-domain finite element simulation under the condition of sudden change in terminal voltage and obtains the two-dimensional cross-sectional magnetic density cloud map of the motor at different time points during the dynamic process.
[0079] The permeability freezing module is used to select multiple sample time points from the dynamic process, obtain the rotor position angle corresponding to each sample time point, and freeze the permeability of the stator and rotor in the two-dimensional plane of the motor at each sample time point.
[0080] The superconducting loop parameter calculation module is used to assign the frozen permeability at each sample time point to a new motor model at the same rotor position angle. Based on the superconducting loop principle, the rotor position, winding connection state and material electromagnetic properties are set, and DC excitation is applied to the three-phase stator windings. Steady-state current and flux linkage data are collected. The d-axis and q-axis reactance parameters corresponding to each sample time point are calculated according to the ratio of the per-unit reactance value to the per-unit current value of the corresponding axis flux linkage.
[0081] The surrogate model training module is used to construct a neural network composed of a convolutional neural network (CNN) and an extreme learning machine (ELM). The two-dimensional cross-sectional magnetic flux density map of the motor corresponding to each sample time point is used as the input dataset, and the corresponding d-axis and q-axis reactance parameters are used as the output training set to train the neural network and generate a surrogate model to characterize the correspondence between the motor magnetic flux density map and the reactance parameters.
[0082] The time-varying parameter identification module is used to input the two-dimensional cross-sectional magnetic density cloud map of the motor corresponding to other time points in the dynamic process into the proxy model, obtain the d-axis and q-axis reactance parameter values at the corresponding time points, and form the time-varying characteristic curve of the synchronous motor reactance parameter changing with the dynamic process.
[0083] The superconducting circuit parameter calculation module includes a rotor position setting unit, a winding state setting unit, a material property setting unit, a three-phase DC excitation unit, and a flux acquisition unit.
[0084] The rotor position setting unit is used to align the rotor magnetic pole axis with the A-phase axis or the rotor inter-pole axis with the A-phase axis. The winding state setting unit is used to set the excitation winding to an open-circuit or short-circuit state. The material property setting unit is used to set the electromagnetic properties of the solid rotor material, slot wedge material, or damping winding material according to the synchronous motor type. The three-phase DC excitation unit is used to apply DC power to the A, B, and C phase windings respectively, so that the three-phase currents in steady state are respectively... , , The flux linkage acquisition unit is used to acquire the flux linkage of phase A in steady state and, together with the per-unit value of the phase A current, calculates the corresponding per-unit value of the reactance parameter.
[0085] A computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the aforementioned method for identifying time-varying reactance parameters of a synchronous motor based on the superconducting circuit principle. The computer-readable storage medium can be a hard disk, solid-state drive, read-only memory, random access memory, portable storage device, or other media capable of storing a computer program. The computer program can be deployed in a motor finite element simulation workstation, a power system digital simulation platform, or a synchronous motor parameter identification calculation platform.
[0086] Example 1: Example of parameter identification for sudden three-phase short circuit at the generator terminals of a large salient-pole hydro-turbine synchronous generator under no-load conditions.
[0087] This embodiment takes a large salient-pole hydro-turbine synchronous generator as the object and performs time-varying identification of reactance parameters during the dynamic process of a sudden three-phase short circuit under no-load conditions at the generator terminals. The reactance parameters to be identified include d-axis synchronous reactance, q-axis synchronous reactance, d-axis transient reactance, d-axis ultra-transient reactance, and q-axis ultra-transient reactance.
[0088] First, based on the stator core dimensions, rotor pole structure, air gap length, stator winding connection method, excitation winding parameters, damping winding parameters, and core material magnetization curve of the large salient-pole hydro-turbine synchronous generator, a two-dimensional electromagnetic transient finite element solution model was established in finite element simulation software. The generator was set to operate under no-load rated voltage conditions, and a sudden three-phase short circuit was applied at the generator terminals at a preset time for time-domain finite element simulation. During the simulation, two-dimensional cross-sectional magnetic flux density cloud maps of the motor were output at each time point during the short-circuit dynamic process according to a preset time step. A uniform upper limit value, lower limit value, and magnetic flux density change step size were applied to all magnetic flux density cloud maps.
[0089] Secondly, 45 sample time points were selected from the short-circuit dynamic process where the rotor magnetic pole axis was aligned with the A-phase axis, and 45 sample time points were selected where the rotor inter-pole axis was aligned with the A-phase axis. Specifically, the sample time points were selected at shorter time intervals during the initial stage after the short circuit, and at longer time intervals during subsequent stages of the short circuit. For each sample time point, the corresponding rotor position angle was extracted, and the spatial distribution of stator and rotor permeability at that time point was frozen.
[0090] Then, the permeability values frozen at each sample time point were imported into the new motor model, and the rotor position, excitation winding connection state, and material electromagnetic properties were set according to the reactance type to be calculated. When calculating the d-axis synchronous reactance, the rotor pole axis was aligned with the A-phase axis, the excitation winding was open-circuited, and the damping winding material was set to air. When calculating the q-axis synchronous reactance, the rotor inter-pole axis was aligned with the A-phase axis, and the damping winding material was set to air. When calculating the d-axis transient reactance, the rotor pole axis was aligned with the A-phase axis, the excitation winding was short-circuited and its resistance was 0, and the damping winding material was set to air. When calculating the d-axis ultra-transient reactance, the rotor pole axis was aligned with the A-phase axis, the excitation winding was short-circuited and its resistance was 0, and the damping winding material resistance was set to 0. When calculating the q-axis ultra-transient reactance, the rotor inter-pole axis was aligned with the A-phase axis, and the damping winding material resistance was set to 0.
[0091] In the calculation of each reactance parameter, the motor speed is set to 0, and DC power is applied to the A, B, and C phase windings respectively, so that the three-phase currents in steady state are respectively , , After the current reaches a steady state, the A-phase flux linkage is collected. Based on the alignment relationship between the rotor pole axis or the rotor inter-pole axis and the A-phase axis, the A-phase flux linkage is equivalent to the corresponding d-axis flux linkage, q-axis flux linkage, d-axis transient flux linkage, d-axis ultra-transient flux linkage, or q-axis ultra-transient flux linkage. Then, the per-unit value of the reactance parameter corresponding to each sample time point is calculated.
[0092] Subsequently, a CNN-ELM combined neural network was constructed. The two-dimensional cross-sectional magnetic flux density maps of the motor corresponding to the 90 sample time points were used as the input dataset, and the per-unit values of the five types of reactance parameters calculated accordingly were used as the output training set to train the CNN-ELM combined neural network. The CNN extracts the spatial magnetic field distribution features from the magnetic flux density maps, and the ELM outputs the corresponding per-unit values of the reactance parameters based on these spatial magnetic field distribution features. After training, a surrogate model capable of representing the correspondence between the magnetic flux density maps and the reactance parameters is obtained.
[0093] Finally, the two-dimensional cross-sectional magnetic flux density maps of the motor at other time points during the short-circuit dynamic process are input into the trained surrogate model to obtain the per-unit values of the d-axis synchronous reactance, q-axis synchronous reactance, d-axis transient reactance, d-axis ultra-transient reactance, and q-axis ultra-transient reactance at each time point, and these are then used to form corresponding time-varying characteristic curves in chronological order. These time-varying characteristic curves reflect the influence of changes in the saturation state of the motor's internal magnetic circuit during the short-circuit dynamic process on various reactance parameters, and can be used for synchronous motor dynamic performance analysis, power system transient simulation parameter correction, and excitation control and protection setting verification.
[0094] The above embodiments are merely preferred technical solutions of the present invention and should not be considered as limitations on the present invention. The scope of protection of the present invention should be limited to the technical solutions described in the claims, including equivalent substitutions of the technical features described in the claims. That is, equivalent substitutions and improvements within this scope are also within the scope of protection of the present invention.
Claims
1. A method for identifying time-varying reactance parameters of a synchronous motor based on the principle of superconducting circuits, characterized in that, Includes the following steps: S1. Based on the electromagnetic structure parameters of the synchronous motor, establish the electromagnetic transient finite element solution model of the synchronous motor. Perform time-domain finite element simulation under the condition of sudden change in terminal voltage to obtain the two-dimensional cross-sectional magnetic density cloud map of the motor at different time points during the dynamic process. S2. Select multiple sample time points from the dynamic process, obtain the rotor position angle corresponding to each sample time point, and freeze the permeability of the stator and rotor in the two-dimensional plane of the motor at each sample time point. S3. The frozen permeability at each sample time point is assigned to a new motor model at the same rotor position angle. Based on the superconducting circuit principle, the rotor position, winding connection state and material electromagnetic properties are set. DC excitation is applied to the three-phase stator windings, steady-state current and flux linkage data are collected, and the d-axis and q-axis reactance parameters corresponding to each sample time point are calculated. S4. Construct a neural network composed of a convolutional neural network (CNN) and an extreme learning machine (ELM). Use the two-dimensional cross-sectional magnetic flux density map of the motor corresponding to each sample time point as the input dataset and the corresponding d-axis and q-axis reactance parameters as the output training set to train the neural network and generate a surrogate model to characterize the correspondence between the motor magnetic flux density map and the reactance parameters. S5. Input the two-dimensional cross-sectional magnetic density cloud map of the motor corresponding to other time points in the dynamic process into the proxy model to obtain the d-axis and q-axis reactance parameter values at the corresponding time points, and form the time-varying characteristic curve of the synchronous motor reactance parameter changing with the dynamic process.
2. The method for identifying time-varying reactance parameters of a synchronous motor based on the principle of a superconducting circuit according to claim 1, characterized in that, In step S1, when obtaining the magnetic flux density cloud map of the two-dimensional cross section of the motor at different time points, the magnetic flux density cloud map of each motor adopts the same upper limit value, lower limit value and magnetic flux density change step size to ensure that the image features of the input dataset have a uniform scale.
3. The method for identifying time-varying reactance parameters of a synchronous motor based on the principle of a superconducting circuit as described in claim 1, characterized in that, In step S2, the multiple sample time points include sample time points where the rotor magnetic pole axis is aligned with the A-phase axis, and sample time points where the rotor inter-pole axis is aligned with the A-phase axis; wherein, the sample time point selection density at the beginning of the transient process is greater than the sample time point selection density in the subsequent stages of the transient process.
4. The method for identifying time-varying reactance parameters of a synchronous motor based on the principle of a superconducting circuit according to claim 3, characterized in that, The number of sample time points for rotor magnetic pole axis alignment with phase A axis is 30 to 60, and the number of sample time points for rotor inter-pole axis alignment with phase A axis is 30 to 60.
5. The method for identifying time-varying reactance parameters of a synchronous motor based on the principle of a superconducting circuit according to claim 1, characterized in that, In step S3, the d-axis and q-axis reactance parameters include at least one of the following: d-axis synchronous reactance, q-axis synchronous reactance, d-axis transient reactance, d-axis ultra-transient reactance, and q-axis ultra-transient reactance.
6. The method for identifying time-varying reactance parameters of a synchronous motor based on the principle of a superconducting circuit according to claim 5, characterized in that, When calculating the d-axis synchronous reactance, import the stator and rotor permeability corresponding to when the rotor pole axes are aligned with the A-phase axis. Set the motor speed to 0 and the excitation winding to an open circuit. If the synchronous motor is a salient-pole synchronous motor, set the solid rotor material to be magnetically permeable but non-conductive and the slot wedge material to air. If the synchronous motor is a salient-pole synchronous motor, set the damping winding material to air. Apply DC power to the A, B, and C phase windings respectively, so that the three-phase currents in steady state are respectively... , , At this point, the A-phase current is equivalent to the d-axis current, the A-phase flux linkage is equivalent to the d-axis flux linkage, and the per-unit value of the d-axis synchronous reactance is calculated according to the following formula: ; In the formula, This is the per-unit value of the d-axis synchronous reactance. This represents the per-unit value corresponding to the nominal value of the d-axis flux linkage when the armature current is stable. The armature current has a nominal value The corresponding per-unit value.
7. The method for identifying time-varying reactance parameters of a synchronous motor based on the principle of a superconducting circuit according to claim 5, characterized in that, When calculating the q-axis synchronous reactance, import the stator and rotor permeability corresponding to the rotor pole axes aligned with the A-phase axis, and set the motor speed to 0. If the synchronous motor is a salient-pole synchronous motor, set the solid rotor material to be magnetically permeable but non-conductive, and the slot wedge material to be air. If the synchronous motor is a salient-pole synchronous motor, set the damping winding material to air. Apply DC power to the A, B, and C phase windings respectively, so that the three-phase currents in steady state are respectively... , , At this point, the A-phase current is equivalent to the q-axis current, and the A-phase flux linkage is equivalent to the q-axis flux linkage. The per-unit value of the q-axis synchronous reactance is calculated according to the following formula: ; In the formula, This is the per-unit value of the q-axis synchronous reactance. This is the per-unit value corresponding to the nominal value of the q-axis flux linkage when the armature current is stable. The armature current has a nominal value The corresponding per-unit value.
8. The method for identifying time-varying reactance parameters of a synchronous motor based on the principle of a superconducting circuit according to claim 5, characterized in that, When calculating the d-axis transient reactance, import the stator and rotor permeability corresponding to when the rotor pole axes are aligned with the A-phase axis. Set the motor speed to 0, the excitation winding to a short-circuited state, and the excitation winding resistance to 0. If the synchronous motor is a salient-pole synchronous motor, set the solid rotor material to be magnetically permeable but non-conductive, and the slot wedge material to be air. If the synchronous motor is a salient-pole synchronous motor, set the damping winding material to air. Apply DC power to the A, B, and C phase windings respectively, so that the three-phase currents in steady state are respectively... , , At this point, the A-phase current is equivalent to the d-axis current, and the A-phase flux linkage is equivalent to the d-axis transient flux linkage. The per-unit value of the d-axis transient reactance is calculated according to the following formula: ; In the formula, This represents the per-unit value of the d-axis transient reactance. This represents the per-unit value corresponding to the nominal value of the transient flux linkage along the d-axis when the armature current is stable. The armature current has a nominal value The corresponding per-unit value.
9. The method for identifying time-varying reactance parameters of a synchronous motor based on the principle of a superconducting circuit according to claim 5, characterized in that, When calculating the d-axis transient reactance, import the stator and rotor permeability corresponding to the rotor magnetic pole axis being aligned with the A-phase axis, set the motor speed to 0, set the excitation winding to a short-circuited state, and set the excitation winding resistance to 0. If the synchronous motor is a salient pole synchronous motor, then set the resistance of the solid rotor material to 0 and the resistance of the slot wedge material to 0. If the synchronous motor is a salient-pole synchronous motor, then set the resistance of the damping winding material to 0; apply a DC power supply to the A, B, and C phase windings respectively, so that the three-phase currents in steady state are respectively , , At this point, the A-phase current is equivalent to the d-axis current, and the A-phase flux linkage is equivalent to the d-axis ultra-transient flux linkage. The per-unit value of the d-axis ultra-transient reactance is calculated according to the following formula: ; In the formula, This represents the per-unit value of the d-axis ultratransient reactance. This represents the per-unit value corresponding to the named value of the d-axis transient flux linkage when the armature current is stable. The armature current has a nominal value The corresponding per-unit value.
10. The method for identifying time-varying reactance parameters of a synchronous motor based on the principle of a superconducting circuit according to claim 5, characterized in that, When calculating the q-axis transient reactance, import the stator and rotor permeability corresponding to the rotor pole axis being aligned with the A-phase axis, and set the motor speed to 0. If the synchronous motor is a salient pole synchronous motor, then set the resistance of the solid rotor material to 0 and the resistance of the slot wedge material to 0. If the synchronous motor is a salient-pole synchronous motor, then set the resistance of the damping winding material to 0; apply a DC power supply to the A, B, and C phase windings respectively, so that the three-phase currents in steady state are respectively , , At this point, the A-phase current is equivalent to the q-axis current, and the A-phase flux linkage is equivalent to the q-axis transient flux linkage. The per-unit value of the q-axis transient reactance is calculated according to the following formula: ; In the formula, This represents the per-unit value of the q-axis transient reactance. This represents the per-unit value corresponding to the nominal value of the q-axis transient flux linkage when the armature current is stable. The armature current has a nominal value The corresponding per-unit value.
11. The method for identifying time-varying reactance parameters of a synchronous motor based on the principle of a superconducting circuit, as described in any one of claims 6 to 10, is characterized in that... The three-phase current formed by applying a DC power supply to the A, B, and C three-phase windings. , , Used to form an equivalent magnetic field direction corresponding to the A-phase axis on the stator side; when the rotor pole axis is aligned with the A-phase axis, the A-phase current and A-phase flux linkage are used as d-axis current and d-axis flux linkage in reactance calculation, respectively; when the rotor pole axis is aligned with the A-phase axis, the A-phase current and A-phase flux linkage are used as q-axis current and q-axis flux linkage in reactance calculation, respectively.
12. The method for identifying time-varying reactance parameters of a synchronous motor based on the principle of a superconducting circuit according to claim 1, characterized in that, In step S4, the convolutional neural network (CNN) is used to extract the spatial magnetic field distribution features in the magnetic density cloud map of the two-dimensional cross-section of the motor, and the extreme learning machine (ELM) is used to output the corresponding d-axis and q-axis reactance parameters based on the spatial magnetic field distribution features.
13. The method for identifying time-varying reactance parameters of a synchronous motor based on the principle of a superconducting circuit according to claim 1, characterized in that, In step S5, the time-varying characteristic curves of the synchronous motor reactance parameters as a function of the dynamic process include at least one of the following: d-axis synchronous reactance time-varying curve, q-axis synchronous reactance time-varying curve, d-axis transient reactance time-varying curve, d-axis ultra-transient reactance time-varying curve, and q-axis ultra-transient reactance time-varying curve.
14. A system for identifying time-varying reactance parameters of a synchronous motor based on the principle of superconducting circuits, characterized in that, The method for identifying time-varying reactance parameters of a synchronous motor based on the superconducting loop principle, as described in any one of claims 1-13, includes a finite element simulation module, a permeability freezing module, a superconducting loop parameter calculation module, a surrogate model training module, and a time-varying parameter identification module that are sequentially and interactively linked: wherein: The finite element simulation module is used to establish a transient electromagnetic finite element solution model of the synchronous motor based on the electromagnetic structure parameters of the synchronous motor. It performs time-domain finite element simulation under the condition of sudden change in terminal voltage and obtains the two-dimensional cross-sectional magnetic density cloud map of the motor at different time points during the dynamic process. The permeability freezing module is used to select multiple sample time points from the dynamic process, obtain the rotor position angle corresponding to each sample time point, and freeze the permeability of the stator and rotor in the two-dimensional plane of the motor at each sample time point. The superconducting loop parameter calculation module is used to assign the frozen permeability at each sample time point to a new motor model at the same rotor position angle. Based on the superconducting loop principle, the rotor position, winding connection state and material electromagnetic properties are set, and DC excitation is applied to the three-phase stator windings. Steady-state current and flux linkage data are collected. The d-axis and q-axis reactance parameters corresponding to each sample time point are calculated according to the ratio of the per-unit reactance value to the per-unit current value of the corresponding axis flux linkage. The surrogate model training module is used to construct a neural network composed of a convolutional neural network (CNN) and an extreme learning machine (ELM). The two-dimensional cross-sectional magnetic flux density map of the motor corresponding to each sample time point is used as the input dataset, and the corresponding d-axis and q-axis reactance parameters are used as the output training set to train the neural network and generate a surrogate model to characterize the correspondence between the motor magnetic flux density map and the reactance parameters. The time-varying parameter identification module is used to input the two-dimensional cross-sectional magnetic density cloud map of the motor corresponding to other time points in the dynamic process into the proxy model, obtain the d-axis and q-axis reactance parameter values at the corresponding time points, and form the time-varying characteristic curve of the synchronous motor reactance parameter changing with the dynamic process.
15. The synchronous motor time-varying reactance parameter identification system based on the superconducting circuit principle according to claim 14, characterized in that, The superconducting loop parameter calculation module includes a rotor position setting unit, a winding state setting unit, a material property setting unit, a three-phase DC excitation unit, and a flux linkage acquisition unit; wherein: The rotor position setting unit is used to set the rotor magnetic pole axis to align with the A-phase axis or the rotor inter-pole axis to align with the A-phase axis. The winding state setting unit is used to set the excitation winding to an open circuit state or a short circuit state. The material property setting unit is used to set the electromagnetic properties of solid rotor material, slot wedge material, or damping winding material according to the type of synchronous motor. The three-phase DC excitation unit is used to apply DC power to the A, B, and C phase windings respectively, so that the three-phase currents in steady state are respectively , , ; The flux linkage acquisition unit is used to acquire the flux linkage of phase A in steady state and calculate the corresponding per-unit value of reactance parameter in conjunction with the per-unit value of phase A current.
16. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the method for identifying time-varying reactance parameters of a synchronous motor based on the principle of a superconducting circuit, as described in any one of claims 1 to 13.
