A method and system for optimizing motor torque control

Abstract

The present invention discloses a method and system for optimizing motor torque control, which relates to the field of motors. The method for optimizing motor torque control includes the following steps: S1, obtaining initial motor parameters, presetting an initial motor torque prediction scheme, and generating initial motor torque prediction parameters according to the motor torque prediction scheme; S2, presetting initial motor torque control rules according to the initial motor torque prediction parameters, and controlling the motor torque according to the initial motor torque control rules, obtaining initial motor torque actual parameters and motor usage parameters; S3, comparing the initial motor torque actual parameters with the initial motor torque prediction parameters, obtaining motor torque error values, and optimizing the initial motor torque prediction scheme according to the motor torque error values, and obtaining an advanced motor torque prediction scheme. The present invention presets the motor torque control rules by initializing the motor parameters, thereby realizing precise control of the motor torque.

Description

Technical Field

[0001] The present invention relates to the field of motors, and in particular to a method and system for optimizing motor torque control. Background Art

[0002] Motor torque, also known as torque, is one of the important performance parameters of the motor. It is the power output by the motor to the mechanical equipment through its axis, which is used to drive the equipment to work. The motor torque is related to the current and the design characteristics of the motor. When designing and selecting the motor, torque is an important parameter to consider. In practical applications, the torque of the motor will also affect the motor's starting performance, normal operating performance and overload capacity.

[0003] Motor torque control refers to controlling the motor's scheduling strategy and electrical parameters to adjust the motor torque and achieve stable operation. The current directly affects the magnetic field generated by the motor, which in turn affects the torque generation. In a DC motor, the motor's torque and current are linearly related. The motor torque can be controlled by adjusting the current. The change in voltage will also affect the motor's speed and torque, thereby changing the motor's speed and torque and improving the effect of the motor's use.

[0004] However, the existing methods and systems for optimizing motor torque control are unable to obtain the actual parameters of the motor in real time and provide feedback during use, resulting in the inability to flexibly control the motor torque, thereby making the precise control efficiency of the motor torque unsatisfactory. In addition, the existing methods and systems for optimizing motor torque control are unable to dynamically optimize the motor torque control method by real-time acquisition of motor torque parameters, comparison predictions, actual parameters and error values, resulting in unsatisfactory update efficiency of the methods and systems for optimizing motor torque control, and unable to perform corresponding optimization according to changes over time, resulting in unsatisfactory effects of the existing methods and systems for optimizing motor torque control when used.

[0005] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention provides a method and system for optimizing motor torque control, which has the advantage of real-time feedback update, thereby solving the problem of unsatisfactory optimization efficiency.

[0007] In order to achieve the above advantages of real-time feedback update, the specific technical solution adopted by the present invention is as follows:

[0008] According to one aspect of the present invention, a method for optimizing motor torque control comprises the following steps:

[0009] S1. Obtaining initial motor parameters, presetting an initial motor torque prediction scheme, and generating initial motor torque prediction parameters according to the motor torque prediction scheme;

[0010] S2. Preset an initial motor torque control rule according to the initial motor torque prediction parameter, and control the motor torque according to the initial motor torque control rule to obtain the initial motor torque actual parameter and the motor use parameter;

[0011] S3, comparing the initial motor torque actual parameters with the initial motor torque prediction parameters to obtain the motor torque error value, and optimizing the initial motor torque prediction scheme according to the motor torque error value to obtain an advanced motor torque prediction scheme;

[0012] S4, generating advanced motor torque prediction parameters according to the advanced motor torque prediction scheme and the motor usage parameters, and generating advanced motor torque control rules according to the advanced motor torque prediction parameters;

[0013] S5, presetting an advanced parameter update rule, and performing motor torque control according to the advanced motor torque control rule to obtain an optimized motor torque actual parameter;

[0014] S6, comparing the optimized motor torque actual parameters with the advanced motor torque prediction parameters, and updating and judging the comparison results between the optimized motor torque actual parameters and the advanced motor torque prediction parameters according to the advanced parameter update rule;

[0015] S7. Update the advanced motor torque prediction scheme according to the update judgment result, and generate an optimized motor torque control rule according to the updated advanced motor torque prediction scheme.

[0016] As a preferred solution, an initial motor torque control rule is preset according to the initial motor torque prediction parameter, and the motor torque is controlled according to the initial motor torque control rule. Acquiring the initial motor torque actual parameter and the motor usage parameter includes the following steps:

[0017] S21, performing data cleaning on the initial motor torque prediction parameters, and constructing a fluctuation diagram of the initial motor torque parameter changes;

[0018] S22, extracting the initial motor torque characteristic value according to the initial motor torque parameter change fluctuation diagram, and constructing the initial motor torque change parameter;

[0019] S23, generating an initial motor torque control rule according to the initial motor torque change parameter, and adjusting the motor torque according to the initial motor torque control rule;

[0020] S24, collecting motor parameters after motor torque adjustment in real time, and extracting initial motor torque actual parameters and motor usage parameters for feedback.

[0021] As a preferred solution, extracting the initial motor torque characteristic value according to the initial motor torque parameter change fluctuation diagram and constructing the initial motor torque change parameter includes the following steps:

[0022] S221, presetting a motor torque feature extraction rule, and extracting an initial motor torque feature value from the initial motor torque parameter change fluctuation diagram according to the motor torque feature extraction rule;

[0023] S222, constructing a motor torque frequency domain model according to the initial motor torque characteristic value;

[0024] S223, training the motor torque frequency domain model, and verifying the trained motor torque frequency domain model;

[0025] S224, optimizing according to the verification result of the motor torque frequency domain model, and calculating the initial motor torque change parameter according to the optimized motor torque frequency domain model.

[0026] As a preferred solution, constructing a motor torque frequency domain model according to the initial motor torque characteristic value includes the following steps:

[0027] S2221, performing Fourier transform on the initial motor torque characteristic value to obtain the motor torque frequency domain;

[0028] S2222. Construct a motor torque spectrum diagram according to the motor torque frequency domain, and obtain motor torque evaluation parameters;

[0029] S2223, presetting a model matching rule, and performing model matching on the motor torque evaluation parameter according to the model matching rule;

[0030] S2224. Construct a motor torque frequency domain model based on the model matching result.

[0031] As a preferred solution, the initial motor torque characteristic value is subjected to Fourier transform to obtain the calculation formula of the motor torque frequency domain:

[0032]

[0033] Wherein, D(m) is the motor torque data in the time domain of m;

[0034] D(y) is the motor torque data in the y frequency domain;

[0035] k is the frequency index value of the motor torque;

[0036] n is the time domain index value of the motor torque;

[0037] b is the number of samples of motor torque;

[0038] j is the imaginary unit of motor torque;

[0039] e is a natural constant.

[0040] As a preferred solution, the initial motor torque actual parameters are compared with the initial motor torque prediction parameters to obtain the motor torque error value, and the initial motor torque prediction scheme is optimized according to the motor torque error value to obtain the advanced motor torque prediction scheme, which includes the following steps:

[0041] S31, using a root mean square error algorithm to calculate a motor torque error value between an initial motor torque actual parameter and an initial motor torque prediction parameter;

[0042] S32, presetting an optimized characteristic parameter extraction rule, and extracting optimized characteristic parameters from the motor torque error value according to the optimized characteristic parameter extraction rule to obtain optimized characteristic parameters;

[0043] S33, adjusting parameters of the initial motor torque prediction scheme according to the optimized characteristic parameters, and verifying the adjusted initial motor torque prediction scheme;

[0044] S34, updating according to the verification result of the adjusted initial motor torque prediction scheme to obtain an advanced motor torque prediction scheme.

[0045] As a preferred solution, adjusting the parameters of the initial motor torque prediction scheme according to the optimized characteristic parameters and verifying the adjusted initial motor torque prediction scheme include the following steps:

[0046] S331, comparing the optimized characteristic parameters with the initial motor torque prediction scheme, and making parameter adjustment judgments based on the comparison results;

[0047] S332, adjusting the internal parameters of the initial motor torque prediction scheme accordingly according to the parameter adjustment judgment result;

[0048] S333, cross-validating the internal parameters of the adjusted initial motor torque prediction scheme, and outputting the cross-validated initial motor torque prediction scheme.

[0049] As a preferred solution, updating the adjusted initial motor torque prediction solution according to the verification result to obtain the advanced motor torque prediction solution includes the following steps:

[0050] S341, analyzing and judging the verification result of the adjusted initial motor torque prediction scheme, and obtaining the updated parameters of the motor torque prediction scheme;

[0051] S342, presetting a motor torque prediction scheme update strategy, and comparing the motor torque prediction scheme update parameters with the motor torque prediction scheme update strategy, and generating an update scheme according to the comparison result;

[0052] S343, updating the initial motor torque prediction scheme according to the update scheme, and evaluating the updated initial motor torque prediction scheme;

[0053] S344, performing feedback adjustment according to the updated evaluation result of the initial motor torque prediction scheme, and outputting the adjusted evaluation result of the initial motor torque prediction scheme as an advanced motor torque prediction scheme.

[0054] As a preferred solution, the actual parameters of the optimized motor torque are compared with the advanced motor torque prediction parameters, and the comparison result of the actual parameters of the optimized motor torque and the advanced motor torque prediction parameters is updated and judged according to the advanced parameter update rule, including the following steps:

[0055] S61, comparing the optimized motor torque actual parameter with the advanced motor torque prediction parameter by using a mean square error algorithm to obtain a motor torque difference measurement value;

[0056] S62, presetting a motor torque optimization update rule, and comparing the motor torque difference measurement value with the motor torque optimization update rule;

[0057] S63, performing an update judgment based on a comparison result between the motor torque difference measurement value and the motor torque optimization update rule, and performing feedback adjustment based on the update judgment result.

[0058] According to another aspect of the present invention, a system for optimizing motor torque control is provided, the system comprising:

[0059] A motor parameter acquisition module is used to acquire initial motor parameters, preset an initial motor torque prediction scheme, and generate initial motor torque prediction parameters according to the motor torque prediction scheme;

[0060] An initial torque prediction module is used to preset an initial motor torque control rule according to the initial motor torque prediction parameter, and control the motor torque according to the initial motor torque control rule to obtain the initial motor torque actual parameter and the motor use parameter;

[0061] The advanced torque prediction module is used to compare the initial motor torque actual parameters with the initial motor torque prediction parameters to obtain the motor torque error value, and optimize the initial motor torque prediction scheme according to the motor torque error value to obtain the advanced motor torque prediction scheme;

[0062] An advanced torque control module, used to generate advanced motor torque prediction parameters according to the advanced motor torque prediction scheme and motor usage parameters, and to generate advanced motor torque control rules according to the advanced motor torque prediction parameters;

[0063] The torque parameter optimization module is used to preset the advanced parameter update rules and perform motor torque control according to the advanced motor torque control rules to obtain the optimized motor torque actual parameters;

[0064] The optimized torque comparison module is used to compare the optimized motor torque actual parameters with the advanced motor torque prediction parameters, and to update and judge the comparison results between the optimized motor torque actual parameters and the advanced motor torque prediction parameters according to the advanced parameter update rules;

[0065] The optimized torque update module is used to update the advanced motor torque prediction scheme according to the update judgment result, and generate an optimized motor torque control rule according to the updated advanced motor torque prediction scheme.

[0066] Compared with the prior art, the present invention provides a method and system for optimizing motor torque control, which has the following beneficial effects:

[0067] (1) The present invention presets the motor torque control rules by initializing the motor parameters, and obtains the actual motor parameters in real time for feedback. It can be generally applied to various motor systems and can be flexibly controlled according to actual conditions, thereby realizing precise control of the motor torque. In addition, by real-time acquisition of motor torque parameters, comparison of predictions and actual parameters, generation of error values, and optimization of the prediction scheme, automated optimization is achieved in each step, thereby improving the effect of motor torque control optimization.

[0068] (2) The entire scheme of the present invention is designed with a complete set of feedback and update mechanisms, which compares the results for judgment and updates the prediction scheme and control rules according to the judgment results, so that the system can continuously learn and evolve to achieve better control effects. In addition, the frequency domain and error are accurately analyzed through advanced methods such as Fourier transform, frequency domain model, and root mean square error algorithm, so as to more accurately control the motor torque. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0070] Figure 1 is a method flow chart of a method for optimizing motor torque control according to an embodiment of the present invention;

[0071] Figure 2 4 is a system block diagram of a motor torque control optimization system according to an embodiment of the present invention.

[0072] In the figure:

[0073] 1. Motor parameter acquisition module; 2. Initial torque prediction module; 3. Advanced torque prediction module; 4. Advanced torque control module; 5. Optimized torque parameter module; 6. Optimized torque comparison module; 7. Optimized torque update module. DETAILED DESCRIPTION

[0074] The specific implementation of the present invention is further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0075] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0076] According to an embodiment of the present invention, a method and system for optimizing motor torque control are provided.

[0077] The present invention is further described with reference to the accompanying drawings and specific embodiments. According to one embodiment of the present invention, Figure 1 As shown, the motor torque control optimization method according to an embodiment of the present invention includes the following steps:

[0078] S1. Obtaining initial motor parameters, presetting an initial motor torque prediction scheme, and generating initial motor torque prediction parameters according to the motor torque prediction scheme;

[0079] Specifically, consult the motor specification sheet and record the basic parameters of the motor, such as motor type, rated power, rated speed, number of pole pairs, etc., and conduct a no-load test to record parameters such as the motor's no-load current and no-load speed to reflect the motor's mechanical loss and copper loss. Then conduct a load resistance test to apply different loads to the motor and record curves such as current, speed, and output torque to reflect the motor's magnetic field characteristics. Construct an initial mathematical model of the motor, such as a model based on a motor equivalent circuit, use typical model reference parameters, and make corrections based on the actual motor parameters to obtain an initial torque prediction model. Under given operating point conditions, calculate the current and speed based on the initial model, and then calculate the motor's initial torque prediction value based on the electromagnetic torque formula.

[0080] S2. Preset an initial motor torque control rule according to the initial motor torque prediction parameter, and control the motor torque according to the initial motor torque control rule to obtain the initial motor torque actual parameter and the motor use parameter;

[0081] Specifically, the initial motor torque control rule is preset according to the initial motor torque prediction parameter, and the motor torque is controlled according to the initial motor torque control rule, and the initial motor torque actual parameter and the motor use parameter are obtained, which includes the following steps:

[0082] S21, performing data cleaning on the initial motor torque prediction parameters, and constructing a fluctuation diagram of the initial motor torque parameter changes;

[0083] Specifically, collect the original data of the initial motor torque prediction parameters, including abnormal values, missing values ​​and other problems, pre-process the data, set the threshold, delete the abnormal data points beyond the threshold range, use the mean, median and other methods to fill the missing data, convert the data of different dimensions into the same dimension, such as converting the torque value into the standard deviation, analyze the overall trend and fluctuation of the data, and draw the curve of the parameter change over time, observe the overall upward and downward trend, and segment or cluster the data according to business needs, for example, divide the data within a cycle into several time periods, and then draw a parameter change fluctuation chart with time as the horizontal axis and the parameter value as the vertical axis, draw the change curve of the cleaned data, analyze the fluctuation chart, and identify the fluctuation pattern and characteristic values, such as fluctuation amplitude and periodicity.

[0084] S22, extracting the initial motor torque characteristic value according to the initial motor torque parameter change fluctuation diagram, and constructing the initial motor torque change parameter;

[0085] Specifically, extracting the initial motor torque characteristic value according to the initial motor torque parameter change fluctuation diagram and constructing the initial motor torque change parameter includes the following steps:

[0086] S221, presetting a motor torque feature extraction rule, and extracting an initial motor torque feature value from the initial motor torque parameter change fluctuation diagram according to the motor torque feature extraction rule;

[0087] Specifically, preset motor torque feature extraction rules are set according to the physical meaning of motor torque parameters, data characteristics, application scenarios, algorithm feasibility and correlation between features. For example, according to the different physical meanings represented by different parameters, it is necessary to distinguish which parameters can better reflect the motor operating status, such as torque, speed, etc., and according to the characteristics of the parameter data itself, such as whether it changes periodically and the amplitude of fluctuations, appropriate statistical features and morphological features are designed. According to the feature values ​​that will be used for downstream tasks such as monitoring and prediction, features that can better serve these application scenarios are designed. Considering the feasibility of the feature extraction algorithm, features with low computational complexity but strong information expression capabilities are selected to avoid designing repeated features with too high correlation and reduce feature dimensions. At the same time, it is necessary to consider the extraction of numerical and categorical features, as well as time and frequency domain features.

[0088] When extracting feature values ​​from the fluctuation graph, calculate statistical features such as the mean and standard deviation, use the peak detection algorithm to find the peak points, use the valley detection algorithm to find the valley points, apply the automatic recognition algorithm to identify the waveform type, and use the period detection algorithm to identify the time length of a complete period. Perform Fourier decomposition on the waveform, extract the energy of each frequency component as a feature, organize and save all the extracted feature values ​​to form a motor torque feature value library, and the feature value library can be applied to downstream tasks such as motor status monitoring and parameter prediction. Appropriate features need to be selected based on the actual application scenario.

[0089] S222, constructing a motor torque frequency domain model according to the initial motor torque characteristic value;

[0090] Specifically, constructing a motor torque frequency domain model according to the initial motor torque characteristic value includes the following steps:

[0091] S2221, performing Fourier transform on the initial motor torque characteristic value to obtain the motor torque frequency domain;

[0092] Specifically, the initial motor torque characteristic value is Fourier transformed to obtain the calculation formula of the motor torque frequency domain:

[0093]

[0094] Wherein, D(m) is the motor torque data in the time domain of m;

[0095] D(y) is the motor torque data in the y frequency domain;

[0096] k is the frequency index value of the motor torque;

[0097] n is the time domain index value of the motor torque;

[0098] b is the number of samples of motor torque;

[0099] j is the imaginary unit of motor torque;

[0100] e is a natural constant.

[0101] S2222. Construct a motor torque spectrum diagram according to the motor torque frequency domain, and obtain motor torque evaluation parameters;

[0102] Specifically, the frequency is the horizontal axis and the power spectrum density is the vertical axis, and then the characteristic frequency parameters are extracted from the spectrum diagram, such as the peak frequency, the frequency corresponding to the peak point of the power spectrum density, the energy center frequency, the frequency of the total energy center of the signal, the bandwidth, and the frequency range in which the power spectrum density exceeds a certain threshold.

[0103] S2223, presetting a model matching rule, and performing model matching on the motor torque evaluation parameter according to the model matching rule;

[0104] Specifically, preset model matching rules, and preset sinusoidal models, exponential models, power-law models and polynomial models, etc. According to the rules, the frequency parameters are normalized, and then the normalized parameters are fit to each preset model using the least squares method, and the model with the smallest fitting error is selected as the matching model. An error threshold can also be set, and models with errors below the threshold are considered to be matched, and the successfully matched model and its parameters are used as the motor torque characteristics.

[0105] S2224. Construct a motor torque frequency domain model based on the model matching result.

[0106] Specifically, the model matching results are collected, including each successfully matched frequency parameter and its corresponding best matching model name and model parameter value, and a frequency parameter model library table is constructed to record the model information corresponding to each frequency parameter. The matching results are sorted out, and the frequency parameters are analyzed to see whether they match different models in different time periods. According to the matching rules, corresponding models are constructed for each frequency parameter in different time periods. For example, a certain parameter matches a sinusoidal model during a period, and an exponential model during other periods, or a certain parameter matches only one model throughout the entire process, such as a polynomial model. Each constructed frequency parameter-time period-model relationship is recorded in the frequency domain model library, and can also be integrated into a multivariate frequency domain model. The frequency domain model library is used for tasks such as motor operation status identification and parameter prediction.

[0107] S223, training the motor torque frequency domain model, and verifying the trained motor torque frequency domain model;

[0108] Specifically, a large amount of sampling data of motor torque frequency parameters and time is collected as a data set for training and verification, and the data set is randomly divided into a training data set and a test data set. According to the constructed frequency domain model structure, such as a multivariable function model, the model parameters are trained using the training data set, and the linear model parameters are trained using the least squares method. After the training, a frequency domain model containing all parameter values ​​is obtained, and the test data set is input into the trained model for prediction. The error between the prediction result and the true value is calculated, and a suitable error function is selected. The degree of conformity between the prediction curve and the true curve is observed, the prediction effect is analyzed, the model structure or hyperparameters are adjusted, the training and verification are repeated, and the model with the best effect is selected to obtain a frequency domain model that has been verified to be effective under a given data set.

[0109] S224, optimizing according to the verification result of the motor torque frequency domain model, and calculating the initial motor torque change parameter according to the optimized motor torque frequency domain model.

[0110] Specifically, analyze the error sources of the frequency domain model during training and verification, identify the deficiencies of the model, and optimize the model structure according to the error analysis results, such as increasing the input features or the number of hidden layer nodes, enhancing the model representation ability, selecting a better activation function or loss function, adding regularization terms to prevent overfitting, repeating the training and verification process, and observing whether the optimized model can effectively reduce the error. Multiple optimization iterations are required, and the best model is selected as the final frequency domain model. The final optimized model is applied to actual data, and the model is used to calculate the change trend prediction curve of each frequency parameter in a certain period of time in the future and the predicted values ​​of the characteristic values ​​such as the peak value and average value of the parameter under different conditions. Compare with the actual monitoring data, evaluate the accuracy of the calculation results, and further improve and update the model according to the error analysis. A frequency domain model for practical application is obtained.

[0111] S23, generating an initial motor torque control rule according to the initial motor torque change parameter, and adjusting the motor torque according to the initial motor torque control rule;

[0112] Specifically, the characteristics of the initial motor torque change parameters are analyzed, the key influencing factors and change laws are identified, and the initial value control strategy is designed based on the influencing factors and change laws. For example, the torque upper limit is dynamically adjusted according to the load size, and the torque output time is appropriately advanced or delayed according to the temperature change. According to the motor operation history, an adaptive adjustment strategy is adopted, and the control strategy is divided into several initial value control rules. The control rules are encoded into a format that can be recognized and called by the motor control system. In the motor control system, the corresponding control rules are called according to the real-time conditions, and the rule calling results output the corresponding torque control instructions. The motor drive system adjusts the motor torque according to the control instructions, monitors the adjustment effect, and optimizes the control rules based on error feedback.

[0113] S24, collecting motor parameters after motor torque adjustment in real time, and extracting initial motor torque actual parameters and motor usage parameters for feedback.

[0114] Specifically, a parameter acquisition module is set in the motor control system to collect operating parameters such as torque and speed from the motor sensor in real time, pre-process the collected parameter data, ensure sampling synchronization, process multi-channel data with different sampling rates, compress the data, extract features to replace the original data transmission, and extract the initial motor torque actual parameters, such as the adjusted torque peak and average value, and motor usage parameters, such as load size, cumulative running time, from the collected data according to preset rules. The extracted actual parameters and usage parameters are classified and stored, and some parameters are selected as needed and fed back to the control system in the form of messages or files. The control system receives feedback data to evaluate the control effect, optimize the control rules, and improve the motor usage model, provide a usage reference, identify motor state changes, and predict maintenance time.

[0115] S3, comparing the initial motor torque actual parameters with the initial motor torque prediction parameters to obtain the motor torque error value, and optimizing the initial motor torque prediction scheme according to the motor torque error value to obtain an advanced motor torque prediction scheme;

[0116] Specifically, the initial motor torque actual parameters are compared with the initial motor torque prediction parameters to obtain the motor torque error value, and the initial motor torque prediction scheme is optimized according to the motor torque error value to obtain the advanced motor torque prediction scheme, which includes the following steps:

[0117] S31, using a root mean square error algorithm to calculate a motor torque error value between an initial motor torque actual parameter and an initial motor torque prediction parameter;

[0118] Specifically, a set of actual parameter values ​​of the motor torque and a set of predicted parameter values ​​of the motor torque are collected, the error between each actual parameter value and the corresponding predicted parameter value is calculated, the square of each error and the sum of all squares of errors are calculated, and then the ratio of the sum of squares of errors to the number of samples is calculated to obtain the error value between the actual parameter and the predicted parameter.

[0119] S32, presetting an optimized characteristic parameter extraction rule, and extracting optimized characteristic parameters from the motor torque error value according to the optimized characteristic parameter extraction rule to obtain optimized characteristic parameters;

[0120] Specifically, preset optimization feature parameter extraction rules, for example, the average error represents the overall prediction accuracy, the standard deviation represents the degree of error distribution, the maximum error predicts the worst data point, the minimum error predicts the best data point, the error change trend, the error change pattern over time, the error distribution type, normal distribution, uniform distribution, etc. According to the rules, the motor torque error value is subjected to feature extraction calculation. In addition to statistical features, other features can also be extracted, such as extracting the error fluctuation period as a feature, performing Fourier analysis on the error, extracting frequency information, and using all extracted feature values ​​as optimized feature parameters. The feature parameters can be used for model optimization and can also be used as quality control indicators.

[0121] S33, adjusting parameters of the initial motor torque prediction scheme according to the optimized characteristic parameters, and verifying the adjusted initial motor torque prediction scheme;

[0122] Specifically, adjusting the parameters of the initial motor torque prediction scheme according to the optimized characteristic parameters and verifying the adjusted initial motor torque prediction scheme includes the following steps:

[0123] S331, comparing the optimized characteristic parameters with the initial motor torque prediction scheme, and making parameter adjustment judgments based on the comparison results;

[0124] Specifically, collect the optimized feature parameters and related feature parameters produced by the prediction scheme, match the two sets of parameters according to the feature type, such as average error corresponding to average error, etc., quantitatively calculate the degree of difference between each corresponding feature parameter, and judge the pros and cons of the prediction scheme according to the degree of difference. For example, if the difference is small, the prediction scheme is effective; if the difference in a certain feature is large, it indicates that the prediction in this aspect is inaccurate. Analyze the potential causes of the difference, such as improper selection or setting of the prediction model, imperfect feature extraction rules, and training data quality problems. According to the cause analysis, adjust the prediction scheme, such as modifying or replacing the prediction model, optimizing the feature extraction rules, adding more training data, adjusting the model hyperparameters, etc., retrain the prediction, observe the adjustment effect, and improve the matching degree between the prediction features and the optimized features through iterative comparison and adjustment.

[0125] S332, adjusting the internal parameters of the initial motor torque prediction scheme accordingly according to the parameter adjustment judgment result;

[0126] Specifically, analyze the parameter adjustment judgment results, determine the internal modules such as the prediction model or algorithm that need to be adjusted, and take corresponding adjustment methods according to different internal modules. If the prediction model needs to be adjusted, you can adjust the number of layers and nodes in the model structure, the learning rate and regularization term strength in the model parameters, the optimization algorithm and loss function in the training algorithm, etc. If feature extraction needs to be adjusted, you can adjust the feature selection rules, feature extraction algorithm, and feature dimensions. If data processing needs to be adjusted, you can adjust the data cleaning rules and thresholds, data normalization methods, and data sampling strategies. For each internal parameter that needs to be adjusted, set an adjustment range or value set. Through the grid method, traverse the adjustment parameter value combinations, retrain the prediction model for each group of adjustment parameters, evaluate the effect, select the adjustment parameter combination with the best effect, and update the internal parameters of the prediction scheme.

[0127] S333, cross-validating the internal parameters of the adjusted initial motor torque prediction scheme, and outputting the cross-validated initial motor torque prediction scheme.

[0128] Specifically, the complete motor torque data is collected as a sample set, and the sample set is randomly divided into several non-overlapping subsets, one subset is set as a validation set, and the rest are used as training sets. The training set is used to train the prediction model to obtain a first model, and the first model is used to make predictions on the validation set, and the prediction error is calculated. The validation set and the training set are exchanged to obtain a second model, and the error of the second model is calculated. This is repeated several times so that each subset is used as a validation set, and the average error is calculated. The set of internal parameters that minimize the average error is selected as the optimal parameters, and an initial motor torque prediction scheme with the optimal parameters is output.

[0129] S34, updating according to the verification result of the adjusted initial motor torque prediction scheme to obtain an advanced motor torque prediction scheme.

[0130] Specifically, updating the adjusted initial motor torque prediction scheme according to the verification result to obtain the advanced motor torque prediction scheme includes the following steps:

[0131] S341, analyzing and judging the verification result of the adjusted initial motor torque prediction scheme, and obtaining the updated parameters of the motor torque prediction scheme;

[0132] Specifically, collect various evaluation indicators of the prediction scheme during the cross-validation process, such as average error, variance, etc., analyze the indicator results, and determine whether the prediction effect meets the requirements. If the average error is too large, the prediction effect is not ideal and needs to be further optimized. If each indicator is within the allowable range, it means that the prediction effect has met the requirements. If further optimization is needed, analyze the causes of the error, such as whether the prediction model itself needs to be improved, whether the feature selection is reasonable, and whether the hyperparameter setting is insufficiently optimized. According to the error cause analysis, determine the types of parameters that need to be updated, such as model parameters, feature selection rules, hyperparameter settings, and evaluation indicator calculation methods. Feedback these internal parameter information that need to be updated to the prediction scheme and continue to optimize the prediction scheme based on the feedback until the requirements are met.

[0133] S342, presetting a motor torque prediction scheme update strategy, and comparing the motor torque prediction scheme update parameters with the motor torque prediction scheme update strategy, and generating an update scheme according to the comparison result;

[0134] Specifically, different prediction scheme update strategies are preset, for example, model update strategy, reselect model or adjust model structure, parameter update strategy, adjust model parameters or hyperparameters, feature update strategy, adjust feature selection rules or extraction methods, data update strategy, collect new data or optimize data quality, match and compare the prediction scheme parameters that need to be updated with the preset update strategies, select the update strategy with the highest matching degree, formulate specific update operations according to the matching update strategy, if it is a model update strategy, select a new model type, if it is a parameter update strategy, select the parameter value range that needs to be adjusted, if it is a feature update strategy, modify the feature selection rules, etc., perform update operations, such as reselecting models, adjusting parameters or collecting new data, repeating the training and verification process, evaluating the update effect, if the effect is not satisfactory, continue to adjust the update operation or select a new strategy, and finally generate an update scheme that can effectively improve the prediction accuracy.

[0135] S343, updating the initial motor torque prediction scheme according to the update scheme, and evaluating the updated initial motor torque prediction scheme;

[0136] Specifically, according to the determined update plan, the prediction plan is updated accordingly. If it is a model update, a new model type is selected and trained. If it is a parameter update, the corresponding parameter value is adjusted and retrained. If it is a feature update, the feature selection rules are modified and new features are extracted. If it is a data update, new data is collected and the model is retrained. After training the updated prediction plan model, the new data of the actual working conditions is input into the prediction plan for prediction. The prediction results are compared with the actual values. New evaluation indicators are calculated, including mean absolute error, root mean square error, determination coefficient, and performance curve. Whether the indicators are better than the original plan is analyzed to determine the update effect. If the indicators drop, it indicates that the update is unsuccessful and the update plan needs to be redesigned. If the indicators increase, it indicates that the update is successful and can be put into use. The update plan is continuously optimized according to the evaluation results to complete the iterative update of the prediction model.

[0137] S344, performing feedback adjustment according to the updated evaluation result of the initial motor torque prediction scheme, and outputting the adjusted evaluation result of the initial motor torque prediction scheme as an advanced motor torque prediction scheme.

[0138] Specifically, analyze the evaluation results of the prediction scheme to determine whether it meets the accuracy requirements. If not, analyze the reasons based on the differences in the evaluation indicators, determine what needs to be adjusted, whether the model structure needs to be optimized, whether the features need to be improved and extracted, whether the prediction algorithm needs to be improved, whether the evaluation method is scientific, etc. According to the reasons for the adjustment, design a new adjustment plan, make a second adjustment and update the prediction plan, repeat the training and verification, and evaluate the adjustment effect. If satisfied, output the adjusted prediction plan as an advanced prediction plan. If not satisfied, continue to adjust and optimize, repeat the above process, and through continuous adjustment and optimization, steadily improve the prediction accuracy, and the prediction plan will continue to advance with iterations, approaching the theoretical limit level, and output the prediction plan with the best effect in the current stage as the advanced version.

[0139] S4, generating advanced motor torque prediction parameters according to the advanced motor torque prediction scheme and the motor usage parameters, and generating advanced motor torque control rules according to the advanced motor torque prediction parameters;

[0140] Specifically, the motor usage parameters such as load and temperature collected in real time are input into the advanced prediction scheme. The advanced prediction scheme calculates the torque prediction parameters at different time points in the future, such as peak value and average value, based on the usage parameters and internal models. These prediction parameters are organized into an advanced motor torque prediction parameter table. According to the characteristics of the prediction parameters, such as whether there is a periodic change pattern, etc., a new control strategy is designed. The control strategy is converted into several advanced motor torque control rules. For example, the rules may be to dynamically adjust the upper and lower limits of the torque according to the load size, output the torque command in advance according to the temperature prediction, etc. The control rules are encoded into a format recognized by the motor control system. The motor control system calls the corresponding rules according to the real-time conditions and outputs the control command. Through continuous optimization, the control effect is improved.

[0141] S5, presetting an advanced parameter update rule, and performing motor torque control according to the advanced motor torque control rule to obtain an optimized motor torque actual parameter;

[0142] Specifically, advanced parameter update rules are preset, such as dynamically adjusting control rule parameters according to control effects, timely revising control strategies according to environmental changes, continuously improving prediction models according to new data, the motor control system calls advanced control rules according to real-time conditions, outputs control instructions, the motor drive system executes control instructions, adjusts motor torque, monitors motor operating parameters in real time, compares control effects, extracts optimized characteristic parameters based on monitoring data and update rules, adjusts the error between the actual torque value and the predicted value, torque fluctuation characteristic parameters at different stages, load response speed, etc., analyzes characteristic parameters, judges control quality and prediction accuracy, and takes update operations based on the analysis results, revises control rules, improves prediction models, optimizes update rules, repeats control to perform monitoring and update cycles, continuously optimizes motor torque quality, and outputs the optimized actual motor torque parameters.

[0143] S6, comparing the optimized motor torque actual parameters with the advanced motor torque prediction parameters, and updating and judging the comparison results between the optimized motor torque actual parameters and the advanced motor torque prediction parameters according to the advanced parameter update rule;

[0144] Specifically, comparing the optimized motor torque actual parameters with the advanced motor torque prediction parameters, and updating the comparison results of the optimized motor torque actual parameters and the advanced motor torque prediction parameters according to the advanced parameter update rule includes the following steps:

[0145] S61, comparing the optimized motor torque actual parameter with the advanced motor torque prediction parameter by using a mean square error algorithm to obtain a motor torque difference measurement value;

[0146] Specifically, collect the optimized actual motor torque parameter set, collect the advanced prediction parameter set, calculate the error between each actual parameter and the predicted parameter, calculate the square of each error, calculate the sum of all squared errors, calculate the ratio of the sum of squared errors to the number of samples, and then obtain the difference metric value.

[0147] S62, presetting a motor torque optimization update rule, and comparing the motor torque difference measurement value with the motor torque optimization update rule;

[0148] Specifically, different motor torque optimization update rules are preset, for example, model update rules, adjust the prediction model structure or algorithm, parameter update rules, optimize model parameters or hyperparameters, feature update rules, improve feature selection or extraction methods, data update rules, collect more training data, collect motor torque difference measurement values, match and compare the difference measurement values ​​with each update rule, if the difference measurement value is large, it may match the model or parameter update rule, if some feature errors are large, it may match the feature update rule, if the overall error fluctuates greatly, it may match the data update rule, select the update rule with the highest matching degree, formulate specific optimization operation plans according to the matching rules, such as adjusting the model structure, optimizing the learning rate, etc., perform optimization operations, repeat training and verification, and through iterative optimization, gradually reduce the difference until satisfaction is achieved.

[0149] S63, performing an update judgment based on a comparison result between the motor torque difference measurement value and the motor torque optimization update rule, and performing feedback adjustment based on the update judgment result.

[0150] Specifically, analyze the difference metric and matching rules, and judge the current prediction effect. If the difference metric is within the allowable range, it indicates that the prediction effect has met the requirements and no update is required. If the difference metric is greater than the threshold, it is necessary to continue to optimize and make an update judgment to determine whether the matching update rules need to be modified. Analyze the main reasons for the difference and determine what needs to be adjusted, such as models and parameters. According to the update judgment results, design a feedback adjustment plan, modify the update rules, redesign the optimization operation plan, adjust the evaluation index calculation method, and perform feedback adjustment, such as selecting a new model or adjusting the learning rate range. Repeat training and verification, and observe the adjustment effect. If the RMSE decreases after adjustment, it indicates that the update is successful. Otherwise, continue to adjust. Through continuous evaluation, judgment and feedback adjustment, the prediction accuracy is steadily improved. Repeat this process until the model effect is satisfactory.

[0151] S7. Update the advanced motor torque prediction scheme according to the update judgment result, and generate an optimized motor torque control rule according to the updated advanced motor torque prediction scheme.

[0152] Specifically, according to the update judgment result, determine the content that needs to be adjusted, such as model, parameters, etc., design specific prediction scheme update operations, such as selecting a new model, adjusting the learning rate, etc., execute the prediction scheme update operation, retrain the prediction model, collect a new batch of actual motor operation data as a verification set, input the verification set into the updated prediction model for prediction, calculate a new round of prediction error, judge the update effect, if the error decreases, it indicates that the update is successful, otherwise continue to adjust, define the updated prediction model as an optimized advanced prediction scheme, design a new control strategy based on the prediction parameters output by the optimized prediction scheme, use the control strategy as an optimized motor torque control rule, encode the control rule into a callable format for the motor control system, and through continuous optimization, achieve further improvement in control effect.

[0153] According to another embodiment of the present invention, Figure 2 As shown, a system for optimizing motor torque control is provided, the system comprising:

[0154] The motor parameter acquisition module 1 is used to acquire the initial motor parameters, preset the initial motor torque prediction scheme, and generate the initial motor torque prediction parameters according to the motor torque prediction scheme;

[0155] The initial torque prediction module 2 is used to preset the initial motor torque control rule according to the initial motor torque prediction parameter, and control the motor torque according to the initial motor torque control rule to obtain the initial motor torque actual parameter and the motor use parameter;

[0156] The advanced torque prediction module 3 is used to compare the initial motor torque actual parameters with the initial motor torque prediction parameters to obtain the motor torque error value, and optimize the initial motor torque prediction scheme according to the motor torque error value to obtain the advanced motor torque prediction scheme;

[0157] An advanced torque control module 4, used to generate advanced motor torque prediction parameters according to the advanced motor torque prediction scheme and motor usage parameters, and to generate advanced motor torque control rules according to the advanced motor torque prediction parameters;

[0158] The torque parameter optimization module 5 is used to preset the advanced parameter update rule, and perform motor torque control according to the advanced motor torque control rule to obtain the optimized motor torque actual parameter;

[0159] The optimized torque comparison module 6 is used to compare the optimized motor torque actual parameters with the advanced motor torque prediction parameters, and to update and judge the comparison result between the optimized motor torque actual parameters and the advanced motor torque prediction parameters according to the advanced parameter update rule;

[0160] The optimized torque update module 7 is used to update the advanced motor torque prediction scheme according to the update judgment result, and generate an optimized motor torque control rule according to the updated advanced motor torque prediction scheme.

[0161] To sum up, with the help of the above-mentioned technical scheme of the present invention, the present invention presets the motor torque control rules by initializing the motor parameters, and obtains the actual parameters of the motor in real time for feedback. It can be generally used in various motor systems and can be flexibly controlled according to actual conditions, thereby realizing precise control of the motor torque. Moreover, by real-time acquisition of motor torque parameters, comparison of predictions and actual parameters, generation of error values, and optimization of the prediction scheme accordingly, automated optimization is achieved in each step, thereby improving the effect of motor torque control optimization.

[0162] In addition, the entire scheme of the present invention has designed a complete feedback and update mechanism to make judgments based on the comparison results, and update the prediction scheme and control rules based on the judgment results, so that the system can continuously learn and evolve to achieve better control effects, and through advanced methods such as Fourier transform, frequency domain model, root mean square error algorithm, etc., the frequency domain and error are accurately analyzed, thereby more accurately controlling the motor torque.

[0163] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for optimizing motor torque control, characterized in that: The following steps are involved: S1. Obtaining initial motor parameters, presetting an initial motor torque prediction scheme, and generating initial motor torque prediction parameters according to the motor torque prediction scheme; S2. Preset an initial motor torque control rule according to the initial motor torque prediction parameter, and control the motor torque according to the initial motor torque control rule to obtain the initial motor torque actual parameter and the motor use parameter; S3, comparing the initial motor torque actual parameters with the initial motor torque prediction parameters to obtain the motor torque error value, and optimizing the initial motor torque prediction scheme according to the motor torque error value to obtain an advanced motor torque prediction scheme; S4, generating advanced motor torque prediction parameters according to the advanced motor torque prediction scheme and the motor usage parameters, and generating advanced motor torque control rules according to the advanced motor torque prediction parameters; S5, presetting an advanced parameter update rule, and performing motor torque control according to the advanced motor torque control rule to obtain an optimized motor torque actual parameter; S6, comparing the optimized motor torque actual parameters with the advanced motor torque prediction parameters, and updating and judging the comparison results between the optimized motor torque actual parameters and the advanced motor torque prediction parameters according to the advanced parameter update rule; S7, updating the advanced motor torque prediction scheme according to the update judgment result, and generating an optimized motor torque control rule according to the updated advanced motor torque prediction scheme; wherein, the initial motor torque control rule is preset according to the initial motor torque prediction parameter, and the motor torque is controlled according to the initial motor torque control rule, and obtaining the initial motor torque actual parameter and the motor use parameter includes the following steps: S21, performing data cleaning on the initial motor torque prediction parameters, and constructing a fluctuation diagram of the initial motor torque parameter changes; S22, extracting the initial motor torque characteristic value according to the initial motor torque parameter change fluctuation diagram, and constructing the initial motor torque change parameter; S23, generating an initial motor torque control rule according to the initial motor torque change parameter, and adjusting the motor torque according to the initial motor torque control rule; S24, collecting motor parameters after motor torque adjustment in real time, and extracting initial motor torque actual parameters and motor usage parameters for feedback; The extracting of the initial motor torque characteristic value according to the initial motor torque parameter change fluctuation diagram and constructing the initial motor torque change parameter comprises the following steps: S221, presetting a motor torque feature extraction rule, and extracting an initial motor torque feature value from the initial motor torque parameter change fluctuation diagram according to the motor torque feature extraction rule; S222, constructing a motor torque frequency domain model according to the initial motor torque characteristic value; S223, training the motor torque frequency domain model, and verifying the trained motor torque frequency domain model; S224, optimizing according to the verification result of the motor torque frequency domain model, and calculating the initial motor torque change parameter according to the optimized motor torque frequency domain model.

2. The method for optimizing motor torque control according to claim 1, characterized in that: The step of constructing the motor torque frequency domain model according to the initial motor torque characteristic value comprises the following steps: S2221, performing Fourier transform on the initial motor torque characteristic value to obtain the motor torque frequency domain; S2222. Construct a motor torque spectrum diagram according to the motor torque frequency domain, and obtain motor torque evaluation parameters; S2223, presetting a model matching rule, and performing model matching on the motor torque evaluation parameter according to the model matching rule; S2224. Construct a motor torque frequency domain model based on the model matching result.

3. The method for optimizing motor torque control according to claim 2, characterized in that: The calculation formula for obtaining the motor torque frequency domain by Fourier transforming the initial motor torque characteristic value is: Wherein, D(m) is the motor torque data in the time domain of m; D(y) is the motor torque data in the y frequency domain; k is the frequency index value of the motor torque; n is the time domain index value of the motor torque; b is the number of samples of motor torque; j is the imaginary unit of motor torque; e is a natural constant.

4. The method for optimizing motor torque control according to claim 1, characterized in that: The method of comparing the initial motor torque actual parameters with the initial motor torque prediction parameters to obtain the motor torque error value, and optimizing the initial motor torque prediction scheme according to the motor torque error value to obtain the advanced motor torque prediction scheme includes the following steps: S31, using a root mean square error algorithm to calculate a motor torque error value between an initial motor torque actual parameter and an initial motor torque prediction parameter; S32, presetting an optimized characteristic parameter extraction rule, and extracting optimized characteristic parameters from the motor torque error value according to the optimized characteristic parameter extraction rule to obtain optimized characteristic parameters; S33, adjusting parameters of the initial motor torque prediction scheme according to the optimized characteristic parameters, and verifying the adjusted initial motor torque prediction scheme; S34, updating according to the verification result of the adjusted initial motor torque prediction scheme to obtain an advanced motor torque prediction scheme.

5. The method for optimizing motor torque control according to claim 4, characterized in that: The step of adjusting the parameters of the initial motor torque prediction scheme according to the optimized characteristic parameters and verifying the adjusted initial motor torque prediction scheme comprises the following steps: S331, comparing the optimized characteristic parameters with the initial motor torque prediction scheme, and making parameter adjustment judgments based on the comparison results; S332, adjusting the internal parameters of the initial motor torque prediction scheme accordingly according to the parameter adjustment judgment result; S333, cross-validating the internal parameters of the adjusted initial motor torque prediction scheme, and outputting the cross-validated initial motor torque prediction scheme.

6. The method for optimizing motor torque control according to claim 4, characterized in that: The updating according to the verification result of the adjusted initial motor torque prediction scheme to obtain the advanced motor torque prediction scheme comprises the following steps: S341, analyzing and judging the verification result of the adjusted initial motor torque prediction scheme, and obtaining the updated parameters of the motor torque prediction scheme; S342, presetting a motor torque prediction scheme update strategy, and comparing the motor torque prediction scheme update parameters with the motor torque prediction scheme update strategy, and generating an update scheme according to the comparison result; S343, updating the initial motor torque prediction scheme according to the update scheme, and evaluating the updated initial motor torque prediction scheme; S344, performing feedback adjustment according to the updated evaluation result of the initial motor torque prediction scheme, and outputting the adjusted evaluation result of the initial motor torque prediction scheme as an advanced motor torque prediction scheme.

7. The method for optimizing motor torque control according to claim 1, characterized in that: The step of comparing the optimized motor torque actual parameter with the advanced motor torque prediction parameter and updating the comparison result between the optimized motor torque actual parameter and the advanced motor torque prediction parameter according to the advanced parameter update rule comprises the following steps: S61, comparing the optimized motor torque actual parameter with the advanced motor torque prediction parameter by using a mean square error algorithm to obtain a motor torque difference measurement value; S62, presetting a motor torque optimization update rule, and comparing the motor torque difference measurement value with the motor torque optimization update rule; S63, performing an update judgment based on a comparison result between the motor torque difference measurement value and the motor torque optimization update rule, and performing feedback adjustment based on the update judgment result.

8. A motor torque control optimization system, used to implement the motor torque control optimization method according to any one of claims 1 to 7, characterized in that: The system includes: A motor parameter acquisition module (1) is used to acquire initial motor parameters, preset an initial motor torque prediction scheme, and generate initial motor torque prediction parameters according to the motor torque prediction scheme; An initial torque prediction module (2) is used to preset an initial motor torque control rule according to the initial motor torque prediction parameter, and control the motor torque according to the initial motor torque control rule to obtain the initial motor torque actual parameter and the motor use parameter; An advanced torque prediction module (3) is used to compare the initial motor torque actual parameters with the initial motor torque prediction parameters to obtain a motor torque error value, and optimize the initial motor torque prediction scheme according to the motor torque error value to obtain an advanced motor torque prediction scheme; An advanced torque control module (4), used for generating advanced motor torque prediction parameters according to the advanced motor torque prediction scheme and motor usage parameters, and generating advanced motor torque control rules according to the advanced motor torque prediction parameters; An optimized torque parameter module (5) is used to preset an advanced parameter update rule and perform motor torque control according to the advanced motor torque control rule to obtain an optimized motor torque actual parameter; An optimized torque comparison module (6) is used to compare the optimized motor torque actual parameter with the advanced motor torque prediction parameter, and to update and judge the comparison result between the optimized motor torque actual parameter and the advanced motor torque prediction parameter according to the advanced parameter update rule; The optimized torque update module (7) is used to update the advanced motor torque prediction scheme according to the update judgment result, and generate an optimized motor torque control rule according to the updated advanced motor torque prediction scheme.

Images