A method for evaluating the harmonic attenuation effect of radial electromagnetic force on motor load

Abstract

This invention provides a method for evaluating the harmonic attenuation effect of radial electromagnetic force on a motor, belonging to the field of electromagnetic force harmonic technology. The method includes: first, collecting sensor data on the motor's magnetic flux density, vibration, current, and torque; constructing a measurement vector; and segmenting the data according to a sliding time window. Then, establishing a coupled electromagnetic field and mechanical structure model, discretizing it using the finite element method, and calibrating the model parameters using the measurement data. Calculating the harmonic coefficient matrix of the radial electromagnetic force using the magnetic flux density data, and establishing a dynamic model of the motor rotor system. Inputting the standard harmonic coefficients and wave harmonic coefficients into the dynamic model to obtain the corresponding vibration response matrices, and defining the harmonic attenuation coefficient matrix. Finally, training a neural network model that can quickly predict the attenuation effect of standard harmonics and wave harmonics, perform quantitative analysis, and output a comprehensive evaluation result. This method solves the problem that existing evaluations often involve extensive experiments, which are time-consuming and labor-intensive.

Description

Technical Field

[0001] This invention belongs to the field of electromagnetic force harmonic technology, and specifically relates to a method for evaluating the reduction effect of radial electromagnetic force harmonics on the load of a motor. Background Technology

[0002] As a widely used power conversion device, the performance of electric motors is crucial for various industrial applications. During motor operation, the electromagnetic interaction between the rotor and stator generates periodic radial electromagnetic forces, which exhibit complex harmonic characteristics. These radial electromagnetic force harmonics can cause vibration and fatigue losses in the motor bearings, thus affecting the motor's reliability and service life. Therefore, effectively mitigating radial electromagnetic force harmonics has become one of the key issues in motor design and operation management.

[0003] Existing technologies for reducing harmonics in the radial electromagnetic force of motors mainly include electromagnetic structure optimization, current waveform control, and vibration absorbers. Electromagnetic structure optimization adjusts the magnetic field distribution of the motor by optimizing the geometric parameters of the stator and rotor, such as the cogging structure and winding distribution, thereby reducing harmonics. Current waveform control modulates the stator current to suppress harmonic components, thus reducing higher-order harmonics in the radial electromagnetic force. Vibration absorbers add damping devices to the motor structure, using the principle of dynamic vibration absorption to reduce vibration response. These technologies have achieved certain results, but they also have some limitations. For example, electromagnetic structure optimization requires complex modeling and simulation processes, and the optimization design space is limited; current waveform control has high requirements for the drive circuit and can only suppress harmonics of specific frequencies; while vibration absorbers can effectively reduce vibration response, they require additional structural design and debugging. Currently, evaluating the effectiveness of radial electromagnetic force harmonic reduction in motors often involves extensive experiments and requires engineers with extensive experience, which is time-consuming and labor-intensive. Therefore, there is an urgent need for a more systematic and intelligent method for evaluating the reduction effect of radial electromagnetic force harmonics in motors, which can comprehensively analyze the vibration characteristics during motor operation, quickly evaluate the effects of different reduction measures, and provide a basis for motor design and operation optimization. Summary of the Invention

[0004] In view of this, the present invention provides a method for evaluating the reduction effect of radial electromagnetic force harmonics of a motor load, which can solve the technical problem that the current evaluation of the reduction effect of radial electromagnetic force harmonics of a motor often requires a large number of experiments and engineers' experience, which is time-consuming and labor-intensive.

[0005] This invention is implemented as follows:

[0006] This invention provides a method for evaluating the harmonic attenuation effect of radial electromagnetic force on a motor load, comprising the following steps:

[0007] S10. Collect sensor data including magnetic flux density, vibration, current and torque, and construct a measurement vector; process the collected data in segments according to a preset sliding time window; at the same time, establish the matrix equation of the coupling model of the motor electromagnetic field and mechanical structure, and use the data in the measurement vector to calibrate and update the model parameters.

[0008] S20. Discretize the coupled model using the finite element method to obtain the electromagnetic field distribution matrix, and verify and correct it using magnetic flux density data measured by the Hall sensor; calculate the radial electromagnetic force distribution vector in the air gap of the motor based on the Maxwell stress tensor method and the magnetic flux density data measured by the Hall sensor.

[0009] S30. Perform a fast Fourier transform on the radial electromagnetic force vector to obtain the harmonic coefficient matrix, and decompose it into the standard harmonic coefficient matrix and the wave harmonic coefficient matrix; at the same time, establish the dynamic model matrix equation of the motor rotor system, and use the measurement data of the acceleration sensor to calibrate and update the relevant matrices.

[0010] S40. Input the standard harmonic coefficient matrix and the wave harmonic coefficient matrix into the dynamic model as excitation, and calculate the standard vibration response matrix and the wave vibration response matrix; verify and correct these response matrices using the measured data from the accelerometer.

[0011] S50. Define the standard harmonic attenuation coefficient matrix and the wave harmonic attenuation coefficient matrix, based on the vibration response matrix and the vibration response matrix when no attenuation measures are taken; calculate the standard evaluation index vector and the wave evaluation index vector, and use the weight matrix to reflect the importance of different harmonic orders.

[0012] S60. Establish a training dataset containing multiple sets of historical data, including the measurement vector, standard harmonic coefficient matrix, fluctuating harmonic coefficient matrix, and evaluation exponent vector for each sliding time window; train a neural network model to obtain an evaluation model for the harmonic attenuation effect of load radial electromagnetic force.

[0013] S70. For new motor load conditions, obtain the measurement vector; use the trained load radial electromagnetic force harmonic attenuation effect evaluation model to predict the standard evaluation index vector and the fluctuation evaluation index vector, so as to achieve fast and accurate harmonic attenuation effect evaluation.

[0014] S80. Based on the components of the standard evaluation index vector and the fluctuation evaluation index vector, perform a quantitative analysis of the harmonic attenuation effect of the radial electromagnetic force of the motor load; evaluate the attenuation degree of the standard harmonics and the fluctuation harmonics, and identify the main influencing factors.

[0015] S90 outputs comprehensive evaluation results, including standard harmonic attenuation effect and fluctuation harmonic attenuation effect.

[0016] Based on the above technical solution, the method for evaluating the load radial electromagnetic force harmonic attenuation effect of a motor according to the present invention can be further improved as follows:

[0017] The measurement vector is obtained through a sensor system, which includes:

[0018] A Hall sensor array is uniformly arranged on the stator core to measure the air gap magnetic flux density distribution;

[0019] An acceleration sensor is installed on the rotor bearing housing to measure rotor vibration;

[0020] Current sensors are installed at key points in the stator winding to measure the stator current.

[0021] A torque sensor is installed on the shaft to measure the load torque.

[0022] Furthermore, the sampling frequency of the sensor system is 1kHz-10kHz.

[0023] Furthermore, the sensor system preprocesses the collected data, including noise reduction and normalization.

[0024] Furthermore, the preset sliding time window length is 10 to 20 times the sampling period.

[0025] Furthermore, the method for performing a fast Fourier transform on the radial electromagnetic force vector is the radix-2 FFT algorithm.

[0026] Furthermore, the neural network model adopts a convolutional neural network structure.

[0027] Specifically, step S10 involves acquiring sensor data including magnetic flux density, vibration, current, and torque to construct a measurement vector {M}. First, a Hall sensor array is uniformly arranged on the motor stator core to measure the air gap magnetic flux density distribution. An accelerometer is installed on the rotor bearing housing to measure rotor vibration. Current sensors are installed at key points in the stator windings to measure stator current. A torque sensor is installed on the shaft to measure load torque. The sampling frequency is typically set to 1kHz-10kHz, and data from each sensor is acquired synchronously during motor operation. The acquired data is preprocessed, such as denoising and normalization, to construct the measurement vector {M}. The purpose of this step is to obtain key physical quantity data during motor operation.

[0028] Next, the collected data is segmented according to a preset sliding time window. Simultaneously, the matrix equation [A]{X}={B} of the coupled electromagnetic field and mechanical structure model of the motor is established. The initial value of the model matrix [A] is obtained through finite element analysis and subsequently updated through iterative optimization. The state vector {X} and input vector {B} are obtained through experimental measurements. The model error ∈ is initially set to zero. During the iterative optimization process, the model parameters are calibrated and updated using data from the measurement vector {M}. The purpose of this step is to establish a mathematical model of the motor's dynamic behavior and optimize the model parameters using measured data.

[0029] Step S20 involves discretizing the coupled model using the finite element method to obtain the electromagnetic field distribution matrix [E]. The specific steps are as follows: First, establish the motor geometric model; set material properties and boundary conditions; perform mesh generation; then solve Maxwell's equations. After obtaining the [E] matrix, verify and correct [E] using magnetic flux density data measured by a Hall sensor. Based on the Maxwell stress tensor method and the magnetic flux density data measured by the Hall sensor, calculate the radial electromagnetic force distribution vector {F} in the motor air gap. r The purpose of this step is to obtain the radial electromagnetic force distribution in the air gap of the motor, laying the foundation for subsequent harmonic analysis.

[0030] The specific steps of step S30 are to adjust the radial electromagnetic force vector {F}. r Perform a Fast Fourier Transform on {F} to obtain the harmonic coefficient matrix [H]. The specific steps are as follows: First, perform a Fast Fourier Transform on {F}. r The sample number N is padded with zeros to make it a power of 2; then [H] is calculated using the radix-2 FFT algorithm. [H] is then decomposed into the standard harmonic coefficient matrix [H]. std ] and the wave harmonic coefficient matrix [H fluc Simultaneously, the dynamic model matrix equations of the motor rotor system are established. The correlation matrices [M], [C], and [K] are calibrated and updated using measurement data from the accelerometer. The purpose of this step is to obtain the harmonic characteristics of the radial electromagnetic force of the motor and to establish a dynamic model of the motor rotor system.

[0031] The specific step of step S40 is to convert the standard harmonic coefficient matrix [H] std ] and the wave harmonic coefficient matrix [H fluc The excitation inputs are respectively used as excitation inputs into the dynamic model to calculate the standard vibration response matrix [R]. std ] and wave vibration response matrix [R fluc The specific calculation process is as follows: First, construct the frequency response function (FRF) matrix ([K]-ω). 2 [M]+jω[C]) -1 Then [H] std] and [H fluc Using [R] as input, calculate the output response [R] respectively. std ] and [R fluc These response matrices are verified and corrected using measured data from an accelerometer. The purpose of this step is to obtain the vibration response characteristics of the motor under standard harmonic and fluctuating harmonic excitation.

[0032] The specific steps in step S50 are to define the standard harmonic attenuation coefficient matrix [S std ] and the wave harmonic attenuation coefficient matrix [S fluc ]. [S std The formula for calculating [S] is [S] std ] = [I] - [R std [R0] -1 , where [I] is the identity matrix, and [R0] is the vibration response matrix without mitigation measures. [S fluc The formula for calculating [S] is [S] fluc ] = [I] - [R fluc [R0] -1 Next, the standard evaluation index vector {E} is calculated. std} and volatility assessment index vector {E fluc}. {E std The formula for calculating} is {E} std} = [W std ][S std ]{1}, where [W std [E] is the weight matrix. fluc The formula for calculating} is {E} fluc} = [W fluc ][S fluc ]{1}, where [W fluc [ ] represents the weight matrix. The purpose of this step is to quantify the attenuation effect of standard harmonics and wave harmonics.

[0033] The specific step of step S60 is to establish a training dataset containing multiple sets of historical data, including the measurement vector {M} for each sliding time window and the standard harmonic coefficient matrix [H]. std ], Wave harmonic coefficient matrix [H fluc Evaluation index vector {E} std} and {E fluc Then, a neural network is trained to obtain a model that can quickly evaluate the harmonic attenuation effect. The forward propagation formula of the neural network is: Where y is the output, f is the activation function, and w is the activation function. i As the weight, x i Let be the input and b be the bias vector. The loss function of the neural network is... Where m is the number of samples, yi For the true value, λ is the predicted value, and λ is the regularization coefficient. The purpose of this step is to train an intelligent model that can quickly evaluate the harmonic attenuation effect.

[0034] The specific steps of step S70 are as follows: under the new motor load condition, obtain the measurement vector {M}; then, using the trained evaluation model, predict the standard evaluation index vector {E}. std} and volatility assessment index vector {E fluc The prediction formula for the evaluation model is as follows: in Let [W] be the predicted evaluation index vector, [W] be the weight matrix, b be the bias vector, and f be the activation function. The purpose of this step is to quickly and accurately evaluate the harmonic attenuation effect under new operating conditions using the trained model.

[0035] The specific steps of step S80 are based on {E std} and {E fluc The components of} are used to quantitatively analyze the harmonic attenuation effect on the radial electromagnetic force of the motor load. The quantitative index Q for the standard harmonic attenuation effect is used. std The calculation formula is Where N represents the dimension of the evaluation index, E std,i Let Q be the i-th component of the standard evaluation index vector. The quantitative index for the wave harmonic attenuation effect is Q. fluc The calculation formula is Overall evaluation results Q total The formula for calculating Q is total =w std Q std +w fluc Q fluc , where w std and w fluc These are the weighting coefficients for standard harmonics and oscillating harmonics, respectively. The purpose of this step is to quantitatively evaluate the attenuation effect of standard harmonics and oscillating harmonics and to provide a comprehensive harmonic attenuation effect.

[0036] Furthermore, the expression for the measurement vector M is as follows:

[0037] M = [m1, m2, ..., m n ] T ;

[0038] In the formula, m i Let M represent the measurement value of the i-th sensor, and n be the total number of sensors. M is obtained experimentally, with the specific steps as follows:

[0039] Step 1: Install high-precision sensors, including Hall effect sensors, acceleration sensors, current sensors, and torque sensors;

[0040] Step 2: Set the sampling frequency, usually 1kHz-10kHz;

[0041] Step 3: During motor operation, synchronously collect data from each sensor;

[0042] Step 4: Preprocess the collected data, such as denoising and normalization.

[0043] Matrix equations of the coupled model of the electromagnetic field and mechanical structure of the electric motor:

[0044] [A]X=B+∈;

[0045] In the formula, [A] is the system matrix, X is the state vector, B is the input vector, and ∈ is the error vector. The initial value of [A] is obtained through finite element analysis and subsequently updated through iterative optimization. X and B are obtained through experimental measurements. ∈ represents the model error, initially set as a zero vector.

[0046] The update formula for the system matrix [A] is as follows:

[0047]

[0048] In the formula, k represents the number of iterations, and η is the learning rate. The value of η is usually in the range of 0.01-0.1, and the optimal value is determined by cross-validation.

[0049] Discretized expression of the electromagnetic field distribution matrix [E]:

[0050]

[0051] Where, N e [N] is the number of finite element elements, [D] is the shape function matrix, [D] is the electromagnetic material property matrix, and Ω is the finite element number. e The integral domain is [E]. This is obtained through calculation using finite element software (such as ANSYS). The calculation steps are as follows:

[0052] Step 1: Establish the geometric model of the motor;

[0053] Step 2: Set material properties and boundary conditions;

[0054] Step 3: Perform mesh generation;

[0055] Step 4: Solve Maxwell's equations.

[0056] Radial electromagnetic force distribution vector F r The calculation formula is as follows:

[0057]

[0058] In the formula, μ0 is the free permeability, which takes the value of 4π × 10⁻⁶. -7H / m; B n and B t The normal and tangential magnetic induction intensities are respectively obtained by measuring them using a Hall sensor.

[0059] Radial electromagnetic force vector F r Fast Fourier Transform:

[0060]

[0061] Where k is the frequency index, N is the number of sampling points, and j is the imaginary unit. H(k) is calculated using the FFT algorithm, with the specific steps as follows:

[0062] Step 1: For F r Pad with zeros so that N is an integer power of 2;

[0063] Step 2: Calculate H(k) using the radix-2 FFT algorithm.

[0064] Matrix equations of the dynamic model of the motor rotor system:

[0065]

[0066] In the formula, [M] is the mass matrix, [C] is the damping matrix, [K] is the stiffness matrix, x is the displacement vector, F(t) is the external force vector, and ξ(t) is the random disturbance. [M], [C], and [K] are obtained through modal analysis and system identification. F(t) is calculated from the electromagnetic force. ξ(t) represents the unmodeled disturbance, assumed to be Gaussian white noise.

[0067] Standard vibration response matrix [R] std The calculation formula for ] is:

[0068] [R std ]=([K]-ω 2 [M]+jω[C]) -1 [H std ];

[0069] Where ω is the angular frequency. [R] std This was obtained through frequency domain analysis, and the specific steps are as follows:

[0070] Step 1: Construct the frequency response function (FRF) matrix ([K]-ω) 2 [M]+jω[C]) -1 ;

[0071] Step 2: Place [H] std Using [R] as input, calculate the output response [R]. std ].

[0072] Wave vibration response matrix [R] flucThe calculation formula for ] is:

[0073] [R fluc ]=([K]-ω 2 [M]+jω[C]) -1 [H fluc ];

[0074] [R fluc The calculation method of [R] is the same as that of [R]. std The same, except the input is changed to [H] fluc ].

[0075] Standard harmonic attenuation coefficient matrix [S] std Definition of ]:

[0076] [S std ] = [I] - [R std [R0] -1 ;

[0077] In the formula, [I] is the identity matrix, and [R0] is the vibration response matrix without any mitigation measures. [R0] is obtained through benchmark tests.

[0078] Wave harmonic attenuation coefficient matrix [S] fluc Definition of ]:

[0079] [S fluc ] = [I] - [R fluc [R0] -1 ;

[0080] [S fluc The calculation method of [S] and [S] std The same, except that [R] is used instead of [R]. std Replace ] with [R fluc ].

[0081] Standard evaluation index vector E std Calculation:

[0082] E std =[W std ][S std ]1;

[0083] Among them, [W std [W] represents the weight matrix, and 1 represents a column vector consisting entirely of 1s. std Determined through expert experience or optimization algorithms.

[0084] Volatility Assessment Index Vector E fluc Calculation:

[0085] F fluc =[W fluc ][S fluc ]1;

[0086] E fluc The calculation method and E std Same, except that [w std ] and [S Std Replace [W] with [W] respectively fluc ] and [S fluc ].

[0087] Forward propagation formula for neural networks:

[0088]

[0089] In the formula, y is the output, f is the activation function (such as ReLU or sigmoid), and w i As the weight, x i Let b be the input, b be the bias vector, and n be the number of input features. The weights and biases are obtained through backpropagation optimization.

[0090] Loss function of neural networks:

[0091]

[0092] Where m is the number of samples, y i For the true value, Let be the predicted value, and λ be the regularization coefficient. λ is determined through cross-validation and typically ranges from 0.001 to 0.1.

[0093] The prediction formula for the evaluation model:

[0094]

[0095] In the formula, Let [W] be the predicted evaluation index vector, [W] be the weight matrix, M be the measurement vector, b be the bias vector, and f be the activation function. [W] and b are obtained through neural network training.

[0096] Quantitative indicators of standard harmonic attenuation effect:

[0097]

[0098] Where N represents the dimension of the evaluation index, and E std,i This is the i-th component of the standard evaluation index vector.

[0099] Quantitative indicators of the attenuation effect of wave harmonics:

[0100]

[0101] Q fluc The calculation method and Q std Same, except E std,iReplace with F fluc,i .

[0102] The formula for calculating the comprehensive evaluation results is as follows:

[0103] Q total =w std Q std +w fluc Q fluc ;

[0104] In the formula, Q total To reduce the overall effect, w std and w fluc These are the weighting coefficients for the standard harmonics and the wave harmonics, respectively. The weighting coefficients are determined using multi-objective optimization methods, such as genetic algorithms or particle swarm optimization algorithms.

[0105] Compared with existing technologies, the present invention provides a method for evaluating the harmonic attenuation effect of radial electromagnetic force on a motor under load. This method collects key physical quantities such as magnetic flux density, vibration, current, and torque during motor operation to construct an electromagnetic-mechanical coupling model. Using frequency domain analysis and dynamic modeling, it calculates the attenuation effect on standard harmonics and wave harmonics, respectively. Finally, the method trains a neural network-based intelligent evaluation model that can quickly and accurately predict the harmonic attenuation effect under motor load conditions.

[0106] Compared with existing technologies for reducing harmonics in the radial electromagnetic force of motors, the method of this invention has the following significant advantages:

[0107] 1. Comprehensive analysis of motor operating characteristics. Through multi-sensor data fusion, rich physical quantity information during motor operation can be obtained, including magnetic flux density distribution, vibration response, current waveform, and load torque. This lays a solid foundation for subsequent electromagnetic-mechanical coupling modeling and harmonic attenuation effect evaluation.

[0108] 2. Precise calculation of harmonic attenuation effect. Based on frequency domain analysis and dynamic modeling methods, this invention can calculate the attenuation effect of standard harmonics and wave harmonics separately, and provide quantitative indicators. This refined evaluation method can better guide the optimization of harmonic attenuation measures in engineering practice.

[0109] 3. Enables rapid intelligent assessment. The intelligent assessment model built using neural networks can quickly predict harmonic attenuation effects based on a limited amount of input data. This data-driven assessment approach significantly improves the efficiency of diagnosis and optimization.

[0110] 4. Supports the evaluation of multiple harmonic mitigation measures. The method of this invention is applicable to evaluating the effects of various harmonic mitigation measures such as electromagnetic structure optimization, current waveform control, and vibration absorption, providing comprehensive support for motor design and operation optimization.

[0111] In summary, this invention achieves rapid and accurate assessment of the harmonic reduction effect of radial electromagnetic force on motor load through multi-source sensor data fusion, electromagnetic-mechanical coupling modeling, and intelligent evaluation model. It solves the technical problem that the current assessment of the harmonic reduction effect of radial electromagnetic force on motor often requires a large number of experiments and engineers' experience, which is time-consuming and labor-intensive. Attached Figure Description

[0112] Figure 1 A flowchart of the method provided by the present invention;

[0113] Figure 2 It is an air gap magnetic flux density diagram;

[0114] Figure 3 This is a rotor vibration acceleration diagram;

[0115] Figure 4 This is a stator current diagram;

[0116] Figure 5 This is a rotor torque diagram;

[0117] Figure 6 This is a diagram showing the radial electromagnetic force distribution in the air gap of the motor.

[0118] Figure 7 These are harmonic amplitude distribution diagrams under standard and fluctuating operating conditions;

[0119] Figure 8 This is the frequency response function diagram under standard operating conditions;

[0120] Figure 9 It is a frequency response function graph under fluctuating operating conditions. Detailed Implementation

[0121] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0122] like Figure 1 The diagram shown is a flowchart of a method for evaluating the harmonic attenuation effect of radial electromagnetic force on a motor according to the present invention. This method includes the following steps:

[0123] S10. Collect sensor data including magnetic flux density, vibration, current and torque, and construct a measurement vector; process the collected data in segments according to a preset sliding time window; at the same time, establish the matrix equation of the coupling model of the motor electromagnetic field and mechanical structure, and use the data in the measurement vector to calibrate and update the model parameters.

[0124] S20. Discretize the coupled model using the finite element method to obtain the electromagnetic field distribution matrix, and verify and correct it using magnetic flux density data measured by the Hall sensor; calculate the radial electromagnetic force distribution vector in the air gap of the motor based on the Maxwell stress tensor method and the magnetic flux density data measured by the Hall sensor.

[0125] S30. Perform a fast Fourier transform on the radial electromagnetic force vector to obtain the harmonic coefficient matrix, and decompose it into the standard harmonic coefficient matrix and the wave harmonic coefficient matrix; at the same time, establish the dynamic model matrix equation of the motor rotor system, and use the measurement data of the acceleration sensor to calibrate and update the relevant matrices.

[0126] S40. Input the standard harmonic coefficient matrix and the wave harmonic coefficient matrix into the dynamic model as excitation, and calculate the standard vibration response matrix and the wave vibration response matrix; verify and correct these response matrices using the measured data from the accelerometer.

[0127] S50. Define the standard harmonic attenuation coefficient matrix and the wave harmonic attenuation coefficient matrix, based on the vibration response matrix and the vibration response matrix when no attenuation measures are taken; calculate the standard evaluation index vector and the wave evaluation index vector, and use the weight matrix to reflect the importance of different harmonic orders.

[0128] S60. Establish a training dataset containing multiple sets of historical data, including the measurement vector, standard harmonic coefficient matrix, fluctuating harmonic coefficient matrix, and evaluation exponent vector for each sliding time window; train a neural network model to obtain an evaluation model for the harmonic attenuation effect of load radial electromagnetic force.

[0129] S70. For new motor load conditions, obtain the measurement vector; use the trained load radial electromagnetic force harmonic attenuation effect evaluation model to predict the standard evaluation index vector and the fluctuation evaluation index vector, so as to achieve fast and accurate harmonic attenuation effect evaluation.

[0130] S80. Based on the components of the standard evaluation index vector and the fluctuation evaluation index vector, perform a quantitative analysis of the harmonic attenuation effect of the radial electromagnetic force of the motor load; evaluate the attenuation degree of the standard harmonics and the fluctuation harmonics, and identify the main influencing factors.

[0131] S90 outputs comprehensive evaluation results, including standard harmonic attenuation effect and fluctuation harmonic attenuation effect.

[0132] The specific implementation methods of the above steps are described in detail below:

[0133] The specific implementation of step S10 involves acquiring sensor data including magnetic flux density, vibration, current, and torque to construct a measurement vector {M}. The expression for the measurement vector M is as follows:

[0134]

[0135] Where, m i Let M represent the measurement value of the i-th sensor, where n is the total number of sensors. This measurement vector M is obtained experimentally, with the specific steps as follows:

[0136] Step 1: Install high-precision sensors, including Hall sensors, accelerometers, current sensors, and torque sensors.

[0137] Step 2: Set the sampling frequency, usually 1kHz-10kHz.

[0138] Step 3: Collect data from each sensor synchronously during motor operation.

[0139] Step 4: Preprocess the collected data, such as denoising and normalization.

[0140] The collected data is segmented according to a preset sliding time window. Simultaneously, the matrix equation [A]{X}={B} for the coupled model of the motor's electromagnetic field and mechanical structure is established, where [A] is the system matrix, {X} is the state vector, and {B} is the input vector. The initial value of [A] is obtained through finite element analysis and subsequently updated through iterative optimization. {X} and {B} are obtained through experimental measurements. The model error ∈ is initially set to zero. The update formula for the system matrix [A] is:

[0141]

[0142] Where k represents the number of iterations and η is the learning rate. The value of η is usually in the range of 0.01-0.1, and the optimal value is determined by cross-validation.

[0143] The specific implementation of step S20 involves discretizing the coupled model using the finite element method to obtain the electromagnetic field distribution matrix [E]. The discretized expression for the electromagnetic field distribution matrix [E] is as follows:

[0144]

[0145] Where, N e [N] is the number of finite element elements, [D] is the shape function matrix, [D] is the electromagnetic material property matrix, and Ω is the finite element number. e The integral domain is defined by [E]. This is obtained through calculations using finite element software such as ANSYS.

[0146] Based on the Maxwell stress tensor method and magnetic flux density data measured by a Hall sensor, the radial electromagnetic force distribution vector {F} in the air gap of the motor is calculated. r Radial electromagnetic force distribution vector {F} r The formula for calculating} is:

[0147]

[0148] Where μ0 is the free permeability, with a value of 4π × 10⁻⁶. -7 H / m; B n and B t The normal and tangential magnetic induction intensities are respectively obtained by measuring them using a Hall sensor.

[0149] The specific implementation of step S30 is to work on the radial electromagnetic force vector {F} r Perform a Fast Fourier Transform to obtain the harmonic coefficient matrix [H]. Radial electromagnetic force vector {F} r The fast Fourier transform expression for} is:

[0150]

[0151] Where k is the frequency index, N is the number of sampling points, and j is the imaginary unit. H(k) is calculated using the FFT algorithm.

[0152] Simultaneously, the dynamic model matrix equations of the motor rotor system are established:

[0153]

[0154] Where [M] is the mass matrix, [C] is the damping matrix, [K] is the stiffness matrix, x is the displacement vector, F(t) is the external force vector, and ξ(t) is the random disturbance. [M], [C], and [K] are obtained through modal analysis and system identification. F(t) is calculated from the electromagnetic force. ξ(t) represents the unmodeled disturbance, assumed to be Gaussian white noise.

[0155] The specific implementation of step S40 is to use the standard harmonic coefficient matrix [H] std ] and the wave harmonic coefficient matrix [H fluc The excitation inputs are respectively used as excitation inputs into the dynamic model to calculate the standard vibration response matrix [R]. std ] and wave vibration response matrix [R fluc Standard vibration response matrix [R] std The formula for calculating ] is:

[0156] [R std ]=([K]-ω 2 [M]+jω[C]) -1 [H std ];

[0157] Wave vibration response matrix [R] fluc The formula for calculating ] is:

[0158] [R fluc ]=([K]-ω 2[M]+jω[C]) -1 [H fluc ];

[0159] Where ω is the angular frequency.

[0160] The specific implementation of step S50 is to define a standard harmonic attenuation coefficient matrix [S std ] and the wave harmonic attenuation coefficient matrix [S fluc Standard harmonic attenuation coefficient matrix [S] std The definition of ] is:

[0161] [S std ] = [I] - [R std [R0] -1 ;

[0162] Wave harmonic attenuation coefficient matrix [S] fluc The definition of ] is:

[0163] [S fluc ] = [I] - [R fluc [R0] -1 ;

[0164] Where [I] is the identity matrix and [R0] is the vibration response matrix when no mitigation measures are taken.

[0165] Next, the standard evaluation index vector {E} is calculated. std} and volatility assessment index vector {E fluc}. Standard evaluation index vector {E std The formula for calculating} is:

[0166] {E std} = [W std ][S std ]{1};

[0167] Volatility Assessment Index Vector {E fluc The formula for calculating} is:

[0168] {E fluc} = [W fluc ][S fluc ]{1};

[0169] Among them, [W std ] and [W fluc [ ] is the weight matrix.

[0170] The specific implementation of step S60 involves establishing a training dataset containing multiple sets of historical data and training a convolutional neural network. The forward propagation formula of the neural network is:

[0171]

[0172] Where y is the output, f is the activation function (such as ReLU or sigmoid), and w i As the weight, x i Let b be the input, b be the bias vector, and n be the number of input features. The loss function of the neural network is:

[0173]

[0174] Where m is the number of samples, y i For the true value, Let be the predicted value, and λ be the regularization coefficient. λ is determined through cross-validation and typically ranges from 0.001 to 0.1.

[0175] The specific implementation of step S70 is to use the trained evaluation model to predict the standard evaluation index vector {E}. std} and volatility assessment index vector {E fluc The prediction formula for the evaluation model is:

[0176]

[0177] in, Let [W] be the predicted evaluation index vector, [W] be the weight matrix, M be the measurement vector, b be the bias vector, and f be the activation function.

[0178] The specific implementation of step S80 is based on {E std} and {E fluc The components of} are used to quantitatively analyze the harmonic attenuation effect on the radial electromagnetic force of the motor load. The quantitative index Q for the standard harmonic attenuation effect is used. std The calculation formula is:

[0179]

[0180] Q, a quantitative indicator of the attenuation effect of wave harmonics fluc The calculation formula is:

[0181]

[0182] Overall evaluation results Q total The calculation formula is:

[0183] Q total =w std Q std +w fluc Q fluc ;

[0184] Among them, w std and w fluc These are the weighting coefficients for standard harmonics and wave harmonics, respectively.

[0185] In summary, this method collects motor operation data through multiple sensors, establishes an electromagnetic-mechanical coupling model, uses frequency domain analysis and dynamic modeling methods to calculate the attenuation effect of standard harmonics and wave harmonics, and finally trains an intelligent evaluation model based on neural networks to achieve a fast and accurate evaluation of harmonic attenuation effect.

[0186] Specifically, the principle of this invention is:

[0187] 1. Multi-sensor data fusion to construct measurement vectors. Sensor data containing key physical quantities such as magnetic flux density, vibration, current, and torque are collected to establish a measurement vector {M}. These physical quantities comprehensively reflect the electromagnetic and mechanical characteristics of the motor during operation.

[0188] 2. Establish an electromagnetic-mechanical coupling model. The collected data is segmented according to a preset sliding time window. Simultaneously, the matrix equations for the coupling between the motor's electromagnetic field and mechanical structure are established: [A]{X}={B}. This coupling model can describe the dynamic response process of the motor system under electromagnetic excitation.

[0189] 3. Calculate the harmonic characteristics of the radial electromagnetic force. The coupled model is discretized using the finite element method to obtain the electromagnetic field distribution matrix [E]. Combining Maxwell's stress tensor theory and magnetic flux density data measured by a Hall sensor, the radial electromagnetic force distribution vector {F} in the motor air gap can be calculated. r Further, regarding {F} r Perform a Fast Fourier Transform to obtain the standard harmonic coefficient matrix [H]. std ] and the wave harmonic coefficient matrix [H fluc ].

[0190] 4. Analyze the vibration response characteristics. [H] std ] and [H fluc The excitation inputs are respectively applied to the dynamic model of the motor rotor system, and the standard vibration response matrix [R] is calculated. std ] and wave vibration response matrix [R fluc This allows us to predict the vibration response characteristics of the motor under standard harmonic and fluctuating harmonic excitation.

[0191] 5. Quantization of harmonic attenuation effect. Based on [R std ]、[R fluc The standard harmonic attenuation coefficient matrix [S] is calculated from the baseline response [R0] without attenuation measures. std ] and the wave harmonic attenuation coefficient matrix [S fluc Further define the standard evaluation index vector {E}. std} and volatility assessment index vector {E fluc} is used to quantitatively describe the effect of harmonic attenuation.

[0192] 6. Train the intelligent evaluation model. Establish a training set containing historical data, including the measurement vector {M} and the harmonic coefficient matrix [H]. std ]、[H fluc and the evaluation index vector {E} std} and {E fluc Training a neural network-based model can quickly predict the harmonic attenuation effect under new operating conditions.

[0193] To better understand and implement this invention, a specific application scenario is provided below: A three-phase asynchronous motor manufactured by a certain company is widely used in production facilities in industries such as metallurgy and chemicals. The motor has a rated power of 800kW and a rated voltage of 6kV. To evaluate the harmonic reduction effect of the motor's radial electromagnetic force under load, the company used the method proposed in this invention for testing and analysis.

[0194] First, 18 Hall effect sensors were evenly arranged on the stator core of the motor to measure the air gap magnetic flux density distribution. Three acceleration sensors were installed on the rotor bearing housing to measure rotor vibration. Current sensors were installed at six key points in the stator windings to measure stator current. One torque sensor was installed on the shaft to measure load torque. The sampling frequency was set to 5kHz, and data from each sensor was collected synchronously under the motor's rated load conditions. The collected raw data was denoised and normalized to construct a measurement vector {M}. Typical waveforms of each physical quantity in the measurement vector {M} are shown below. Figure 2-5 As shown in the figure, it can be seen that the waveforms of each physical quantity have obvious harmonic components, which will have an adverse effect on the operating performance of the motor.

[0195] Based on the data in the measurement vector {M}, the matrix equation [A]{X}={B} for the coupling of the motor's electromagnetic field and mechanical structure was first established. The initial parameters of this coupled model were obtained through finite element analysis, and the model parameters were subsequently calibrated and updated using measurement data through iterative optimization.

[0196] Next, the coupled model is discretized using the finite element method to obtain the electromagnetic field distribution matrix [E]. By solving the [E] matrix and combining it with the magnetic flux density data measured by the Hall sensor, the radial electromagnetic force distribution vector {F} in the air gap of the motor can be calculated. r} Figure 6 The distribution of radial electromagnetic force in the air gap of the motor as a function of rotor position is given.

[0197] It can be seen that the radial electromagnetic force exhibits obvious periodic fluctuations, indicating the presence of complex harmonic components.

[0198] To further analyze the harmonic characteristics, for {F r Perform a Fast Fourier Transform to obtain the standard harmonic coefficient matrix [H]. std ] and the wave harmonic coefficient matrix [H fluc ]. Figure 7 The amplitude distribution of each harmonic order is presented under standard and fluctuating operating conditions. As can be seen from the figure, the standard operating condition mainly contains 3rd, 5th, and 7th order harmonic components, while the fluctuating operating condition contains more higher order harmonic components. These harmonic components will adversely affect the vibration characteristics of the motor.

[0199] To quantitatively evaluate the harmonic attenuation effect, the method of this invention establishes the dynamic model matrix equations of the motor rotor system. The parameters [M], [C], and [K] of this dynamic model were obtained through modal analysis and system identification. [H] std ] and [H fluc Using these as excitation inputs, the standard vibration response matrix [R] is calculated. std ] and wave vibration response matrix [R fluc ]. Figure 8 and Figure 9 The frequency response functions of the motor rotor system under standard and fluctuating operating conditions are presented. As can be seen from the figure, the main resonant frequencies of the motor system under standard operating conditions are concentrated around 300Hz, while under fluctuating operating conditions, there are more resonant peaks and a wider frequency range. This indicates that fluctuating harmonics have a more significant impact on the motor's vibration response.

[0200] Next, the method of the present invention defines the standard harmonic attenuation coefficient matrix [S]. std ] and the wave harmonic attenuation coefficient matrix [S fluc ]. [S] std ] and [S fluc ] and the preset weight matrix [W std ] and [W fluc Multiplying them together yields the standard evaluation exponent vector {E}. std} and volatility assessment index vector {E fluc Table 1 presents the evaluation index results under different operating conditions.

[0201] Table 1. Evaluation index results under different working conditions

[0202] Operating conditions <![CDATA[E std,1 ]]> <![CDATA[E std,2 ]]> <![CDATA[E std,3 ]]> <![CDATA[E fluc,1 ]]> <![CDATA[E fluc,2 ]]> <![CDATA[E fluc,3 ;]]> standard 0.72 0.81 0.75 0.85 0.88 0.91 fluctuation 0.61 0.68 0.63 0.78 0.83 0.86

[0203] As can be seen from the table, the evaluation index under standard operating conditions is generally high, indicating a good reduction effect on standard harmonics. However, the evaluation index under fluctuating operating conditions is relatively low, indicating that the reduction effect on fluctuating harmonics needs further improvement. Overall, the reduction effect of radial electromagnetic force harmonics under this motor load requires further optimization.

[0204] To quickly evaluate the harmonic attenuation effect under new operating conditions, the method of this invention also trains an intelligent evaluation model based on a neural network. The input of this model is the measurement vector {M}, and the output is {E}. std} and {E fluc The model comprises three branches, used to predict the standard harmonic evaluation index, the fluctuating harmonic evaluation index, and the comprehensive evaluation index, respectively. Each branch consists of three fully connected layers, using ReLU as the activation function. The model was trained using the Adam optimization algorithm, with the loss function being a weighted sum of mean squared error and L2 regularization. Cross-validation determined the optimal regularization coefficient λ = 0.05.

[0205] The prediction accuracy of the intelligent evaluation model on the test set after training is shown in Table 2.

[0206] Table 2 Prediction accuracy of neural network models

[0207] index Predicted value actual value relative error <![CDATA[E std,1 ]]> 0.70 0.72 2.8％ <![CDATA[E std,2 ]]> 0.79 0.81 2.5％ <![CDATA[E std,3 ]]> 0.73 0.75 2.7％ <![CDATA[E fluc,1 ]]> 0.82 0.85 3.5％ <![CDATA[E fluc,2 ]]> 0.86 0.88 2.3％ <![CDATA[E fluc,3 ]]> 0.89 0.91 2.2％ <![CDATA[Q total ]]> 0.78 0.80 2.5％

[0208] As can be seen from the table, the intelligent evaluation model can accurately predict the attenuation effect of standard harmonics and wave harmonics, with the average relative error controlled within 3%, which meets the requirements of engineering applications.

[0209] Based on the above analysis, the following conclusions can be drawn:

[0210] 1. Under standard operating conditions, this motor mainly exhibits 3rd, 5th, and 7th order harmonic components, while under fluctuating operating conditions, it exhibits even more high-order harmonic components. These harmonics have a significant impact on the motor's vibration characteristics.

[0211] 2. Under standard operating conditions, the main resonant frequency of the motor system is concentrated around 300Hz, while under fluctuating operating conditions, there are more resonant peaks and a wider frequency range. This indicates that fluctuating harmonics have a more significant impact on the motor's vibration response.

[0212] 3. Through quantitative analysis of the evaluation index, the attenuation effect of standard harmonics is relatively good, while the attenuation effect of fluctuating harmonics needs further optimization. Overall, the attenuation effect of radial electromagnetic force harmonics on the motor load needs to be improved.

[0213] 4. The intelligent evaluation model based on neural networks can quickly and accurately predict the harmonic attenuation effect under new operating conditions, providing effective support for subsequent motor design optimization.

[0214] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for evaluating the harmonic attenuation effect of radial electromagnetic force on a motor load, characterized in that, Includes the following steps: S10. Collect sensor data including magnetic flux density, vibration, current and torque, and construct a measurement vector; process the collected data in segments according to a preset sliding time window; at the same time, establish the matrix equation of the coupling model of the motor electromagnetic field and mechanical structure, and use the data in the measurement vector to calibrate and update the model parameters. S20. Discretize the coupled model using the finite element method to obtain the electromagnetic field distribution matrix, and verify and correct it using magnetic flux density data measured by the Hall sensor; calculate the radial electromagnetic force vector in the air gap of the motor based on the Maxwell stress tensor method and the magnetic flux density data measured by the Hall sensor. S30. Perform a fast Fourier transform on the radial electromagnetic force vector to obtain the harmonic coefficient matrix, and decompose it into a standard harmonic coefficient matrix and a fluctuating harmonic coefficient matrix; simultaneously, establish the dynamic model matrix equations of the motor rotor system: Where [M] is the mass matrix, [C] is the damping matrix, [K] is the stiffness matrix, x is the displacement vector, F(t) is the external force vector, and ξ(t) is the random disturbance; The correlation matrix is ​​calibrated and updated using measurement data from the accelerometer. S40. Input the standard harmonic coefficient matrix and the wave harmonic coefficient matrix into the dynamic model as excitation, and calculate the standard vibration response matrix and the wave vibration response matrix; verify and correct these response matrices using the measured data from the accelerometer. S50. Define the standard harmonic attenuation coefficient matrix and the wave harmonic attenuation coefficient matrix, based on the vibration response matrix and the vibration response matrix when no attenuation measures are taken; calculate the standard evaluation index vector and the wave evaluation index vector, and use the weight matrix to reflect the importance of different harmonic orders. S60. Establish a training dataset containing multiple sets of historical data, including the measurement vector, standard harmonic coefficient matrix, fluctuating harmonic coefficient matrix, and evaluation exponent vector for each sliding time window; train a neural network model to obtain an evaluation model for the harmonic attenuation effect of load radial electromagnetic force. S70. For new motor load conditions, obtain the measurement vector; use the trained load radial electromagnetic force harmonic attenuation effect evaluation model to predict the standard evaluation index vector and the fluctuation evaluation index vector, so as to achieve fast and accurate harmonic attenuation effect evaluation. S80. Based on the components of the standard evaluation index vector and the fluctuation evaluation index vector, perform a quantitative analysis of the harmonic attenuation effect of the radial electromagnetic force of the motor load; evaluate the attenuation degree of the standard harmonics and the fluctuation harmonics, and identify the main influencing factors. S90, Output comprehensive evaluation results, including standard harmonic attenuation effect and fluctuation harmonic attenuation effect; The calibration and update steps specifically involve: segmenting the collected data according to a preset sliding time window; simultaneously, establishing the matrix equation [A]{X}={B} for the coupled model of the motor's electromagnetic field and mechanical structure, where [A] is the system matrix, {X} is the state vector, and {B} is the input vector; the update formula for the system matrix [A] is: Where k represents the number of iterations and η is the learning rate.

2. The method for evaluating the harmonic attenuation effect of radial electromagnetic force on a motor according to claim 1, characterized in that, The measurement vector is obtained through a sensor system, which includes: A Hall sensor array is uniformly arranged on the stator core to measure the air gap magnetic flux density distribution; An acceleration sensor is installed on the rotor bearing housing to measure rotor vibration; Current sensors are installed at key points in the stator winding to measure the stator current. A torque sensor is installed on the shaft to measure the load torque.

3. The method for evaluating the harmonic attenuation effect of radial electromagnetic force on a motor load according to claim 2, characterized in that, The sampling frequency of the sensor system is 1kHz-10kHz.

4. The method for evaluating the harmonic attenuation effect of radial electromagnetic force on a motor load according to claim 3, characterized in that, The sensor system preprocesses the collected data, including noise reduction and normalization.

5. The method for evaluating the harmonic attenuation effect of radial electromagnetic force on a motor according to claim 4, characterized in that, The preset sliding time window length is 10 to 20 times the sampling period.

6. The method for evaluating the harmonic attenuation effect of radial electromagnetic force on a motor according to claim 5, characterized in that, The method for performing a fast Fourier transform on the radial electromagnetic force vector is the radix-2 FFT algorithm.

7. The method for evaluating the harmonic attenuation effect of radial electromagnetic force on a motor according to claim 6, characterized in that, The neural network model adopts a convolutional neural network structure.

8. The method for evaluating the harmonic attenuation effect of radial electromagnetic force on a motor according to claim 7, characterized in that, The quantitative indicators of standard harmonic attenuation effect are as follows: In the formula, N represents the dimension of the evaluation index, and E std,i This is the i-th component of the standard evaluation index vector.

9. The method for evaluating the harmonic attenuation effect of radial electromagnetic force on a motor according to claim 8, characterized in that, The quantitative indicators of the harmonic attenuation effect are as follows: In the formula, E fluc,i Let i be the i-th component of the volatility assessment index vector.

10. The method for evaluating the harmonic attenuation effect of radial electromagnetic force on a motor according to claim 9, characterized in that, The formula for calculating the comprehensive evaluation results is as follows: Q total =w std Q std +w fluc Q fluc ; In the formula, Q total To reduce the overall effect, w std and w fluc These are the weighting coefficients for standard harmonics and wave harmonics, respectively.

Images