A closed-loop adaptive simulation system for dynamic evolution of inter-turn short circuits in motors

Abstract

This invention discloses a closed-loop adaptive dynamic evolution simulation system for inter-turn short circuits in motors, relating to the field of motor fault simulation technology. It includes a feature extraction module for calculating and generating a fault excitation index and winding chaotic entropy from acquired motor operating data; a fault evolution modeling module for solving a preset fault evolution model based on the fault excitation index and winding chaotic entropy to generate an updated fault resistance; a multi-level parameter dynamic correction module for correcting the parameters of the simulation model according to the winding chaotic entropy; and a simulation strategy optimization module for switching between multiple preset simulation modes based on the winding chaotic entropy. This invention, by constructing a closed-loop adaptive system and integrating interdisciplinary theories, achieves dynamic simulation of the entire lifecycle of inter-turn short circuits in motors. Through multi-level correction and adaptive strategies, it balances high fidelity, physical interpretability, and computational efficiency.

Description

Technical Field

[0001] This invention relates to the field of motor fault simulation technology, specifically to a closed-loop adaptive dynamic evolution simulation system for inter-turn short circuits in motors. Background Technology

[0002] Electric motors are indispensable core power equipment in modern industrial production. Their operational stability and reliability directly affect the normal operation of the entire production system. Among the many types of motor failures, stator winding inter-turn short circuits are of particular concern due to their high incidence, strong concealment, and rapid destructiveness. This type of fault usually begins with minor damage to the winding insulation but can rapidly deteriorate within a very short time. Its development process exhibits highly dynamic and nonlinear characteristics. If its development trend is not detected in time and accurately assessed, it may lead to insulation breakdown, phase-to-phase short circuits, and ultimately motor burnout. This not only causes huge direct economic losses but may also trigger a chain of production safety accidents, posing a serious threat to industrial safety. Therefore, accurate modeling and simulation of stator winding inter-turn short circuits are crucial for ensuring industrial production safety, enabling early warning and predictive maintenance.

[0003] Existing simulation techniques for inter-turn short circuits in motors generally rely on electromagnetic field analysis models based on the finite element method or magnetic circuit method. These traditional models typically treat such faults as a static problem with a fixed number of short-circuit turns and a constant fault resistance. However, this approach has a fundamental flaw: it drastically simplifies a dynamic, nonlinear physical degradation process into a static, linear problem. In the real physical world, the fault resistance of an inter-turn short circuit is not constant; it continuously and dynamically changes due to the combined effects of various factors, such as localized overheating caused by the short-circuit circulating current, electrodynamic impacts, and aging of insulation materials. Because of the fixed core parameters, traditional simulation models cannot realistically simulate the complete lifecycle of a fault from its inception to its final deterioration. This results in significant limitations in the accuracy of fault diagnosis and the guidance for predictive maintenance, making it difficult to meet the high fidelity and predictive capabilities required by modern industry for high-reliability digital twin technology for motors. Summary of the Invention

[0004] The purpose of this invention is to provide a closed-loop adaptive dynamic evolution simulation system for inter-turn short circuits in motors, which solves the problems existing in the background technology.

[0005] To address the aforementioned technical problems, this invention provides a closed-loop adaptive dynamic evolution simulation system for inter-turn short circuits in motors, comprising: a feature extraction module, used to calculate and generate a fault excitation index and winding chaotic entropy from the acquired motor operating data;

[0006] The fault evolution modeling module is used to solve the preset fault evolution model based on the fault excitation index and the winding chaotic entropy to generate the updated fault resistance.

[0007] A multi-level parameter dynamic correction module is used to correct the parameters of the simulation model based on the winding chaotic entropy.

[0008] The simulation strategy optimization module is used to switch between multiple preset simulation modes based on the winding chaotic entropy.

[0009] Preferably, the process by which the feature extraction module calculates and generates the fault excitation index includes:

[0010] The motor speed and load torque are obtained from the motor operating data, and the operating condition adjustment function is determined accordingly.

[0011] Obtain the number of short-circuited turns from the motor operating data;

[0012] Obtain the root mean square value of the fault phase current from the motor operating data;

[0013] Obtain the current fault resistance value from the motor operating data;

[0014] The fault excitation index is calculated by combining the operating condition adjustment function, the number of short-circuit turns, the root mean square value of the fault phase current, and the current fault resistance value.

[0015] Preferably, the process by which the feature extraction module calculates and generates the winding chaotic entropy includes:

[0016] A fast Fourier transform is performed on the stator current signal obtained from the motor operation data to obtain multiple harmonic components;

[0017] Based on multiple harmonic components, calculate the normalized proportion of the energy of each harmonic component to the total harmonic energy;

[0018] Based on each normalized weight, the winding chaotic entropy is calculated using the information entropy calculation method.

[0019] Preferably, the process by which the fault evolution modeling module generates the updated fault resistance includes:

[0020] Obtain the number of short-circuit turns, the root mean square value of the fault phase current, the current fault resistance value, the preset rated current, the preset single-turn winding health resistance, and the operating condition adjustment function.

[0021] Based on the parameters and functions obtained above, the dimensionless fault stress factor is calculated and generated.

[0022] By combining the fault stress factor and the winding chaotic entropy, a preset fault evolution model is solved to generate an updated fault resistance.

[0023] Preferably, it is also used for:

[0024] The updated fault resistance is used as the current fault resistance value in the next simulation time step and provided to the feature extraction module to form a closed-loop feedback.

[0025] Preferably, the process by which the multi-level parameter dynamic correction module corrects the parameters of the simulation model includes:

[0026] The winding chaotic entropy is compared with the preset entropy increase trigger threshold;

[0027] Based on the comparison results, adaptive weights are generated by calculating using a preset logical function;

[0028] Adaptive weights are applied to the three-level progressive architecture of "turn-phase-system" to correct the parameters of the simulation model.

[0029] Preferably, the output characteristics of the logic function are as follows: when the winding chaotic entropy is less than the entropy increase trigger threshold, the value of the adaptive weight is close to 0; when the winding chaotic entropy is equal to or greater than the entropy increase trigger threshold, the value of the adaptive weight is close to 1, so as to activate parameter correction.

[0030] Preferably, the process of switching simulation modes by the simulation strategy optimization module includes:

[0031] The winding chaotic entropy is compared with the preset first threshold and second threshold;

[0032] The second threshold is greater than the first threshold;

[0033] Based on the comparison results, the system switches between early warning mode, fine diagnosis mode, and failure prediction mode.

[0034] Preferably, the mode switching specifically involves:

[0035] When the winding chaotic entropy is less than the first threshold, switch to early warning mode and use the lumped parameter model for simulation;

[0036] When the winding chaotic entropy is greater than or equal to the first threshold and less than the second threshold, switch to fine diagnosis mode, use field-circuit coupling model for simulation and activate meso-level parameter correction.

[0037] When the winding chaotic entropy is greater than or equal to the second threshold, switch to failure prediction mode, start multiphysics coupling simulation and activate macroscopic layer parameter correction.

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

[0039] 1. This invention constructs a closed-loop adaptive system that includes feature extraction, fault evolution, parameter correction, and strategy optimization. It can continuously and with high fidelity dynamically reproduce the entire life cycle of inter-turn short-circuit faults from initiation to development and eventual failure. It introduces a fault excitation index that comprehensively quantifies electromagnetic thermal shock intensity and a winding chaotic entropy that characterizes the disorder of the system. This provides a solid physical quantitative foundation for accurately perceiving the instantaneous state and cumulative impact of faults, enabling the simulation model to have unprecedented dynamic insight capabilities.

[0040] 2. This invention deeply integrates interdisciplinary theories, creatively mathematizing the entropy increase theory of physics and the broken windows theory of sociology into calculable winding chaotic entropy and evolutionary equations that can drive simulation. This allows the description of the nonlinear deterioration process of damage acceleration to be directly coupled with the real-time electromagnetic and thermal stress and chaotic state of the system. By feeding back the updated fault resistance to the input to form a closed loop, fault evolution is no longer a pre-set script, but a dynamic process that interacts with the system state in real time and is self-driven, greatly improving the physical realism and predictive ability of the simulation.

[0041] 3. This invention constructs a unique three-level progressive parameter correction architecture of "turn-phase-system", and drives an adaptive weight by winding chaotic entropy to intelligently activate and adjust the magnitude of parameter correction. When the fault is not significant, the correction mechanism remains silent to avoid unnecessary computational overhead; when the fault worsens and the system entropy value exceeds the preset threshold, the correction function is activated and enhanced with the severity of the fault. This not only ensures that the simulation model can maintain high fidelity throughout the entire fault life cycle, but also cleverly optimizes the balance between simulation accuracy and computational cost.

[0042] 4. This invention proposes an adaptive simulation optimization strategy based on entropy threshold, which can intelligently switch between the most computationally efficient lumped parameter model, the high-precision field-circuit coupling model, and the most comprehensive multiphysics coupling model according to the actual severity of the fault. Compared with the limitation of a single model that cannot meet the needs of all scenarios, it can dynamically match the most suitable simulation resource configuration for different fault stages in the entire monitoring and prediction task, which greatly enhances the engineering practical value and economy of the system. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1This is a structural diagram of a closed-loop adaptive inter-turn short-circuit dynamic evolution simulation system for motors according to the present invention. Detailed Implementation

[0045] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0046] Example 1:

[0047] Please see Figure 1 The present invention provides a closed-loop adaptive dynamic evolution simulation system for inter-turn short circuits in motors, comprising: a feature extraction module, used to calculate and generate a fault excitation index and winding chaotic entropy from the acquired motor operating data;

[0048] The fault evolution modeling module is used to solve the preset fault evolution model based on the fault excitation index and the winding chaotic entropy to generate the updated fault resistance.

[0049] A multi-level parameter dynamic correction module is used to correct the parameters of the simulation model based on the winding chaotic entropy.

[0050] The simulation strategy optimization module switches between multiple preset simulation modes based on the winding chaotic entropy. This system aims to overcome the shortcomings of existing technologies, such as fixed simulation model parameters and the inability to simulate the dynamic evolution of faults. By constructing a closed-loop, adaptive simulation system, the system accurately depicts the complete lifecycle of inter-turn short-circuit faults from initiation to development and deterioration. The system includes a feature extraction module, a fault evolution modeling module, a multi-level parameter dynamic correction module, and a simulation strategy optimization module.

[0051] The feature extraction module aims to calculate core indicators that can quantify the fault state from motor operation data. This module is configured to calculate and generate the fault excitation index and winding chaos entropy from the acquired real-time or historical motor operation data. These two indicators jointly characterize the instantaneous state and cumulative impact of the fault from two dimensions: electromagnetic thermal shock intensity and system spectrum disorder, and serve as the core input for subsequent modules to perform dynamic modeling and adaptive control.

[0052] The fault evolution modeling module aims to construct and solve a mathematical model that can describe the dynamic changes of the core physical parameters of the fault. This module is configured to solve a pre-defined fault evolution model driven by the broken window effect based on the fault excitation index and winding chaotic entropy output by the aforementioned feature extraction module. The solution output of this model is an updated fault resistance value, which reflects the degradation result of the short-circuit point insulation state in the next time step under the combined action of the current electromagnetic thermal stress and the chaotic state of the system, thereby tightly coupling the fault evolution process with the real-time operating state of the system.

[0053] The purpose of the multi-level parameter dynamic correction module is to ensure that the simulation model maintains high fidelity at different stages of fault development. This module is configured to correct the electrical, magnetic field and thermal parameters in the simulation model in real time based on the winding chaotic entropy calculated by the feature extraction module, in a three-level progressive architecture of "turn-phase-system". By introducing an adaptive weight driven by the winding chaotic entropy, this module can intelligently control the triggering and magnitude of parameter correction, mapping the dynamic changes of microscopic physical parameters to macroscopic performance characterization.

[0054] The simulation strategy optimization module aims to balance the accuracy and efficiency of simulation to meet the analysis needs of different fault stages. This module is configured to adaptively switch between multiple preset simulation modes based on the magnitude of the winding chaos entropy. These modes correspond to different fault severity levels and are matched with simulation models of different complexity and computational resource consumption, thereby providing technical support for the realization of predictive maintenance and high-fidelity digital twins for high-reliability motors.

[0055] This embodiment achieves a paradigm shift from static parameter setting to dynamic evolution through tight data flow coupling and closed-loop feedback of the aforementioned modules. It deeply integrates interdisciplinary theories, mathematically transforming concepts such as entropy increase theory in physics and the broken window effect in sociology into calculable indicators and evolutionary equations that drive simulation. This not only enables dynamic simulation of the entire lifecycle of inter-turn short-circuit faults but also balances multi-scale fidelity, simulation accuracy, and computational efficiency through multi-level correction and adaptive strategy optimization, thereby enhancing the physical interpretability and engineering practical value of the model.

[0056] Example 2:

[0057] The process of calculating and generating the fault excitation index by the feature extraction module includes:

[0058] The motor speed and load torque are obtained from the motor operating data, and the operating condition adjustment function is determined accordingly.

[0059] Obtain the number of short-circuited turns from the motor operating data;

[0060] Obtain the root mean square value of the fault phase current from the motor operating data;

[0061] Obtain the current fault resistance value from the motor operating data;

[0062] The fault excitation index is calculated by combining the operating condition adjustment function, the number of short-circuit turns, the root mean square value of the fault phase current, and the current fault resistance value.

[0063] This embodiment provides a detailed explanation of the process by which the feature extraction module calculates and generates the fault excitation index. The fault excitation index is a composite index used to quantify the comprehensive electromagnetic and thermal shock intensity caused by a short-circuit fault to the system under specific operating conditions. Its role is to provide a quantitative physical quantity that drives the evolution of the fault. It is derived from multiple real-time operating parameters.

[0064] To calculate this index, the feature extraction module obtains the motor speed and load torque from the motor operating data and determines a preset operating condition adjustment function accordingly. This function is a dimensionless function, which is determined by fitting and calibrating aging test data of a specific motor model under different operating conditions to simulate the aggravating effect of different speed and load combinations on the insulation aging rate. The module also obtains the number of short-circuit turns from the motor operating data, which characterizes the spatial extent of the fault; obtains the root mean square value of the fault phase current, which characterizes the current electrical stress level; and obtains the current fault resistance value, which characterizes the instantaneous severity of the fault.

[0065] Based on the above parameters, the module combines the operating condition adjustment function, the number of short-circuit turns, the root mean square value of the fault phase current, and the current fault resistance value to calculate and generate the fault excitation index using the following preset formula:

[0066]

[0067] in, The fault excitation index is expressed in A / Ω.

[0068] The number of short-circuit turns is dimensionless.

[0069] The root mean square value of the fault phase current obtained from the operating data, in A;

[0070] The current fault resistance value, in Ω;

[0071] A tiny positive constant used to prevent the denominator from being zero; unit: Ω.

[0072] This is a dimensionless operating condition adjustment function, whose input is the motor speed obtained from the operating data. and load torque ;

[0073] To enable those skilled in the art to implement this function, the operating condition adjustment function can take a specific functional form, such as a quadratic polynomial function:

[0074]

[0075] in, This refers to the motor speed;

[0076] This is the load torque;

[0077] , , , , These are the fitting coefficients with specific dimensions;

[0078] The fitting coefficients are dimensionless.

[0079] These coefficients were determined for a specific type of motor by least-squares fitting of its insulation accelerated aging test data under multiple stable speed and load torque combinations; this method can effectively characterize the coupled effect of speed and load on insulation aging rate.

[0080] The index is recalculated at each simulation time step and passed to the fault evolution modeling module;

[0081] This embodiment constructs a composite index that comprehensively reflects current load, fault severity, fault extent, and motor operating conditions, making the fault evolution rate directly related to the actual pressure borne by the motor. For example, when the motor starts under heavy load, the index will increase significantly, thereby accelerating the deterioration of the fault in the model. This is highly consistent with physical reality and greatly improves the realism of fault dynamic evolution simulation.

[0082] Example 3:

[0083] The process of calculating and generating the winding chaotic entropy by the feature extraction module includes:

[0084] A fast Fourier transform is performed on the stator current signal obtained from the motor operation data to obtain multiple harmonic components;

[0085] Based on multiple harmonic components, calculate the normalized proportion of the energy of each harmonic component to the total harmonic energy;

[0086] Based on each normalized weight, the winding chaotic entropy is calculated using the information entropy calculation method.

[0087] This embodiment provides a detailed explanation of the process by which the feature extraction module calculates and generates the winding chaotic entropy. The winding chaotic entropy refers to an objective indicator used to quantify the degree of spectral disorder caused by a fault at the overall system level. Its function is to quantify the physical process of the fault leading to an increase in the disorder of the system into an objective and calculable decision indicator. Its source is the spectral analysis results of the stator current signal.

[0088] To calculate this entropy value, the module performs a fast Fourier transform on the stator current signal obtained from the motor operation data to obtain a series of harmonic components and their amplitudes in the frequency domain.

[0089] Based on multiple harmonic components, the module calculates the normalized proportion of each harmonic component's energy to the total harmonic energy; specifically, if the first harmonic component... The amplitude of the subharmonic is Then its corresponding normalized weight Calculated as:

[0090]

[0091] in, The upper limit of the harmonic analysis is determined by covering all major harmonic frequency ranges caused by inter-turn short circuits. This value is predetermined by analyzing the fault test spectrum of the sample motor.

[0092] Based on each normalized weight, the module calculates and generates the winding chaotic entropy using the information entropy calculation method:

[0093]

[0094] in, The dimensionless entropy of the winding chaos; The first step calculated Normalized proportion of subharmonic energy;

[0095] This entropy value, as a macroscopic, system-level state indicator, is simultaneously transmitted to the fault evolution modeling module, the multi-level parameter dynamic correction module, and the simulation strategy optimization module.

[0096] This embodiment transforms the physical nature of the fault into a robust quantitative indicator. Compared with traditional methods that rely on a single physical quantity, the winding chaotic entropy can more comprehensively and stably reflect the overall health status of the system, providing a global and reliable decision-making basis for the adaptive correction of subsequent modules and the intelligent switching of simulation strategies.

[0097] Example 4:

[0098] The process by which the fault evolution modeling module generates the updated fault resistance includes:

[0099] Obtain the number of short-circuit turns, the root mean square value of the fault phase current, the current fault resistance value, the preset rated current, the preset single-turn winding health resistance, and the operating condition adjustment function.

[0100] Based on the parameters and functions obtained above, the dimensionless fault stress factor is calculated and generated.

[0101] By combining the fault stress factor and the winding chaotic entropy, the preset fault evolution model is solved to generate the updated fault resistance.

[0102] The updated fault resistance is used as the current fault resistance value in the next simulation time step and provided to the feature extraction module to form a closed-loop feedback.

[0103] This embodiment provides a detailed explanation of the process by which the fault evolution modeling module generates the updated fault resistance, and the closed-loop feedback formed with the feature extraction module.

[0104] The workflow of the fault evolution modeling module begins with acquiring a series of parameters, including: the number of short-circuit turns, the root mean square value of the fault phase current, the current fault resistance value provided by the feature extraction module at the current time step, and multiple preset values, including the rated current and single-turn winding health resistance determined according to the motor model, and the aforementioned operating condition adjustment function.

[0105] Based on the parameters and functions obtained above, the module calculates and generates a dimensionless fault stress factor. This factor aims to normalize the overall electrical stress level experienced by the current fault point, thereby enhancing the model's universality and clarity of physical meaning. Its calculation formula is as follows:

[0106]

[0107] in, The fault stress factor is a dimensionless factor. The preset rated current (unit: A); The preset health resistance of a single-turn winding at the reference temperature (unit: Ω); , The definitions and origins of the remaining symbols are the same as before;

[0108] By introducing a benchmark value and Normalization is performed so that It can more universally reflect the relative severity of the fault;

[0109] In the calculation process of this module, the calculated fault stress factor will be... With the winding chaotic entropy from the feature extraction module Using these as inputs, a pre-defined fault evolution model is solved to generate an updated fault resistance. This model is a fault resistance evolution differential equation driven by the broken window effect, used to describe the nonlinear deterioration process of damage accelerating damage. The equation is constructed as follows:

[0110]

[0111] in, Fault resistor Over time The rate of change, in Ω / s;

[0112] The basic characteristic degradation rate, in Ω / s, represents the basic degradation rate under reference stress without considering the system entropy effect. It is calibrated by conducting accelerated aging tests on the motor winding insulation material.

[0113] It is a dimensionless entropy amplification factor, which is also calibrated through experimental data;

[0114] This is a random noise term, measured in Ω / s, used to characterize random impacts; by solving this differential equation, the fault resistance value for the next time step can be obtained. ;

[0115] To enable those skilled in the art to implement this, the random noise term It can be set to a zero-mean Gaussian white noise with variance This reflects the average intensity of the random shock, and the variance is also calibrated by statistical analysis of the random fluctuation components of resistance variation in the experimental data. For solving this differential equation, discretized numerical methods can be used, such as the first-order forward Euler method, at each simulation time step. Internally update the faulty resistor:

[0116]

[0117] The fault resistor at the current time step;

[0118] The fault resistor is updated at the next time step;

[0119] The simulation time step should be small enough to ensure the stability and accuracy of the numerical calculation.

[0120] The system thus achieves crucial closed-loop feedback: the updated fault resistance is used as the current fault resistance value in the next simulation time step and provided to the feature extraction module; this design forms a core feedback loop, ensuring that the updated fault resistance... Changing the electrical characteristics of the motor, thereby affecting the phase current at the next moment. With winding chaotic entropy This, in turn, alters the failure evolution rate; the closed-loop mechanism makes failure evolution a self-consistent dynamic process tightly coupled with the system state.

[0121] The synergistic effect of this embodiment is that it realizes for the first time the dynamic, closed-loop and self-consistent evolution modeling of the core parameters (fault resistance) of inter-turn short circuits; by introducing differential equations based on the broken window effect and deeply coupling their real-time solution with the system operating state, the simulation is no longer a static snapshot stitching, but a continuous and high-fidelity reproduction of the physical degradation process of the fault, which fundamentally improves the predictive ability of the simulation.

[0122] Example 5:

[0123] The process by which the multi-level parameter dynamic correction module corrects the parameters of the simulation model includes:

[0124] The winding chaotic entropy is compared with the preset entropy increase trigger threshold;

[0125] Based on the comparison results, adaptive weights are generated by calculating using a preset logical function;

[0126] Adaptive weights are applied to the three-level progressive architecture of "turn-phase-system" to correct the parameters of the simulation model;

[0127] The output characteristics of the logic function are as follows: when the winding chaotic entropy is less than the entropy increase trigger threshold, the value of the adaptive weight is close to 0; when the winding chaotic entropy is equal to or greater than the entropy increase trigger threshold, the value of the adaptive weight is close to 1 to activate parameter correction.

[0128] This embodiment provides a detailed explanation of the process by which the multi-level parameter dynamic correction module corrects the parameters of the simulation model;

[0129] The core technology of this module lies in introducing an adaptive weight based on cumulative effect to control the correction magnitude of parameters at the meso and macro levels. The correction process compares the winding chaotic entropy from the feature extraction module with a preset entropy increase trigger threshold. This threshold is a preset dimensionless value, which is set based on statistical analysis of a large number of complete experimental data of motors from health to failure, to find the entropy value that can best distinguish the fault initiation and fault development stages.

[0130] Based on the above comparison results, adaptive weights are generated through a preset logistic function. This logistic function is a standard sigmoid function, used to smoothly control the triggering and magnitude of model parameter adjustments. Its calculation formula is as follows:

[0131]

[0132] The weights are dimensionless adaptive weights in the range (0,1);

[0133] It is the chaotic entropy of the winding;

[0134] This refers to the aforementioned entropy increase trigger threshold;

[0135] It is the dimensionless slope of the function, the value of which is determined through experimental data analysis to determine the severity of the state transition;

[0136] The output characteristics of this function are explicitly defined: when the winding chaotic entropy is less than the entropy increase trigger threshold ( ), The value is close to 0; when the winding chaotic entropy is equal to or greater than the entropy increase trigger threshold ( ), The value increases rapidly and approaches 1 to activate parameter correction;

[0137] The module applies this adaptive weight to a three-level progressive architecture of "turn-phase-system" to correct the parameters of the simulation model; for example, at the meso-level, i.e., the phase winding, the correction amount for the equivalent inductance or resistance of the phase winding can be expressed as a factor related to... and All are related functions; in the early stages of a fault, When the value is close to 0, parameter corrections almost never occur; when the fault develops... Exceeding the threshold back, When activated, the magnitude of parameter correction varies. The increase is significant;

[0138] To enable those skilled in the art to implement this, the adaptive weights It can be used to correct parameters at the meso- and macro-levels; for example, for the equivalent resistance of phase windings at the meso-level. and equivalent inductance The correction is achieved through the following formula:

[0139]

[0140]

[0141] and These are the phase resistance and phase inductance under the initial healthy state, respectively;

[0142] and It is a dimensionless correction coefficient, calibrated through experimental data, used to adjust the sensitivity of chaotic entropy to the influence of parameters;

[0143] This correction is activated when the simulation strategy enters fine-tuning diagnostic mode;

[0144] For macroscopic parameters, such as the equivalent moment of inertia of the entire machine The correction can be expressed as:

[0145]

[0146] It is the initial moment of inertia;

[0147] It is a macroeconomic correction coefficient;

[0148] This correction reflects the increased mechanical vibration or imbalance that may be caused by the development of the fault, and is activated when the simulation strategy enters the failure prediction mode.

[0149] The gain effect brought by this embodiment is that it constructs an intelligent multi-scale parameter correction mechanism. Through an adaptive weighting mechanism controlled by a system-level index (chaotic entropy), it introduces necessary corrections when most needed (during the significant development stage of the fault), thereby ensuring the fidelity of the model throughout its entire life cycle while avoiding unnecessary complex calculations in the early stages of the fault, thus achieving an optimal balance between fidelity and computational cost.

[0150] Example 6:

[0151] The process of switching simulation modes in the simulation strategy optimization module includes:

[0152] The winding chaotic entropy is compared with the preset first threshold and second threshold;

[0153] The second threshold is greater than the first threshold;

[0154] Based on the comparison results, the system switches between early warning mode, fine diagnosis mode and failure prediction mode.

[0155] The specific mode switching is as follows:

[0156] When the winding chaotic entropy is less than the first threshold, switch to early warning mode and use the lumped parameter model for simulation;

[0157] When the winding chaotic entropy is greater than or equal to the first threshold and less than the second threshold, switch to fine diagnosis mode, use field-circuit coupling model for simulation and activate meso-level parameter correction.

[0158] When the winding chaotic entropy is greater than or equal to the second threshold, switch to failure prediction mode, start multiphysics coupling simulation and activate macroscopic layer parameter correction.

[0159] This embodiment provides a detailed description of the process by which the simulation strategy optimization module switches simulation modes;

[0160] The decision-making basis of this module is the winding chaotic entropy calculated by the feature extraction module. The module compares the winding chaotic entropy with a preset first threshold. Second threshold A comparison is performed; these two thresholds are dimensionless preset values, and their setting logic is based on statistical analysis of the motor fault database. Corresponding to the typical entropy value at which the fault can be initially detected, This corresponds to the typical entropy value when the fault has significantly affected the motor performance, and ;

[0161] Based on the above comparison results, the system automatically switches between early warning mode, fine diagnosis mode, and failure prediction mode; the switching rules for the modes are specified as follows:

[0162] When the winding chaotic entropy is less than the first threshold ( The system switches to early warning mode; this state is interpreted as the system being in a healthy or very early fault state. The system uses the most computationally efficient lumped parameter model for rapid simulation in order to achieve real-time monitoring of potential anomalies.

[0163] When the winding chaotic entropy is greater than or equal to the first threshold and less than the second threshold ( The system switches to fine diagnostic mode; this state is interpreted as the fault being confirmed to have formed and begun to develop. The system automatically uses a high-precision field-circuit coupling model for simulation and activates the parameter correction mechanism at the meso-level (phase winding level).

[0164] When the winding chaotic entropy is greater than or equal to the second threshold ( The system switches to failure prediction mode; this state is interpreted as the failure has entered a stage of severe deterioration. The system initiates multi-physics field coupled simulation with thermal field and stress field, and activates parameter correction at the macroscopic level (system level) to conduct comprehensive performance evaluation and life prediction.

[0165] The gain effect of this embodiment is that it proposes an adaptive simulation strategy based on the actual severity of the fault. This strategy can intelligently match the most suitable simulation model and resource configuration for different fault stages, avoiding the use of overly complex models in early faults or overly simplified models in severe faults. Thus, in the entire fault monitoring and prediction task, it takes into account both the accuracy and efficiency of the simulation, and greatly improves the practicality and economy of the system.

[0166] Compared with existing technologies that use fixed-parameter simulation models and cannot simulate the dynamic evolution of faults, this invention has the following advantages:

[0167] This invention provides a closed-loop adaptive dynamic evolution simulation system for inter-turn short circuits in motors, which fundamentally overcomes the static and isolated defects of existing simulation models. By constructing a closed-loop adaptive system that includes feature extraction, fault evolution modeling, multi-level parameter dynamic correction, and simulation strategy optimization, this solution achieves continuous and high-fidelity dynamic reproduction of the entire life cycle of inter-turn short circuit faults from initiation to failure.

[0168] The feature extraction module of this system can calculate and generate fault excitation index and winding chaos entropy from motor operation data. The fault excitation index comprehensively quantifies the combined impact intensity of the motor's real-time operating conditions, fault space breadth, and electrical stress level on the short-circuit point. The winding chaos entropy objectively measures the degree of system disorder caused by the fault from a global perspective of the system spectrum. The introduction of these two core indicators enables the model to accurately perceive the instantaneous state and cumulative impact of the fault for the first time, providing a solid physical quantitative foundation for subsequent dynamic evolution.

[0169] The system's fault evolution modeling module solves a preset fault evolution model based on the aforementioned fault excitation index and winding chaotic entropy to generate an updated fault resistance. This design reflects the nonlinear deterioration process of damage acceleration, directly coupled with the system's real-time electromagnetic thermal stress and chaotic state. More importantly, the updated fault resistance is fed back to the feature extraction module in the next simulation time step, forming a self-consistent core closed loop. This mechanism makes fault evolution no longer a preset script, but a dynamic process that interacts with the system state in real time and is self-driven, fundamentally improving the physical realism and predictive ability of the simulation.

[0170] The system's multi-level parameter dynamic correction module dynamically corrects the simulation model parameters based on the winding chaotic entropy within a three-level progressive architecture of "turn-phase-system". This module intelligently activates and adjusts the magnitude of parameter correction through an adaptive weight driven by the winding chaotic entropy. When the fault is insignificant, correction hardly occurs, avoiding unnecessary computational overhead. When the fault worsens and causes the system entropy value to exceed a preset threshold, correction is activated and strengthens as the entropy value increases. This design ensures that the simulation model maintains high fidelity throughout the entire lifecycle of the fault and optimizes the balance between simulation accuracy and computational cost.

[0171] The simulation strategy optimization module of this invention is also based on winding chaotic entropy, and adaptively switches between multiple modes such as early warning, fine diagnosis and failure prediction. It can intelligently match the most computationally efficient lumped parameter model, the high-precision field-circuit coupling model or the most comprehensive multi-physics coupling model according to the actual severity of the fault. Compared with the limitations of existing technologies where a single model cannot meet the needs of all scenarios, this strategy can dynamically and intelligently balance the accuracy and efficiency of simulation in the entire monitoring and prediction task, which greatly enhances the engineering practical value of the system.

[0172] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A closed-loop adaptive dynamic evolution simulation system for inter-turn short circuits in a motor, characterized in that, include: The feature extraction module is used to calculate and generate the fault excitation index and winding chaotic entropy from the acquired motor operation data. The fault evolution modeling module is used to solve the preset fault evolution model based on the fault excitation index and the winding chaotic entropy to generate the updated fault resistance. A multi-level parameter dynamic correction module is used to correct the parameters of the simulation model based on the winding chaotic entropy. The simulation strategy optimization module is used to switch between multiple preset simulation modes based on the winding chaotic entropy. The process by which the feature extraction module calculates and generates the fault excitation index includes: The motor speed and load torque are obtained from the motor operating data, and the operating condition adjustment function is determined accordingly. Obtain the number of short-circuited turns from the motor operating data; Obtain the root mean square value of the fault phase current from the motor operating data; Obtain the current fault resistance value from the motor operating data; Combining the operating condition regulation function, the number of short-circuit turns, the root mean square value of the fault phase current, and the current fault resistance value, the fault excitation index is calculated using the following preset formula: in, The fault excitation index is expressed in A / Ω. The number of short-circuit turns is dimensionless. The root mean square value of the fault phase current obtained from the operating data, in A; The current fault resistance value, in Ω; A tiny positive constant used to prevent the denominator from being zero; unit: Ω. This is a dimensionless operating condition adjustment function, whose input is the motor speed obtained from the operating data. and load torque ; in, This refers to the motor speed; This is the load torque; , , , , These are the fitting coefficients with specific dimensions; The fitting coefficients are dimensionless. These coefficients were determined for a specific model of motor by least-squares fitting of its insulation accelerated aging test data under multiple stable speed and load torque combinations. The process by which the feature extraction module calculates and generates the winding chaotic entropy includes: A fast Fourier transform is performed on the stator current signal obtained from the motor operation data to obtain multiple harmonic components; Based on multiple harmonic components, calculate the normalized proportion of the energy of each harmonic component to the total harmonic energy; Based on each normalized weight, the chaotic entropy of the generated winding is calculated using the information entropy calculation method. The process by which the fault evolution modeling module generates the updated fault resistance includes: Obtain the number of short-circuit turns, the root mean square value of the fault phase current, the current fault resistance value, the preset rated current, the preset single-turn winding health resistance, and the operating condition adjustment function. Based on the parameters and functions obtained above, the dimensionless fault stress factor is calculated and generated. The formula for calculating the fault stress factor is as follows: in, The fault stress factor is a dimensionless factor. This is the preset rated current; The preset health resistance of a single-turn winding at the reference temperature; By combining the fault stress factor and the winding chaotic entropy, the preset fault evolution model is solved to generate the updated fault resistance. The equations for the fault evolution model are constructed as follows: in, Fault resistor Over time The rate of change, in Ω / s; The rate of degradation of basic features; It is a dimensionless entropy amplification factor; It is a random noise term; It is the dimensionless entropy of the winding chaos.

2. The closed-loop adaptive dynamic evolution simulation system for inter-turn short circuits in motors according to claim 1, characterized in that, Also used for: The updated fault resistance is used as the current fault resistance value in the next simulation time step and provided to the feature extraction module to form a closed-loop feedback.

3. The closed-loop adaptive dynamic evolution simulation system for inter-turn short circuits of a motor according to claim 1, characterized in that, The process by which the multi-level parameter dynamic correction module corrects the parameters of the simulation model includes: The winding chaotic entropy is compared with the preset entropy increase trigger threshold; Based on the comparison results, adaptive weights are generated by calculating using a preset logical function; Adaptive weights are applied to the three-level progressive architecture of "turn-phase-system" to correct the parameters of the simulation model.

4. The closed-loop adaptive dynamic evolution simulation system for inter-turn short circuits of a motor according to claim 3, characterized in that, The output characteristics of the logic function are as follows: when the winding chaotic entropy is less than the entropy increase trigger threshold, the value of the adaptive weight is close to 0; when the winding chaotic entropy is equal to or greater than the entropy increase trigger threshold, the value of the adaptive weight is close to 1, so as to activate parameter correction.

5. The closed-loop adaptive dynamic evolution simulation system for inter-turn short circuits in motors according to claim 1, characterized in that, The process of switching simulation modes by the simulation strategy optimization module includes: The winding chaotic entropy is compared with the preset first threshold and second threshold; The second threshold is greater than the first threshold; Based on the comparison results, the system switches between early warning mode, fine diagnosis mode, and failure prediction mode.

6. The closed-loop adaptive dynamic evolution simulation system for inter-turn short circuits in motors according to claim 5, characterized in that, The specific mode switching is as follows: When the winding chaotic entropy is less than the first threshold, switch to early warning mode and use the lumped parameter model for simulation; When the winding chaotic entropy is greater than or equal to the first threshold and less than the second threshold, switch to fine diagnosis mode, use field-circuit coupling model for simulation and activate meso-level parameter correction. When the winding chaotic entropy is greater than or equal to the second threshold, switch to failure prediction mode, start multiphysics coupling simulation and activate macroscopic layer parameter correction.

Images