A motor multi-physics field agent model construction method, device, equipment and medium

Abstract

The present application relates to the technical field of model construction, and discloses a motor multi-physical field agent model construction method, device, equipment and medium, the present application firstly carries out a conventional sampling in the preset sample three-dimensional space by using the Latin hypercube sampling method, then determines the new sample point in combination with the field operation condition data set and carries out local secondary sampling, the accuracy of sample space point collection and the fitting degree with field data are improved. Finally, the model is repeatedly trained until the root mean square error meets the requirements and the average error meets the requirements, and the final trained motor multi-physical field target agent model is obtained. Through the new field data sample point used for the precision improvement of the agent model, the dataset error of the initial construction of the agent model is minimized, meanwhile, the fitting with the field operation is increased, and the error caused by the multi-physical field simulation is reduced.

Description

Technical Field

[0001] This invention relates to the field of model building technology, specifically to a method, apparatus, equipment, and medium for building a multiphysics proxy model of an electric motor. Background Technology

[0002] As the complexity of the model increases, the computational load in multiphysics simulation also increases. In the field of digital twin intelligent operation and maintenance of hydropower and pumped storage unit motors, different methods are needed to construct real-time simulation models for different models, thereby achieving dimensionality reduction of numerical calculation order or data fitting, and realizing real-time display of multiphysics of each component through visualization.

[0003] The surrogate model is an efficient mathematical approximation model that has gradually developed into a new computational method due to its significant improvement in computational efficiency. Sampling methods study how to construct a fitting function with the minimum error using the fewest sample points and can account for random errors in the data, while the accuracy of the surrogate model depends on the distribution of the sample points.

[0004] Existing sample space sampling methods include orthogonal design, uniform design, and Latin hypercube design. During the sampling process, a surrogate model is often constructed through a single sampling. However, the distribution of a single sampling often contains errors and uncertainties, and it cannot be optimized for areas with insufficient accuracy. Furthermore, the surrogate model constructed after the initial sampling may not meet the field requirements, thus necessitating secondary sampling. Conventional secondary sampling still adds data based on Latin hypercube sampling points without combining it with actual field operating data, and there is no precise method to determine the most frequently operating conditions for the newly added units. Therefore, in the process of selecting new sample spaces, relying solely on experience to divide the sampling area without considering field operating data is incomplete and does not match the actual long-term operating range of the power plant, resulting in insufficient accuracy of the constructed surrogate model and increasing the time and cost of surrogate model construction. Summary of the Invention

[0005] In view of this, the present invention provides a method, apparatus, equipment and medium for constructing a multiphysics proxy model of an electric motor, in order to solve the problem that in the process of selecting new sample space, it is incomplete to divide the sampling area based solely on experience without relying on field operation data, which does not match the long-term operation range of the actual power plant, resulting in insufficient accuracy of the constructed proxy model, and also increasing the time and cost of constructing the proxy model.

[0006] In a first aspect, the present invention provides a method for constructing a multiphysics proxy model for an electric motor, the method comprising:

[0007] Obtain the on-site operating condition dataset of the target motor; within the preset sample 3D space, perform a first sampling using the Latin hypercube sampling method to determine the first sample point set; based on the first sample point set, after processing with a preset periodic boundary 3D flow thermal coupling model, establish an initial multi-physics proxy model for the motor; sequentially perform root mean square error and average error judgments on the initial multi-physics proxy model for the motor; when the root mean square error meets the requirements but the average error does not, perform a second sampling based on the on-site operating condition dataset to determine the target new sample point set; add the target new sample point set to the first sample point set, and iterate repeatedly to train the model based on the added first sample point set until the root mean square error meets the requirements and the average error meets the requirements, thus obtaining the target multi-physics proxy model for the motor.

[0008] The method for constructing a multiphysics surrogate model for a motor provided by this invention first performs a conventional sampling using the Latin hypercube sampling method within a preset sample three-dimensional space. Then, it combines the field operating condition dataset to determine new sample points and performs local secondary sampling, improving the accuracy of sample space point acquisition and its fit with field data. Finally, the model is repeatedly trained until both the root mean square error and the mean error meet the requirements, resulting in the final trained multiphysics surrogate model for the motor. By adding new field data sample points to improve the accuracy of the surrogate model, the method minimizes the dataset errors in the initial construction of the surrogate model, while increasing the fit with field operation and reducing errors caused by multiphysics simulation.

[0009] In one optional implementation, based on the first set of sample points and processed by a preset periodic boundary three-dimensional flow thermal coupling model, an initial proxy model for the multiphysics field of the motor is established, including:

[0010] Based on the first sample point set, numerical simulation is performed using a pre-set periodic boundary three-dimensional flow-thermal coupling model to obtain the first sample dataset; the first sample dataset is then used to establish an initial proxy model for the multi-physics field of the motor.

[0011] The method for constructing a multiphysics proxy model for a motor provided by the present invention can obtain a first sample dataset corresponding to a first sample point set by performing numerical simulation using a pre-set periodic boundary three-dimensional flow-thermal coupling model, and then construct a corresponding initial proxy model for the multiphysics proxy model of the motor based on the first sample dataset.

[0012] In one optional implementation, the root mean square error and average error of the initial surrogate model of the multiphysics motor are judged sequentially, including:

[0013] The root mean square error (RMSE) of the initial surrogate model of the motor multiphysics field is judged. If the RMSE does not meet the requirements, the model parameters are adjusted and iterative training is repeated based on the adjusted model parameters until the RMSE meets the requirements. The average error of the initial surrogate model of the motor multiphysics field is then judged.

[0014] The method for constructing a multi-physics proxy model for a motor provided by this invention can improve the model accuracy by adjusting the model parameters if the root mean square error does not meet the requirements, thereby making the root mean square error meet the requirements.

[0015] In one optional implementation, after sequentially determining the root mean square error and average error of the initial surrogate model of the multiphysics motor, the method further includes:

[0016] When the root mean square error and the average error meet the requirements, the initial surrogate model of the multiphysics field of the motor is determined as the target surrogate model of the multiphysics field of the motor.

[0017] In one optional implementation, when the root mean square error meets the requirement but the average error does not, secondary sampling is performed based on the field operation condition dataset to determine the target set of newly added sample points, including:

[0018] When the root mean square error meets the requirements but the average error does not, the initial operating range dataset is determined based on the field operating condition dataset; based on the initial operating range dataset, the initial set of newly added sample points is determined by processing with the empirical distribution function method; based on the initial set of newly added sample points and the first set of sample points, the target set of newly added sample points is determined by processing with the metric space method.

[0019] The method for constructing a multi-physics surrogate model for a motor provided by this invention, if the root mean square error of the constructed initial multi-physics surrogate model for the motor meets the requirements but the average error does not, then by combining the on-site operating condition dataset, new sample points are determined and local secondary sampling is performed, improving the accuracy of sample space point acquisition and its fit with the on-site data. Furthermore, by combining the new sample points with the first sample point set, the metric space method can be used to remove duplicate points from the two samplings and determine the final target new sample point set, further improving the accuracy of sample space point acquisition and its fit with the on-site data.

[0020] In one optional implementation, based on the initial on-site operating range dataset, and processed using the empirical distribution function method, an initial set of newly added sample points is determined, including:

[0021] The data in the initial on-site operating range dataset are sorted to obtain the on-site target operating range dataset; based on the on-site target operating range dataset, multiple cumulative proportion values ​​are obtained through empirical distribution function calculation; and the initial set of newly added sample points is determined based on the multiple cumulative proportion values.

[0022] Secondly, the present invention provides a device for constructing a multiphysics proxy model for an electric motor, the device comprising:

[0023] The system comprises the following modules: an acquisition module for acquiring the on-site operating condition dataset of the target motor; a first sampling determination module for performing a first sampling using the Latin hypercube sampling method within a preset sample 3D space and determining the first sample point set; a processing module for establishing an initial multi-physics proxy model of the motor based on the first sample point set and processing it using a preset periodic boundary 3D flow thermal coupling model; a judgment module for sequentially judging the root mean square error and average error of the initial multi-physics proxy model of the motor; a second sampling determination module for performing a second sampling based on the on-site operating condition dataset and determining the target new sample point set when the root mean square error meets the requirements but the average error does not; and a repetition module for adding the target new sample point set to the first sample point set and iteratively training the model based on the added first sample point set until both the root mean square error and the average error meet the requirements and the target multi-physics proxy model of the motor is obtained.

[0024] Thirdly, the present invention provides a computer device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the motor multiphysics proxy model construction method of the first aspect or any corresponding embodiment described above.

[0025] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the method for constructing a multiphysics proxy model of a motor as described in the first aspect or any corresponding embodiment thereof.

[0026] Fifthly, the present invention provides a computer program product, including computer instructions, which are used to cause a computer to execute the method for constructing a multiphysics proxy model of an electric motor according to the first aspect or any corresponding embodiment described above. Attached Figure Description

[0027] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0028] Figure 1 This is a flowchart illustrating the method for constructing a multiphysics proxy model for an electric motor according to an embodiment of the present invention.

[0029] Figure 2 This is a flowchart illustrating another method for constructing a multiphysics proxy model for an electric motor according to an embodiment of the present invention;

[0030] Figure 3 This is a flowchart illustrating another method for constructing a multiphysics proxy model for an electric motor according to an embodiment of the present invention;

[0031] Figure 4 This is a flowchart illustrating the method for constructing a multiphysics proxy model of a motor based on data sampling according to an embodiment of the present invention.

[0032] Figure 5 This is a schematic diagram illustrating the process of determining new sample points according to an embodiment of the present invention;

[0033] Figure 6 This is a structural block diagram of a motor multiphysics proxy model construction device according to an embodiment of the present invention;

[0034] Figure 7 This is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. Detailed Implementation

[0035] 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. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0036] This invention provides a method for constructing a multiphysics surrogate model for an electric motor. By employing a single conventional sampling and a local secondary sampling, the accuracy of sample space point acquisition and its fit with field data are improved. Furthermore, by adding new field data sample points to enhance the accuracy of the surrogate model, the errors in the initial dataset construction of the surrogate model are minimized, while simultaneously increasing the fit with field operation and reducing errors introduced by multiphysics simulation.

[0037] According to an embodiment of the present invention, a method for constructing a multiphysics proxy model of an electric motor is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0038] This embodiment provides a method for constructing a multiphysics proxy model for an electric motor, which can be used in electronic devices such as computers, mobile phones, and tablets. Figure 1 This is a flowchart of a method for constructing a multiphysics proxy model for a motor according to an embodiment of the present invention, such as... Figure 1 As shown, the process includes the following steps:

[0039] Step S101: Obtain the on-site operating condition dataset of the target motor.

[0040] Specifically, the initial dataset of field operating conditions can include multiple field operating condition data of the target motor under normal operating conditions.

[0041] In this embodiment, the temperature field proxy model of the stator winding and core of the pumped storage motor is used as an example. The inlet wind speed V, air temperature T1, stator winding heat source P1, and stator core heat source P2 are selected as independent variables, and the average temperature of the stator winding T2 and the average temperature of the stator core T3 are selected as dependent variables to form the initial dataset of field operating conditions.

[0042] Among them, the stator winding heat source P1 can be calculated using the following relationship (1):

[0043] P1 = P copper +P stray (1)

[0044] In the formula: P copper Copper loss (unit: watt, W) is expressed and calculated using the following relationships (2) and (3):

[0045]

[0046]

[0047] In the formula: I rns R represents the effective value of the current (unit: ampere, A); coil The values ​​represent the winding resistance (in ohms, Ω); ρ represents the resistivity of the conductor (in ohm-meters, Ω-m), which is related to the material and temperature; l represents the length of the conductor (in meters, m); and A represents the cross-sectional area of ​​the conductor (in square meters, m²). 2 ).

[0048] Furthermore, P stray Stray loss (unit: watt, W) is calculated using the following relationships (4) to (6):

[0049]

[0050] B=μ·H (5)

[0051]

[0052] In the formula: k represents a proportionality constant, which depends on the material and structure of the winding; B max V represents the maximum magnetic flux density (unit: Tesla, T); f represents the operating frequency (unit: Hertz, Hz); coil Indicates the winding volume (unit: cubic meters, m). 3 B represents magnetic flux density; μ represents permeability (unit: Henry / meter, H / m); H represents magnetic field strength; N represents the number of turns of the coil; I represents the effective length of the coil.

[0053] Furthermore, the stator core heat source P2 can be calculated using the following relationship (7):

[0054] P2 = P h +P e (7)

[0055] In the formula: P h The hysteresis loss (unit: watt, W) is expressed as shown in the following equation (8); P e The eddy current loss (unit: watt, W) is expressed as shown in the following equation (9):

[0056]

[0057]

[0058] In the formula: k1 represents the hysteresis loss coefficient of the material (unit: W / kg); V represents the volume of the iron core (unit: cubic meters, m³). 3 σ represents the electrical conductivity of the material (unit: Siemens / meter, S / m); t represents the thickness of the iron core.

[0059] Step S102: Within the preset sample three-dimensional space, perform a sampling using the Latin hypercube sampling method and determine the first sample point set.

[0060] Among them, the preset sample three-dimensional space refers to the global sample three-dimensional space composed of data ranges set according to experience.

[0061] Specifically, Latin hypercube sampling is performed within the preset sample three-dimensional space to collect N sample point data, i.e., N first sample points, and form the first sample point set.

[0062] In one example, taking the construction of a proxy model for the temperature field of the stator winding and core of a pumped-storage motor as an example, based on experience, the data ranges for the three variables of wind speed, temperature, and current are set respectively [V]. x V y ]、[T x ,T y ]、[I x ,Iy And establish the corresponding preset sample three-dimensional space.

[0063] Step S103: Based on the first sample point set, and after processing by the preset periodic boundary three-dimensional flow-thermal coupling model, an initial proxy model of the motor multiphysics field is established.

[0064] Specifically, the preset periodic boundary three-dimensional flow thermal coupling model can be used for numerical simulation at different operating conditions. Taking the construction of the temperature field proxy model of the stator winding and iron core of the pumped storage motor as an example, the preset periodic boundary three-dimensional flow thermal coupling model is a periodic boundary three-dimensional flow thermal coupling model based on the stator winding and iron core.

[0065] Specifically, numerical simulations are performed on multiple sample points included in the first sample point set using a pre-defined periodic boundary three-dimensional flow-thermal coupling model, and a corresponding initial proxy model for the multiphysics field of the motor is established.

[0066] Step S104: Perform root mean square error and average error judgments on the initial proxy model of the multiphysics field of the motor in sequence.

[0067] Among them, the root mean square error (RMSE) and the mean error (ME) are important indicators for evaluating the accuracy and reliability of a model.

[0068] Specifically, the accuracy and reliability of the model can be ensured by sequentially judging the root mean square error and the average error.

[0069] Step S105: When the root mean square error meets the requirements but the average error does not, perform secondary sampling based on the on-site operating condition dataset and determine the target new sample point set.

[0070] Specifically, if the root mean square error of the initial surrogate model of the multiphysics motor meets the requirements, but the average error does not, then model optimization is required by adding new sample points.

[0071] In this embodiment, secondary sampling can be performed using the on-site operating condition dataset, improving the accuracy of sample spatial point collection and its fit with on-site data. Furthermore, the final target set of newly added sample points is determined through sampling and used to improve the accuracy of the subsequent surrogate model, minimizing the errors in the dataset initially constructed by the surrogate model, while increasing the fit with on-site operation and reducing errors caused by multiphysics simulation.

[0072] Step S106: Add the newly added target sample point set to the first sample point set, and iterate the model training repeatedly based on the added first sample point set until the root mean square error meets the requirements and the average error meets the requirements, and obtain the multi-physics target proxy model of the motor.

[0073] Specifically, the determined target new sample point set is added to the first sample point set to form a new sample point set. Further, based on the obtained new sample point set, steps S103 to S105 are repeated to iteratively train the model until the root mean square error of the model meets the requirements and the average error meets the requirements, at which point the corresponding trained multi-physics target surrogate model of the motor is obtained.

[0074] The method for constructing a multiphysics surrogate model for a motor provided in this embodiment first performs a conventional sampling using the Latin hypercube sampling method within a preset sample three-dimensional space. Then, it combines the field operating condition dataset to determine new sample points and performs local secondary sampling, improving the accuracy of sample space point acquisition and its fit with field data. Finally, the model is repeatedly trained until both the root mean square error and the mean error meet the requirements, resulting in the final trained multiphysics target surrogate model for the motor. By adding new field data sample points to improve the accuracy of the surrogate model, the errors in the initial dataset construction of the surrogate model are minimized, while increasing the fit with field operation and reducing errors caused by multiphysics simulation.

[0075] This embodiment provides a method for constructing a multiphysics proxy model for an electric motor, which can be used in electronic devices such as computers, mobile phones, and tablets. Figure 2 This is a flowchart of a method for constructing a multiphysics proxy model for a motor according to an embodiment of the present invention, such as... Figure 2 As shown, the process includes the following steps:

[0076] Step S201: Obtain the on-site operating condition dataset of the target motor. For details, please refer to [link to relevant documentation]. Figure 1 Step S101 of the illustrated embodiment will not be described again here.

[0077] Step S202: Within the preset sample 3D space, perform a single sampling using the Latin hypercube sampling method to determine the first sample point set. For details, please refer to [link to details]. Figure 1 Step S102 of the illustrated embodiment will not be described again here.

[0078] Step S203: Based on the first sample point set, and after processing by a preset periodic boundary three-dimensional flow-thermal coupling model, an initial proxy model for the multi-physics field of the motor is established.

[0079] Specifically, step S203 includes:

[0080] Step S2031: Based on the first sample point set, numerical simulation is performed using a preset periodic boundary three-dimensional flow-thermal coupling model to obtain the first sample dataset.

[0081] Specifically, taking the construction of a proxy model for the temperature field of the stator winding and core of a pumped-storage motor as an example, for the m operating points corresponding to the target sample point set, based on the periodic boundary three-dimensional flow-thermal coupling model of the stator winding and core, commercial software is used to realize the numerical simulation of the operating points, generate the average temperature of the stator winding and stator core under different operating conditions, and generate the corresponding target dataset A′, as shown in the following relation (10):

[0082]

[0083] Step S2032: Establish an initial proxy model for the multiphysics field of the motor using the first sample dataset.

[0084] First, a convolutional neural network (CNN) was chosen as the training network for the model.

[0085] Secondly, the number of network layers q and the number of neurons in each layer w were determined through experiments. The activation function Sigmoid was selected, the loss function mean squared error (MSE) was determined, and the Adam optimization algorithm was selected to update the network weights.

[0086] Finally, the neural network was trained using the first sample dataset, the network parameters were adjusted to minimize the loss function, and the trained multiphysics initial surrogate model of the motor was obtained.

[0087] Step S204: Perform root mean square error and average error judgments on the initial proxy model of the multiphysics field of the motor in sequence.

[0088] Specifically, step S204 includes:

[0089] Step S2041: Perform root mean square error judgment on the initial proxy model of the multiphysics field of the motor.

[0090] Specifically, determine whether the root mean square error (MSE) of the initial surrogate model of the multiphysics field of the motor meets the requirements.

[0091] Step S2042: When the root mean square error does not meet the requirements, adjust the model parameters and iterate the training repeatedly based on the adjusted model parameters until the root mean square error meets the requirements, and judge the average error of the initial proxy model of the multi-physics field of the motor.

[0092] Specifically, if the root mean square error does not meet the requirements, the network model parameters can be adjusted by adding a regularization term, using Dropout, adjusting the learning rate, etc., and step S2032 can be repeated based on the adjusted network model parameters until the root mean square error meets the requirements.

[0093] Furthermore, when the root mean square error meets the requirements, a second error judgment is performed on the initial surrogate model of the multiphysics field of the motor, that is, an average error judgment is performed. The specific judgment process is as follows:

[0094] Randomly select n on-site operating conditions and their corresponding output temperature values, and determine the output temperature value Y of the selected sample points. n And the surrogate model predicted value Y ′ The average error value of n ρ ave Is the efficiency value lower than the specified value of 10%?

[0095] If yes, it means the average error meets the requirements; if no, it means the average error does not meet the requirements.

[0096] Wherein, the average error value ρ ave It can be calculated using the following relations (11) and (12):

[0097]

[0098] Step S205: When the root mean square error meets the requirements but the average error does not, perform secondary sampling based on the field operation condition dataset and determine the target set of newly added sample points. For details, please refer to [link to relevant documentation]. Figure 1 Step S105 of the illustrated embodiment will not be described again here.

[0099] Step S206: Add the newly added target sample point set to the first sample point set, and iteratively train the model based on the added first sample point set until the root mean square error and the average error meet the requirements, thus obtaining the multi-physics target surrogate model for the motor. For details, please refer to [link to details]. Figure 1 Step S106 of the illustrated embodiment will not be described again here.

[0100] Step S207: When the root mean square error and the average error meet the requirements, the initial surrogate model of the motor multiphysics field is determined as the target surrogate model of the motor multiphysics field.

[0101] Specifically, if both the root mean square error and the average error meet the requirements, the established initial surrogate model of the motor multiphysics field is directly used as the final target surrogate model of the motor multiphysics field.

[0102] The method for constructing a multiphysics surrogate model for a motor provided in this embodiment first performs a conventional sampling using the Latin hypercube sampling method within a preset sample three-dimensional space. Then, numerical simulation using a preset periodic boundary three-dimensional flow-thermal coupling model yields a first sample dataset corresponding to the first sample point set. Based on this first sample dataset, a corresponding initial multiphysics surrogate model for the motor can be constructed. Further, if the root mean square error (RMSE) of the initial multiphysics surrogate model meets the requirements but the mean error does not, new sample points are determined by combining the field operating condition dataset, and local secondary sampling is performed. The model is then repeatedly trained until both the RMS and mean error meet the requirements, resulting in the final trained multiphysics target surrogate model for the motor. This improves the accuracy of sample space point acquisition and the fit with field data. Simultaneously, by adding new field data sample points to improve the accuracy of the surrogate model, the dataset error in the initial construction of the surrogate model is minimized, while increasing the fit with field operation and reducing errors caused by multiphysics simulation.

[0103] This embodiment provides a method for constructing a multiphysics proxy model for an electric motor, which can be used in electronic devices such as computers, mobile phones, and tablets. Figure 3 This is a flowchart of a method for constructing a multiphysics proxy model for a motor according to an embodiment of the present invention, such as... Figure 3 As shown, the process includes the following steps:

[0104] Step S301: Obtain the on-site operating condition dataset of the target motor. For details, please refer to [link to relevant documentation]. Figure 1 Step S101 of the illustrated embodiment will not be described again here.

[0105] Step S302: Within the preset sample 3D space, perform a single sampling using the Latin hypercube sampling method to determine the first sample point set. For details, please refer to [link to relevant documentation]. Figure 1 Step S102 of the illustrated embodiment will not be described again here.

[0106] Step S303: Based on the first sample point set, and after processing with a pre-defined periodic boundary three-dimensional flow-thermal coupling model, an initial multiphysics proxy model for the motor is established. For details, please refer to [link to relevant documentation]. Figure 2 Step S203 of the illustrated embodiment will not be described again here.

[0107] Step S304 involves sequentially determining the root mean square error and average error of the initial multiphysics surrogate model for the motor. For details, please refer to [link to relevant documentation]. Figure 2 Step S204 of the illustrated embodiment will not be described again here.

[0108] Step S305: When the root mean square error meets the requirements but the average error does not, perform secondary sampling based on the on-site operating condition dataset and determine the target new sample point set.

[0109] Specifically, step S305 includes:

[0110] Step S3051: When the root mean square error meets the requirements but the average error does not meet the requirements, determine the initial operating range dataset based on the field operating condition dataset.

[0111] Specifically, when the root mean square error meets the requirements but the mean error does not, the initial operating range dataset can be determined through the following steps:

[0112] (1) Data on wind speed, temperature and current are usually recorded in time series format, including timestamps and corresponding temperature readings.

[0113] (2) Extract the corresponding physical quantity values ​​at 10-minute intervals and sample at equal time intervals to form a dataset of the field operation range.

[0114] Step S3052: Based on the initial on-site operating range dataset, the initial set of newly added sample points is determined through processing using the empirical distribution function method.

[0115] The empirical distribution function (ECDF) method is a statistical approach used to estimate the probability distribution function of sample data. Based on observations of the sample data, this method constructs a cumulative distribution function by calculating the sum of the probability densities of each data point. This function is called the empirical distribution function (ECDF). The ECDF is a non-parametric estimation method applicable to sample data of any distribution form, requiring no assumptions about the data. It is directly calculated through observation of the sample data and is an estimate of the true distribution function.

[0116] In some optional implementations, step S3052 above includes:

[0117] Step a1: Sort the data in the initial on-site operating range dataset to obtain the target on-site operating range dataset.

[0118] Step a2: Based on the on-site target operating range dataset, multiple cumulative proportion values ​​are obtained through empirical distribution function calculation.

[0119] Step a3: Determine the initial set of new sample points based on multiple cumulative ratio values.

[0120] First, sort the data in the initial on-site operating range dataset in ascending order.

[0121] Secondly, for each data point x i Calculate x less than or equal toi The proportion of the number of data points to the total number of data points. Defined as an empirical distribution function, as shown in the following relationship (13):

[0122]

[0123] Then, calculate F for adjacent data points. n (x) difference.

[0124] Finally, select data point F. n (x) The three points with the largest differences are taken as new sample points. Different parameters can be freely combined to form three sets of new sample points, which form the initial set of new sample points.

[0125] Step S3053: Based on the initial set of newly added sample points and the first set of sample points, the target set of newly added sample points is determined through the metric space method.

[0126] The metric space defines a distance function over a set, such that the distance between any two elements in the set can be calculated using this distance function.

[0127] Specifically, the metric space method is used to determine duplicate points. A distance metric (such as Euclidean distance) d(p,q) is defined as shown in the following relation (14):

[0128]

[0129] In the formula: p and q represent two points in three-dimensional space.

[0130] Furthermore, a threshold is set. If the distance between a new sample in the initial set of newly added sample points and any sample point in the known first set of sample points is less than this threshold, they are considered duplicates. In this case, one point in the initial set of newly added sample points obtained from Latin hypercube sampling is deleted, and the final target set of newly added sample points N is obtained. n .

[0131] Furthermore, the final target set of newly added sample points is obtained as m = N + N. n .

[0132] Step S306: Add the newly added target sample point set to the first sample point set, and iteratively train the model based on the added first sample point set until the root mean square error and the average error meet the requirements, thus obtaining the multi-physics target surrogate model for the motor. For details, please refer to [link to details]. Figure 1 Step S106 of the illustrated embodiment will not be described again here.

[0133] The method for constructing a multiphysics surrogate model for a motor provided in this embodiment first performs a conventional sampling using the Latin hypercube sampling method within a preset sample three-dimensional space to establish an initial multiphysics surrogate model for the motor. If the root mean square error of the constructed initial multiphysics surrogate model for the motor meets the requirements but the mean error does not, then new sample points are determined by combining the field operating condition dataset and local secondary sampling is performed, improving the accuracy of sample space point acquisition and its fit with the field data. Furthermore, by combining the new sample points with the first sample point set, the duplicate points from the two samplings can be removed by processing with the metric space method, and the final target new sample point set can be determined, further improving the accuracy of sample space point acquisition and its fit with the field data.

[0134] In one example, a method for constructing a multiphysics proxy model for a motor based on data sampling is provided, and the specific process is as follows: Figure 4 As shown in the figure. The two error requirements are the root mean square error and the average error, respectively; the process for determining new sample points is as follows. Figure 5 As shown.

[0135] The method for constructing a multiphysics proxy model of a motor based on data sampling provided in this example has the following characteristics:

[0136] Beneficial effects:

[0137] 1. Improve the accuracy of the proxy model dataset:

[0138] By identifying the most frequent operating conditions of the unit through on-site data, and by conducting a routine sampling and a secondary sampling for key local operating conditions, and removing duplicate points, the accuracy of the sample spatial point collection and its fit with the on-site data were improved.

[0139] 2. Enhance the accuracy of the proxy model and achieve adaptive optimization:

[0140] By adjusting the surrogate model parameters and adding new field data sample points, the accuracy of the surrogate model was improved, minimizing the errors in the dataset initially constructed by the surrogate model. At the same time, the fit with field operation was increased, reducing the errors caused by multiphysics simulation.

[0141] Therefore, the technical solution of this method makes the construction of the unit equipment proxy model more efficient and accurate. The beneficial effects of this method are not only reflected in improved work efficiency and accuracy, but also in promoting the intelligent and digital transformation of hydropower and pumped storage engineering, providing support for innovation and development in engineering management.

[0142] This embodiment also provides a multiphysics proxy model construction device for an electric motor. This device is used to implement the above embodiments and preferred embodiments, and details already described will not be repeated. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0143] This embodiment provides a device for constructing a multiphysics proxy model for an electric motor, such as... Figure 6 As shown, the device includes:

[0144] The acquisition module 601 is used to acquire the on-site operating condition dataset of the target motor.

[0145] The first sampling determination module 602 is used to perform a sampling and determine the first sample point set in a preset sample three-dimensional space using the Latin hypercube sampling method.

[0146] The processing module 603 is used to establish an initial proxy model of the multiphysics field of the motor based on the first sample point set and processed by a preset periodic boundary three-dimensional flow thermal coupling model.

[0147] The judgment module 604 is used to sequentially judge the root mean square error and average error of the initial proxy model of the multi-physics field of the motor.

[0148] The second sampling determination module 605 is used to perform secondary sampling and determine the target new sample point set based on the field operation condition dataset when the root mean square error meets the requirements but the average error does not.

[0149] The repeating module 606 is used to add the newly added target sample point set to the first sample point set, and repeatedly iterate the model training based on the added first sample point set until the root mean square error meets the requirements and the average error meets the requirements, and the multi-physics target proxy model of the motor is obtained.

[0150] In some alternative implementations, the processing module 603 includes:

[0151] The simulation submodule is used to perform numerical simulation based on the first sample point set using a preset periodic boundary three-dimensional flow thermal coupling model to obtain the first sample dataset.

[0152] A submodule is created to build an initial proxy model of the motor using the first sample dataset.

[0153] In some optional implementations, the determination module 604 includes:

[0154] The first judgment submodule is used to judge the root mean square error of the initial proxy model of the multiphysics field of the motor.

[0155] The adjustment submodule is used to adjust the model parameters when the root mean square error does not meet the requirements, and to iteratively train the model based on the adjusted model parameters until the root mean square error meets the requirements, and to judge the average error of the initial proxy model of the multiphysics field of the motor.

[0156] In some alternative embodiments, the device further includes:

[0157] The determination module is used to determine the initial surrogate model of the motor multiphysics field as the target surrogate model of the motor multiphysics field when the root mean square error and the average error meet the requirements.

[0158] In some optional implementations, the second sampling determination module 605 includes:

[0159] The determination submodule is used to determine the initial operating range dataset based on the field operating condition dataset when the root mean square error meets the requirements but the average error does not.

[0160] The first processing submodule is used to determine the initial set of newly added sample points based on the initial on-site operating range dataset and processed by the empirical distribution function method.

[0161] The second processing submodule is used to determine the target set of new sample points based on the initial set of new sample points and the first set of sample points, using the metric space method.

[0162] In some alternative implementations, the first processing submodule includes:

[0163] The sorting unit is used to sort the data in the initial on-site operating range dataset to obtain the target on-site operating range dataset.

[0164] The calculation unit is used to calculate multiple cumulative proportion values ​​based on the on-site target operating range dataset and through an empirical distribution function.

[0165] The determination unit is used to determine the initial set of new sample points based on multiple cumulative proportion values.

[0166] Further functional descriptions of the above modules and units are the same as those in the corresponding embodiments described above, and will not be repeated here.

[0167] In this embodiment, the motor multiphysics proxy model construction device is presented in the form of a functional unit. Here, a unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that execute one or more software or fixed programs, and / or other devices that can provide the above functions.

[0168] This invention also provides a computer device having the above-described features. Figure 6 The device shown is for constructing a multiphysics proxy model of an electric motor.

[0169] Please see Figure 7 , Figure 7 This is a schematic diagram of the structure of a computer device provided in an optional embodiment of the present invention, such as... Figure 7 As shown, the computer device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system). Figure 7 Take a processor 10 as an example.

[0170] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GDA), or any combination thereof.

[0171] The memory 20 stores instructions executable by at least one processor 10 to cause at least one processor 10 to perform the method shown in the above embodiments.

[0172] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the computer device. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0173] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0174] The computer device also includes a communication interface 30 for communicating with other devices or communication networks.

[0175] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

[0176] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0177] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for constructing a multiphysics proxy model for an electric motor, characterized in that, The method includes: Obtain the on-site operating condition dataset of the target motor. The on-site operating condition dataset includes inlet wind speed, air temperature, stator winding heat source, stator core heat source, stator winding average temperature, and stator core average temperature. The stator winding heat source is determined by copper loss and stray loss, and the stator core heat source is determined by hysteresis loss. Within a preset sample 3D space, a first sample point set is determined by using the Latin hypercube sampling method. The preset sample 3D space represents a global sample 3D space for the multi-physics characteristics of the motor, composed of a data range set based on experience. Based on the first sample point set, after processing by a preset periodic boundary three-dimensional flow thermal coupling model, an initial proxy model of the motor multiphysics field is established. The preset periodic boundary three-dimensional flow thermal coupling model is a periodic boundary three-dimensional flow thermal coupling model based on the stator winding and the iron core, which is used for numerical simulation for different operating conditions. The root mean square error and average error of the initial multiphysics proxy model of the motor are judged sequentially. When the root mean square error meets the requirements but the average error does not, secondary sampling is performed based on the on-site operating condition dataset to determine the target set of new sample points. The target new sample point set is added to the first sample point set, and the model is trained iteratively according to the added first sample point set until the root mean square error meets the requirements and the average error meets the requirements, and the multi-physics target proxy model of the motor is obtained. The process includes sequentially determining the root mean square error and average error of the initial proxy model for the multiphysics field of the motor, including: The root mean square error of the initial proxy model of the multiphysics field of the motor is determined; When the root mean square error does not meet the requirements, the model parameters are adjusted and iterative training is repeated based on the adjusted model parameters until the root mean square error meets the requirements, and the average error of the initial proxy model of the multiphysics field of the motor is judged. Based on the first set of sample points, and after processing with a preset periodic boundary three-dimensional flow-thermal coupling model, an initial proxy model for the multiphysics field of the motor is established, including: Based on the first sample point set, numerical simulation is performed using the preset periodic boundary three-dimensional flow thermal coupling model to obtain the first sample dataset. The initial proxy model of the motor multiphysics field was established using the first sample dataset; Specifically, when the root mean square error meets the requirements but the average error does not, secondary sampling is performed based on the on-site operating condition dataset to determine the target set of newly added sample points, including: When the root mean square error meets the requirements but the average error does not, the initial operating range dataset is determined based on the field operating condition dataset. Based on the initial on-site operating range dataset, the initial set of newly added sample points was determined using the empirical distribution function method. Based on the initial set of newly added sample points and the first set of sample points, the target set of newly added sample points is determined through the metric space method. The metric space method defines a distance function on a set such that the distance between any two elements in the set can be calculated using this distance function. Specifically, based on the initial on-site operating range dataset, and after processing using the empirical distribution function method, the initial set of newly added sample points is determined, including: Sort the data in the initial on-site operating range dataset to obtain the target on-site operating range dataset; Based on the aforementioned on-site target operating range dataset, multiple cumulative proportion values ​​are obtained through empirical distribution function calculation; The initial set of newly added sample points is determined based on the multiple cumulative ratio values; Specifically, determining the target set of new sample points based on the initial set of new sample points and the first set of sample points, after processing with the metric space method, includes: using the metric space method to calculate the distance between a new sample in the initial set of new sample points and any sample point in the first set of sample points; deleting a new sample in the initial set of new sample points when the calculated distance is less than a preset threshold, and determining the target set of new sample points based on the first set of sample points and the deleted initial set of new sample points.

2. The method according to claim 1, characterized in that, After sequentially determining the root mean square error and average error of the initial surrogate model of the multiphysics field of the motor, the method further includes: When the root mean square error and the average error meet the requirements, the initial surrogate model of the motor multiphysics field is determined as the target surrogate model of the motor multiphysics field.

3. A device for constructing a multiphysics proxy model for an electric motor, characterized in that, The device includes: The acquisition module is used to acquire the on-site operating condition dataset of the target motor. The on-site operating condition dataset includes inlet wind speed, air temperature, stator winding heat source, stator core heat source, stator winding average temperature, and stator core average temperature. The stator winding heat source is determined by copper loss and stray loss, and the stator core heat source is determined by hysteresis loss. The first sampling determination module is used to perform a sampling and determine the first sample point set in a preset sample three-dimensional space using the Latin hypercube sampling method. The preset sample three-dimensional space represents a global sample three-dimensional space for the multi-physics characteristics of the motor, composed of a data range set according to experience. The processing module is used to establish an initial proxy model of the multiphysics field of the motor based on the first sample point set and after processing by a preset periodic boundary three-dimensional flow thermal coupling model. The preset periodic boundary three-dimensional flow thermal coupling model is a periodic boundary three-dimensional flow thermal coupling model based on the stator winding and the iron core, which is used for numerical simulation for different operating conditions. The judgment module is used to sequentially judge the root mean square error and average error of the initial proxy model of the multiphysics field of the motor. The second sampling determination module is used to perform secondary sampling and determine the target new sample point set based on the field operation condition dataset when the root mean square error meets the requirements but the average error does not meet the requirements. The repeat module is used to add the target new sample point set to the first sample point set, and repeatedly iterate the model training according to the added first sample point set until the root mean square error meets the requirements and the average error meets the requirements and the motor multi-physics target proxy model is obtained. The judgment module includes: The first judgment submodule is used to judge the root mean square error of the initial proxy model of the multiphysics field of the motor. The adjustment submodule is used to adjust the model parameters and iterate the training repeatedly based on the adjusted model parameters when the root mean square error does not meet the requirements, until the root mean square error meets the requirements, and to judge the average error of the initial proxy model of the multiphysics field of the motor. The processing module includes: The simulation submodule is used to perform numerical simulation based on the first sample point set using the preset periodic boundary three-dimensional flow thermal coupling model to obtain the first sample dataset. A submodule is established to build the initial proxy model of the motor using the first sample dataset; The second sampling determination module includes: The determination submodule is used to determine the initial operating range dataset based on the field operating condition dataset when the root mean square error meets the requirements but the average error does not meet the requirements. The first processing submodule is used to determine the initial set of newly added sample points based on the initial on-site operating range dataset and processed by the empirical distribution function method. The second processing submodule is used to determine the target new sample point set based on the initial new sample point set and the first sample point set, through the metric space method. The metric space method defines a distance function on a set such that the distance between any two elements in the set can be calculated using this distance function. The first processing submodule includes: The sorting unit is used to sort the data in the initial on-site operating range dataset to obtain the target on-site operating range dataset. The calculation unit is used to calculate multiple cumulative ratio values ​​based on the on-site target operating range dataset and through an empirical distribution function; A determining unit is configured to determine the initial set of newly added sample points based on the plurality of cumulative ratio values; The second processing submodule is specifically used to: calculate the distance between a new sample in the initial set of newly added sample points and any sample point in the first set of sample points using the metric space method; when the calculated distance is less than a preset threshold, delete the new sample in the initial set of newly added sample points, and determine the target set of newly added sample points based on the first set of sample points and the deleted initial set of newly added sample points.

4. A computer device, characterized in that, include: The system includes a memory and a processor, which are interconnected. The memory stores computer instructions, and the processor executes these computer instructions to perform the method for constructing a multiphysics proxy model for an electric motor as described in claim 1 or 2.

5. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the method for constructing a multiphysics proxy model of an electric motor as described in claim 1 or 2.

6. A computer program product, characterized in that, Includes computer instructions for causing a computer to execute the method for constructing a multiphysics proxy model of an electric motor as described in claim 1 or 2.

Images