A high-order sliding mode compound control method for speed regulation of an electric forklift motor

Abstract

The application relates to the technical field of motor control, and provides a high-order sliding mode composite control method for motor speed regulation of an electric forklift, which comprises the following steps: control period signals in motor operation state data are selected to generate control parameter samples; the samples are sent to a sliding mode control model main controller module pre-deployed in a local controller to extract motor speed regulation characteristic data; the data are sent to a parameter adjustment module of a central processor for dynamic compensation to generate regulation and control instructions; and a local controller execution unit responds to the instructions and outputs. The method realizes dynamic optimization of speed regulation parameters through cooperation of the main controller and the central processor, wherein state quantity extraction covers current, torque, rotating speed and other characteristics to realize multi-parameter fusion control. The application can improve the response speed, anti-interference capability and control precision of motor speed regulation of the electric forklift and optimize operation stability under multiple working conditions.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of motor control, and more particularly, to a high-order sliding mode composite control method for motor speed regulation of an electric forklift. BACKGROUND

[0002] In modern industrial production, as an important material handling equipment, the motor speed regulation performance of an electric forklift directly affects the work efficiency and operation accuracy. Traditional motor speed regulation control methods mainly include simple PID control, vector control, etc. These methods can achieve motor speed regulation to some extent, but when facing complex working conditions and dynamic disturbances, they often show slow response speed, low speed regulation accuracy, weak anti-interference ability, etc. For example, PID control relies on accurate system model parameters, and when the system parameters change or there is external disturbance, the control effect will decrease significantly; although vector control can achieve high-performance motor speed regulation, it requires accurate measurement and identification of motor parameters, and when the load changes greatly, the speed regulation performance will also be affected to some extent.

[0003] In recent years, with the development of control theory, sliding mode control as an effective nonlinear control method has been gradually applied to the field of motor speed regulation. Sliding mode control switches the system state on the sliding surface quickly to achieve fast control of the system. However, traditional sliding mode control methods also have some shortcomings in practical application. For example, sliding mode control produces chattering phenomenon in the switching process, which not only affects the stability of motor speed regulation, but also may cause the mechanical wear of the motor to increase. In addition, traditional sliding mode control methods have poor adaptability to system parameter changes and external disturbances, and are difficult to meet the speed regulation requirements of electric forklifts in complex working conditions.

[0004] In the implementation of the embodiments of the present application, the inventors found that at least the following problems or defects exist in the prior art: the existing motor speed regulation control methods have insufficient speed regulation accuracy and anti-interference ability when facing complex working conditions and dynamic disturbances; although traditional sliding mode control methods can achieve fast control, they have chattering phenomenon and poor adaptability to system parameter changes and external disturbances, making it difficult to meet the speed regulation requirements of electric forklifts in complex working conditions. SUMMARY

[0005] The present application provides a high-order sliding mode composite control method for motor speed regulation of an electric forklift, comprising:

[0006] selecting a control period signal in the motor operating state data, and generating a control parameter sample from the control period signal;

[0007] sending the control parameter sample to a main controller module in a pre-constructed sliding mode control model for state quantity extraction to obtain motor speed regulation feature data, wherein the main controller module of the sliding mode control model is pre-deployed in the local controller;

[0008] sending the motor speed regulation feature data to a central processor for dynamic compensation of speed regulation parameters by a parameter adjustment module corresponding to the sliding mode control model in the central processor to obtain a regulation and control instruction of the motor speed regulation;

[0009] in response to the regulation and control instruction returned by the central processor, performing instruction output in an execution unit of the local controller.

[0010] Further, before sending the control parameter sample to the pre-constructed sliding mode control model, the following steps are performed:

[0011] performing noise suppression processing on the control parameter sample to remove high-frequency interference signals therein.

[0012] Further, the method further comprises:

[0013] performing state parameter detection on each sampling period signal in the motor operating state data to generate a state evaluation result set;

[0014] determining an evaluation result in the state evaluation result set that satisfies a preset condition, and determining a sampling signal corresponding to the evaluation result that satisfies the preset condition as a control period signal.

[0015] Further, the method further comprises:

[0016] performing fluctuation detection on the motor operating state data to determine dynamic disturbance components in the state data;

[0017] performing state tracking on the detected dynamic disturbance components, wherein the marked dynamic disturbance components are tracked to obtain a disturbance identification set;

[0018] determining a fluctuation range of the detected dynamic disturbance components in the operating state data to obtain a disturbance fluctuation boundary set;

[0019] based on the disturbance identification set and the disturbance fluctuation boundary set, performing state parameter detection on a sampling signal corresponding to each detected dynamic disturbance component to generate a state evaluation result set.

[0020] Further, the method further comprises:

[0021] In response to determining that there is no evaluation result in the state evaluation result set that meets the preset condition, reselecting the control cycle of the motor operation state data collected by the sensor.

[0022] Further, the method further comprises:

[0023] According to the disturbance fluctuation boundary set, determining a signal signal-to-noise ratio set and an interference suppression ratio set of the sampling signal corresponding to each identification in the disturbance identification set;

[0024] According to the disturbance fluctuation boundary set, determining a signal signal-to-noise ratio set and an interference suppression ratio set of the sampling signal corresponding to each identification in the disturbance identification set;

[0025] According to the signal signal-to-noise ratio set, the interference suppression ratio set and the working condition mode information set, a state evaluation result set corresponding to each disturbance identification is determined.

[0026] Further, the method further comprises:

[0027] According to the disturbance fluctuation boundary set, determining a signal signal-to-noise ratio set and an interference suppression ratio set of the sampling signal corresponding to each identification in the disturbance identification set;

[0028] According to the disturbance fluctuation boundary set, determining a signal signal-to-noise ratio set and an interference suppression ratio set of the sampling signal corresponding to each identification in the disturbance identification set.

[0029] Further, the sliding mode control model is generated by a preset parameter setting method; wherein, the preset parameter setting method comprises:

[0030] Obtaining a training data set, wherein the training samples in the training data set include reference speed signal, disturbance speed signal pairs and corresponding reference feature labels, disturbance feature labels;

[0031] Selecting training samples from the training data set and inputting them to the main controller module of the initial sliding mode control model to generate reference feature matrix and disturbance feature matrix, wherein the parameters of the main controller module in the initial sliding mode control model remain fixed, the parameter adjustment module in the initial sliding mode control model includes a global adjustment unit and a local adjustment unit, and the parameter adjustment module corresponds to a single control loop.

[0032] inputting the reference feature matrix into a parameter adjustment module of the initial sliding mode control model, and outputting a reference prediction label, wherein a local adjustment unit in the parameter adjustment module is configured to perform dynamic adjustment processing on transient features in the reference feature matrix;

[0033] inputting the disturbance feature matrix into a parameter adjustment module of the initial sliding mode control model, and outputting a disturbance prediction label, wherein a global adjustment unit in the parameter adjustment module is configured to perform dynamic adjustment processing on steady-state features in the reference feature matrix;

[0034] generating an adjustment error value according to the reference prediction label, the disturbance prediction label, the reference feature label and the disturbance feature label;

[0035] adjusting a control parameter in the parameter adjustment module according to the adjustment error value.

[0036] Further, the training data set is generated by the following steps:

[0037] receiving a filter removal operation of an operator on a displayed set of control parameters;

[0038] determining a speed regulation parameter corresponding to the filter removal operation as a disturbance speed regulation signal to obtain a disturbance signal group, wherein the disturbance speed regulation signal is an abnormal disturbance signal excluded by the operator;

[0039] determining a tracking identifier corresponding to each disturbance speed regulation signal in the disturbance signal group as a disturbance feature label to obtain a disturbance label group;

[0040] receiving a parameter integration operation of an operator on a displayed set of control parameters;

[0041] adjusting the tracking identifier corresponding to the parameter integration operation to the same identifier, and determining each speed regulation parameter in the set of control parameters as a reference speed regulation signal to obtain a reference signal group;

[0042] determining a tracking identifier corresponding to each reference speed regulation signal in the reference signal group as a reference feature label to obtain a reference label group;

[0043] combining the reference signal group, the reference label group, the disturbance signal group and the disturbance label group to obtain the training data set.

[0044] Further, the state quantity extraction is configured to extract multiple operating features of the motor, including at least one of the following: current features, torque features, and speed features, to perform multi-parameter fusion speed regulation control.

[0045] The above embodiments of the present application have at least the following beneficial effects:

[0046] (1) The high-order sliding mode composite control method for electric forklift motor speed regulation can improve the precision and stability of motor speed regulation. By selecting the control cycle signal in the motor operating state data and generating control parameter samples, and using the sliding mode control model for state quantity extraction and dynamic compensation of speed regulation parameters, the motor speed regulation characteristic data can be accurately extracted and efficient control instructions can be generated. In addition, the method can also perform noise suppression processing on the control parameter samples to remove high-frequency interference signals, thereby further improving the anti-interference ability of the speed regulation control and ensuring that the motor can still operate stably under complex working conditions.

[0047] (2) The parameter adjustment module in the initial sliding mode control model includes a global adjustment unit and a local adjustment unit. The global adjustment unit is used for dynamic adjustment processing of the steady-state characteristics in the reference characteristic matrix, and the local adjustment unit is used for dynamic adjustment processing of the transient characteristics in the reference characteristic matrix. In this way, the parameter adjustment module can simultaneously process the steady-state and transient characteristics of motor speed regulation to achieve more accurate speed regulation control.

[0048] (3) The application can also perform fluctuation detection and state tracking on the motor operating state data, determine the dynamic disturbance component and its fluctuation range, and generate a state evaluation result set based on this, and then realize real-time monitoring and accurate evaluation of the motor operating state. More specifically, when processing the motor operating state data, the fluctuation detection and dynamic disturbance component state tracking steps are further introduced, the disturbance identification set and the disturbance fluctuation boundary set are determined, and based on the disturbance identification set and the disturbance fluctuation boundary set, the state parameter detection of the sampling signal is performed, and the generated state evaluation result set can more comprehensively reflect the operating state of the motor under different working conditions, thereby providing a more accurate basis for speed regulation control.

[0049] (4) When processing the motor operating state data, a feedback mechanism is introduced, that is, when there is no evaluation result in the state evaluation result set that meets the preset condition, the control cycle selection of the motor operating state data is performed again. This mechanism ensures that the selection of the control cycle signal is always based on the actual operating state of the motor, avoiding the failure of speed regulation control due to incorrect or non-compliant signals. Through this feedback mechanism, the selection strategy of the control cycle signal can be dynamically adjusted to adapt to changes in the operating state of the motor. BRIEF DESCRIPTION OF DRAWINGS

[0050] The above and other objects, features and advantages of the exemplary embodiments of the present application will be more apparent from the following detailed description taken in conjunction with the accompanying drawings, in which:

[0051] Figure 1A flowchart of a high-order sliding mode composite control method for electric forklift motor speed regulation is provided for an embodiment of the present application. DETAILED DESCRIPTION

[0052] The principles and spirits of the present application will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are given only to enable those skilled in the art to better understand and implement the present application, and do not limit the scope of the present application in any way. On the contrary, these embodiments are provided to make the present application more thorough and complete, and to fully convey the scope of the present application to those skilled in the art.

[0053] Those skilled in the art know that the embodiments of the present application can be implemented as a system, device, apparatus, method or computer program product. Therefore, the present application can be embodied in the form of a complete hardware, a complete software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0054] It should be noted that any number of elements in the drawings is used for illustration only and not limitation, and any naming is only for distinction and does not have any limiting meaning.

[0055] Reference will be made to the following Figure 1 , Figure 1 A flowchart of a high-order sliding mode composite control method for electric forklift motor speed regulation is provided for an embodiment of the present application. As shown in Figure 1 , a high-order sliding mode composite control method for electric forklift motor speed regulation includes:

[0056] S1, selecting a control cycle signal in motor operating state data, and generating a control parameter sample from the control cycle signal;

[0057] S2, sending the control parameter sample to a main controller module in a pre-constructed sliding mode control model for state quantity extraction to obtain motor speed regulation feature data, wherein the main controller module of the sliding mode control model is pre-deployed in a local controller;

[0058] S3, sending the motor speed regulation feature data to a central processor for dynamic compensation of speed regulation parameters by a parameter adjustment module corresponding to the sliding mode control model in the central processor to obtain a regulation and control instruction of the motor speed regulation;

[0059] S4, in response to the regulation and control instruction returned by the central processor received, performing instruction output in an execution unit of the local controller.

[0060] It should be noted that the present application proposes a high-order sliding mode composite control method for electric forklift motor speed regulation. The core of this method is to realize accurate speed control through the analysis and processing of motor operating state data. Specifically, first, select the control period signal in the motor operating state data, the control period signal refers to the signal with periodic characteristics in the motor running process, such as the periodic change part of the motor speed signal or current signal. These signals can reflect the characteristic information of the motor under different operating conditions. Based on the control period signal, generate control parameter samples, control parameter samples refer to the parameter set extracted from the control period signal that can be used for subsequent control model analysis, such as the average value of motor speed, the peak value of current, etc. These parameter samples will be used as input data for subsequent sliding mode control model processing.

[0061] Specifically, the sliding mode control model is a control model based on sliding mode variable structure control theory, and its main controller module is used to extract state quantities from the input control parameter samples. State quantity extraction refers to extracting data from control parameter samples that can represent motor speed characteristics, such as motor current characteristics, torque characteristics, and speed characteristics, etc. These characteristic data can reflect the running state of the motor under different working conditions, and are the key information to realize accurate speed regulation. The main controller module is pre-deployed in the local controller, which is a controller installed on the electric forklift motor drive device, used to process and execute speed control instructions in real time. After the motor speed characteristic data is extracted, it is sent to the central processor, which is the core computing unit of the electric forklift control system, used to further process the motor speed characteristic data. The parameter adjustment module corresponding to the sliding mode control model in the central processor will dynamically compensate the speed parameters according to the motor speed characteristic data. Dynamic compensation of speed parameters refers to adjusting the speed control parameters in real time according to the actual operating state of the motor to adapt to the operating requirements of the motor under different working conditions. Finally, the central processor generates the motor speed regulation control instruction and returns it to the execution unit of the local controller, and the execution unit outputs specific speed control signals according to the control instruction to realize accurate control of the motor speed.

[0062] Preferably, the construction process of the sliding mode control model is as follows: first, obtain a training data set, which includes reference speed signal, disturbance speed signal pair, and corresponding reference feature label and disturbance feature label. The reference speed signal refers to the signal of motor speed under ideal working conditions, and the disturbance speed signal refers to the signal of motor speed when there is interference or dynamic change. The reference feature label and the disturbance feature label are used to identify the characteristics of the reference speed signal and the disturbance speed signal respectively. Then, the training sample is input to the main controller module of the initial sliding mode control model to generate a reference feature matrix and a disturbance feature matrix. The reference feature matrix and the disturbance feature matrix are used to describe the characteristic information of the reference speed signal and the disturbance speed signal respectively. The parameter adjustment module in the initial sliding mode control model includes a global adjustment unit and a local adjustment unit, the global adjustment unit is used to dynamically adjust the steady-state characteristics in the reference feature matrix, and the local adjustment unit is used to dynamically adjust the transient characteristics in the reference feature matrix. In this way, the parameter adjustment module can simultaneously process the steady-state and transient characteristics of motor speed control, and realize more accurate speed control.

[0063] In some embodiments, before sending the control parameter sample to the pre-constructed sliding mode control model, the following steps are performed:

[0064] The control parameter sample is subjected to noise suppression processing to remove high-frequency interference signals therein.

[0065] It should be noted that when processing the control parameter sample, the present application particularly emphasizes the importance of noise suppression processing. Noise suppression processing refers to removing high-frequency interference signals in the control parameter sample through specific signal processing techniques to improve signal quality and control accuracy. High-frequency interference signals are usually caused by electromagnetic interference, sensor noise or other external factors during motor operation, and these interference signals can seriously affect the accuracy and stability of motor speed control. Through noise suppression processing, the influence of these disturbances can be effectively reduced, thereby providing a more reliable data basis for subsequent speed control.

[0066] Specifically, the noise suppression processing involves the analysis and processing of control parameter samples. Control parameter samples are extracted from motor operating state data and contain various characteristic information of motor operation, such as sampling values of current, speed, torque, etc. High-frequency interference signals usually exhibit rapidly changing signal components with frequencies much higher than the frequency range of normal motor operation signals. Noise suppression processing can be achieved through various methods, such as using a low-pass filter to filter out high-frequency components. A low-pass filter is a signal processing tool that allows low-frequency signals to pass through while attenuating high-frequency signals. In practical applications, the cutoff frequency of the low-pass filter can be set according to the frequency characteristics of the motor operating signal to ensure that high-frequency interference signals are removed while retaining useful information of the motor operating signal.

[0067] Preferably, the specific implementation steps of noise suppression processing can include the following contents: first, determine the frequency characteristics of the control parameter samples, analyze the frequency component distribution in the control parameter samples through Fast Fourier Transform (FFT) or other spectral analysis methods, and determine the frequency range of the high-frequency interference signal. Then, select the appropriate low-pass filter type according to the analysis results, such as Butterworth filter, Chebyshev filter, etc., and set its cutoff frequency. The selection of the cutoff frequency should be based on the highest frequency component of the normal motor operating signal to ensure that the filter will not distort the motor operating signal. Finally, filter the control parameter samples through the low-pass filter to remove high-frequency interference signals. The control parameter samples after noise suppression processing are more pure and can more accurately reflect the actual operating state of the motor, providing high-quality input data for the subsequent sliding mode control model, thereby improving the performance and reliability of motor speed control.

[0068] In some embodiments, the method further comprises:

[0069] detecting state parameters in each sampling period signal in the motor operating state data to generate a set of state evaluation results;

[0070] determining evaluation results in the set of state evaluation results that meet a preset condition, and determining the sampling signal corresponding to the evaluation results that meet the preset condition as a control period signal.

[0071] It should be noted that the application emphasizes the importance of state parameter detection for each sampling period signal when processing motor operating state data. State parameter detection refers to analyzing the sampling period signal in the motor operating state data to generate a state evaluation result set, thereby providing a basis for subsequent control period signal selection. The state evaluation result set is a set of evaluation results obtained by analyzing the sampling period signal, which is used to reflect whether the motor operating state in different sampling periods meets the preset conditions. The preset conditions are set according to the normal operation requirements of the motor, such as whether the motor speed is within the allowed range, whether the current is abnormal, etc. In this way, the selected control period signal can accurately reflect the actual operating state of the motor, thereby improving the accuracy and reliability of speed control.

[0072] Specifically, state parameter detection involves analyzing the sampling period signal in the motor operating state data. The sampling period signal refers to the motor operating data collected at certain time intervals during the operation of the motor, such as the sampling values of current, speed, torque, etc. The purpose of state parameter detection is to determine whether the motor is in a normal operating state in each sampling period by analyzing these sampling period signals. The state evaluation result set is obtained by processing the sampling period signal through a series of detection algorithms, which can include threshold judgment, trend analysis, etc. For example, by setting a current threshold to determine whether the motor is overloaded, or by analyzing the trend of the speed to determine whether the motor is in a stable operating state. The evaluation results that meet the preset conditions refer to those state evaluation results that meet the normal operation requirements of the motor, such as the motor speed being within the allowed range and the current not exceeding the rated value. The sampling signals corresponding to these evaluation results will be determined as control period signals for subsequent speed control.

[0073] Preferably, the specific implementation steps of state parameter detection can be further refined. First, determine the sampling period signal in the motor operating state data, which usually includes the sampling values of the motor's current, speed, torque, etc. Then, set the preset conditions according to the operating requirements of the motor, such as current threshold, speed range, etc. Next, analyze each sampling period signal through detection algorithms to generate state evaluation results. For example, a sliding window algorithm can be used to smooth the speed signal, then calculate its average value and fluctuation range, and determine whether it meets the preset speed range condition. For the current signal, a current threshold can be set, and when the sampled current exceeds the threshold, it is determined that the motor is overloaded. Finally, the sampling signal corresponding to the state evaluation result that meets the preset condition is determined as the control period signal. These control period signals will serve as the basis for subsequent speed control, ensuring that the motor can achieve precise speed control under different operating conditions.

[0074] In some embodiments, the method further comprises:

[0075] detecting fluctuations in the motor operating state data to determine dynamic disturbance components in the state data;

[0076] tracking the detected dynamic disturbance components, wherein the labeled dynamic disturbance components are tracked to obtain a disturbance identification set;

[0077] determining fluctuation ranges of the detected dynamic disturbance components in the operating state data to obtain a disturbance fluctuation boundary set;

[0078] based on the disturbance identification set and the disturbance fluctuation boundary set, detecting state parameters of each detected dynamic disturbance component corresponding to the sampled signal to generate a state evaluation result set.

[0079] It should be noted that the present application further introduces fluctuation detection and dynamic disturbance component tracking steps when processing motor operating state data. Fluctuation detection refers to identifying dynamic disturbance components in the motor operating state data through analysis. These disturbance components may be caused by load changes, power grid fluctuations or other external disturbances. Dynamic disturbance component tracking is to mark and track these disturbance components so that subsequent processing of the impact of these disturbances on motor speed control can be more accurate. By determining the fluctuation range of the dynamic disturbance component, i.e. the disturbance fluctuation boundary set, the amplitude and influence range of the disturbance can be better understood. Based on the disturbance identification set and the disturbance fluctuation boundary set, the state parameter detection of the sampled signal generates a state evaluation result set that can more comprehensively reflect the operating state of the motor under different working conditions, thereby providing a more accurate basis for speed control.

[0080] Specifically, fluctuation detection and dynamic disturbance component tracking involve multiple key concepts. Dynamic disturbance components are abnormal fluctuations in motor operating state data caused by external disturbances or internal changes. These disturbances may be transient or continuous, and are characterized by abnormal changes in parameters such as current, speed or torque. The disturbance identification set is the result of labeling the detected dynamic disturbance components, with each disturbance identification corresponding to a specific disturbance event. The disturbance fluctuation boundary set is the result of defining the fluctuation range of each disturbance component, which describes the amplitude and duration of the disturbance. State parameter detection is a further analysis of the sampled signal, which calculates parameters such as signal-to-noise ratio and interference suppression ratio to evaluate the quality and reliability of the signal. The working condition mode information set is identified based on load characteristics and describes the operating state of the motor under different load conditions, such as no load, half load or full load. The generation of these parameters and information sets provides rich data support for subsequent speed control.

[0081] Preferably, the specific implementation steps of fluctuation detection and dynamic disturbance component state tracking can be further refined. First, through a signal processing algorithm (such as wavelet transform or adaptive filtering), the motor operating state data is subjected to fluctuation detection, and the dynamic disturbance component is identified. Then, each detected disturbance component is labeled to form a disturbance identification set. Next, through a tracking algorithm (such as a Kalman filter), the state of the disturbance component is tracked to determine its fluctuation range and generate a disturbance fluctuation boundary set. On this basis, the state parameters of the sampling signal corresponding to each disturbance identification are detected, the signal-to-noise ratio and interference suppression ratio of the signal are calculated, and the signal quality is evaluated. At the same time, the sampling signal corresponding to each disturbance fluctuation boundary is subjected to load characteristic identification to generate a working condition mode information set, and the operating conditions of the motor under different disturbances are clarified. Finally, combined with the signal-to-noise ratio, interference suppression ratio and working condition mode information, a state evaluation result set is generated to provide accurate data support for subsequent speed control.

[0082] In some embodiments, the method further comprises:

[0083] In response to determining that there is no evaluation result in the state evaluation result set that meets the preset condition, reselecting the motor operating state data collected by the sensor for control cycle selection.

[0084] It should be noted that when the motor operating state data is processed, a feedback mechanism is introduced, i.e., when there is no evaluation result in the state evaluation result set that meets the preset condition, the motor operating state data is reselected for control cycle selection. This mechanism ensures that the selection of the control cycle signal is always based on the actual operating state of the motor, avoiding the failure of speed control due to incorrect or non-compliant signals. The state evaluation result set is generated by state parameter detection of the sampling cycle signal, and is used to determine whether the motor operating state meets the preset speed control requirements. The preset condition is set according to the normal operating parameter range of the motor, such as whether the speed is within the allowed range and whether the current exceeds the rated value. Through this feedback mechanism, the selection strategy of the control cycle signal can be dynamically adjusted to adapt to changes in the operating state of the motor.

[0085] Specifically, the state evaluation result set refers to a group of evaluation results obtained through state parameter detection, and each evaluation result corresponds to the state of a sampling period signal. The preset condition refers to the parameter range or condition that should be met during normal operation of the motor, such as current not exceeding the rated value, speed being within the set interval, etc. When there is no evaluation result in the state evaluation result set that meets these preset conditions, it indicates that the currently selected control period signal cannot reflect the true operating state of the motor or does not meet the speed regulation requirements. Control period selection refers to selecting an appropriate signal period from the motor operating state data as the basis for subsequent speed regulation control. This process needs to be dynamically adjusted according to the actual operation of the motor to ensure that the selected signal can accurately reflect the operating state of the motor. Re-selecting the control period signal means that the motor operating state data needs to be re-analyzed to find a signal period that meets the preset conditions.

[0086] Preferably, the specific implementation steps of re-selecting the control period signal can be further refined. First, when there is no evaluation result in the state evaluation result set that meets the preset conditions, the re-selection mechanism is triggered. At this time, the motor operating state data needs to be re-analyzed, which can be achieved by increasing the sampling frequency or adjusting the sampling time window to obtain more detailed operating data. Then, the re-collected data is subjected to state parameter detection to regenerate the state evaluation result set. This process may require adjusting the parameters of the detection algorithm, such as threshold values or filter parameters, to adapt to changes in the operating state of the motor. Finally, the control period signal is re-determined based on the new state evaluation result set. If a signal that meets the conditions cannot still be found, the preset conditions or the detection algorithm can be further adjusted or optimized to ensure that a suitable control period signal can be finally selected, thereby providing reliable data support for motor speed regulation control.

[0087] In some embodiments, based on the disturbance identification set and the disturbance fluctuation boundary set, the state parameter detection is performed on the sampling signal corresponding to each dynamic disturbance component to generate a state evaluation result set, including:

[0088] According to the disturbance fluctuation boundary set, a signal signal-to-noise ratio set and an interference suppression ratio set of the sampling signal corresponding to each identification in the disturbance identification set are determined;

[0089] Load characteristic identification is performed on the sampling signal corresponding to each fluctuation boundary in the disturbance fluctuation boundary set to generate a working condition mode information set, wherein each working condition mode information in the working condition mode information set corresponds to an operating state, and each working condition mode information includes an operating working condition label of the motor relative to the driving device in the control period signal, and the operating working condition label includes at least one of the following: an idle load working condition label, a half load working condition label, and a full load working condition label;

[0090] According to the signal-to-noise ratio set, the interference suppression ratio set, and the working condition mode information set, a state evaluation result set corresponding to each disturbance identification is determined.

[0091] It should be noted that the present application proposes a state parameter detection method based on the disturbance fluctuation boundary set and the disturbance identification set when processing the sampling signal corresponding to the dynamic disturbance component. This method comprehensively considers the signal-to-noise ratio, the interference suppression ratio, and the working condition mode information to generate a state evaluation result set, thereby more comprehensively evaluating the motor operating state. The signal-to-noise ratio (SNR) and the interference suppression ratio (ISR) are important parameters for measuring signal quality, which respectively reflect the proportion of effective components and noise components in the signal and the suppression ability to interference. The working condition mode information is obtained by identifying the load characteristics of the sampling signal, which is used to describe the characteristics of the motor under different operating conditions, such as no-load, half-load, or full-load. In this way, the operating state of the motor under dynamic disturbance can be more accurately evaluated, providing a more accurate basis for subsequent speed control.

[0092] Specifically, the disturbance fluctuation boundary set refers to the fluctuation range of the dynamic disturbance component, which defines the amplitude and duration of the disturbance. The disturbance identification set is the result of labeling the detected dynamic disturbance component, and each identification corresponds to a specific disturbance event. The signal-to-noise ratio (SNR) is the ratio of signal power to noise power, which is used to measure the quality and reliability of the signal. The interference suppression ratio (ISR) is a parameter that measures the suppression ability of the system to interference signals, reflecting the reduction of interference components in the processed signal. The working condition mode information set is obtained by identifying the load characteristics of the sampling signal, which contains the operating state labels of the motor under different load conditions, such as no-load condition label, half-load condition label, and full-load condition label. These labels are used to describe the operating conditions of the motor relative to the driving device, for no-load, half-load, or full-load state. By combining these parameters and information, a state evaluation result set can be generated, thereby more comprehensively evaluating the operating state of the motor.

[0093] Preferably, the specific implementation steps of the state parameter detection based on the disturbance fluctuation boundary set and the disturbance identification set can be further refined. First, according to the disturbance fluctuation boundary set, the signal-to-noise ratio (SNR) and the interference suppression ratio (ISR) of the sampling signal corresponding to each disturbance identification are determined. This can be achieved by calculating the power spectral density of the sampling signal, taking the ratio of signal power to noise power as the signal-to-noise ratio, and calculating the interference suppression ratio by comparing the signal power before and after processing. Next, the load characteristic identification is performed on the sampling signal corresponding to each fluctuation boundary in the disturbance fluctuation boundary set, and a working condition mode information set is generated. Load characteristic identification can be achieved by analyzing the trend of parameters such as current, speed and torque of the motor, for example, by the size of the current to determine whether the motor is in an idle or full load state. Finally, combined with the signal-to-noise ratio, the interference suppression ratio and the working condition mode information, a state evaluation result set is generated. Each evaluation result in the state evaluation result set reflects the quality of the sampling signal under certain disturbance conditions, the interference suppression effect and the running condition of the motor, providing comprehensive data support for subsequent speed control.

[0094] In some embodiments, the method further comprises:

[0095] determining the signal-to-noise ratio of the sampling signal corresponding to each identification in the disturbance identification set, obtaining a signal-to-noise ratio set;

[0096] determining the interference suppression ratio of the sampling signal corresponding to each identification in the disturbance identification set, obtaining an interference suppression ratio set.

[0097] It should be noted that the present application further refines the analysis process of the sampling signal corresponding to each identification in the disturbance identification set, especially the determination of the signal-to-noise ratio and the interference suppression ratio. The signal-to-noise ratio (SNR) and the interference suppression ratio (ISR) are important indicators for measuring signal quality and system anti-interference ability. By calculating these parameters respectively, the quality of the sampling signal and the interference suppression effect of the system can be more accurately evaluated. The signal-to-noise ratio reflects the proportion of effective components and noise components in the signal, while the interference suppression ratio measures the reduction degree of interference components in the processed signal. The calculation results of these two parameters will provide key data support for subsequent state evaluation, so as to more accurately reflect the running state of the motor under dynamic disturbance conditions.

[0098] Specifically, the disturbance identification set refers to the result of labeling the detected dynamic disturbance components, and each identification corresponds to a specific disturbance event. The sampled signal refers to the signal extracted from the motor operating state data for subsequent analysis and processing. The signal-to-noise ratio (SNR) is the ratio of signal power to noise power, usually determined by calculating the power spectral density of the signal. The interference suppression ratio (ISR) is a parameter that measures the system's ability to suppress interference signals, which can be calculated by comparing the signal power before and after processing. In practical applications, the signal-to-noise ratio can be calculated by performing Fourier transform on the sampled signal and analyzing its spectral characteristics. The interference suppression ratio can be evaluated by comparing the signal power before and after processing, for example, by calculating the ratio of the signal power after low-pass filter processing to the original signal power. The results of these parameters will provide important basis for subsequent state evaluation, helping to more accurately judge the operating state of the motor.

[0099] Preferably, the specific implementation steps for determining the signal-to-noise ratio and the interference suppression ratio can be further refined. First, for each disturbance identification corresponding to the sampled signal, the power spectral density of the signal is calculated by analyzing its spectral characteristics through Fourier transform. Then, the ratio of signal power to noise power is calculated based on the power spectral density to obtain the signal-to-noise ratio (SNR). Next, the sampled signal is processed to suppress noise, such as using a low-pass filter to remove high-frequency interference signals. After processing, the power spectral density of the signal is calculated again, and the ratio of the processed signal power to the original signal power is calculated to obtain the interference suppression ratio (ISR). Finally, the calculated signal-to-noise ratio and interference suppression ratio are stored in the corresponding set to provide data support for subsequent state evaluation. In this way, the quality of the sampled signal and the anti-interference ability of the system can be more accurately evaluated, providing more reliable basis for motor speed control.

[0100] In some embodiments, the sliding mode control model is generated by a preset parameter tuning method; wherein the preset parameter tuning method comprises:

[0101] Obtain a training data set, wherein the training samples in the training data set include reference speed signal, disturbance speed signal pair and corresponding reference feature label, disturbance feature label;

[0102] Select training samples from the training data set and input them to the main controller module of the initial sliding mode control model to generate reference feature matrix and disturbance feature matrix, wherein the parameters of the main controller module in the initial sliding mode control model remain fixed, the parameter adjustment module in the initial sliding mode control model includes a global adjustment unit and a local adjustment unit, and the parameter adjustment module corresponds to a single control loop;

[0103] The reference feature matrix is input into a parameter adjustment module of the initial sliding mode control model, and a reference prediction label is output, wherein a local adjustment unit in the parameter adjustment module is used for dynamically adjusting transient features in the reference feature matrix;

[0104] The disturbance feature matrix is input into the parameter adjustment module of the initial sliding mode control model, and a disturbance prediction label is output, wherein a global adjustment unit in the parameter adjustment module is used for dynamically adjusting steady-state features in the reference feature matrix;

[0105] An adjustment error value is generated according to the obtained reference prediction label, disturbance prediction label, reference feature label and disturbance feature label;

[0106] The control parameters in the parameter adjustment module are adjusted according to the adjustment error value.

[0107] It should be noted that the present application proposes a construction method of a sliding mode control model, which adjusts parameters of an initial sliding mode control model through a training data set to generate a high-order sliding mode composite control model that can adapt to complex working conditions. The sliding mode control model is a model based on sliding mode variable structure control theory, and its core is to realize fast and accurate control of motor speed by designing a sliding mode surface and a parameter adjustment module. The training data set includes reference speed signal, disturbance speed signal pair and corresponding reference feature label and disturbance feature label, and these data are used to train the model to identify speed signal features in different working conditions. Through the global adjustment unit and the local adjustment unit in the parameter adjustment module, the model can dynamically compensate the speed parameters, thereby improving the adaptability and stability of the speed control.

[0108] Specifically, the main controller module of the sliding mode control model is the core part of the model, which is used for state quantity extraction of the input control parameter sample to generate a reference feature matrix and a disturbance feature matrix. The reference feature matrix and the disturbance feature matrix are respectively used to describe the feature information of the reference speed signal and the disturbance speed signal. The parameter adjustment module includes a global adjustment unit and a local adjustment unit, the global adjustment unit is used for dynamically adjusting the steady-state features in the reference feature matrix, and the local adjustment unit is used for dynamically adjusting the transient features. The reference prediction label and the disturbance prediction label are output results obtained by inputting the feature matrix into the parameter adjustment module, and are used to evaluate the prediction performance of the model. The adjustment error value is calculated by comparing the prediction label with the actual feature label, and is used to adjust the control parameters in the parameter adjustment module to optimize the performance of the model.

[0109] Preferably, the construction step of the sliding mode control model can be further refined. First, a training dataset containing reference speed signals and disturbance speed signals reflecting the operating state of the motor under different working conditions is collected. Then, the training samples are input into the main controller module of the initial sliding mode control model to generate a reference feature matrix and a disturbance feature matrix. Next, the reference feature matrix is input into the local adjustment unit of the parameter adjustment module to output a reference prediction label; the disturbance feature matrix is input into the global adjustment unit to output a disturbance prediction label. By comparing the prediction labels with the actual labels, the adjustment error value is calculated. Finally, the control parameters in the parameter adjustment module are adjusted according to the adjustment error value to optimize the model performance. For example, the parameters can be adjusted by gradient descent method or other optimization algorithms to minimize the adjustment error value, thereby improving the adaptability of the model to different working conditions and the speed regulation accuracy.

[0110] In some embodiments, the training dataset is generated by the following steps:

[0111] receiving an operator's filtering removal operation on the displayed speed regulation parameters in the control parameter set;

[0112] determining the speed regulation parameter corresponding to the filtering removal operation as a disturbance speed signal to obtain a disturbance signal group, wherein the disturbance speed signal is an abnormal disturbance signal excluded by the operator;

[0113] determining the tracking identifier corresponding to each disturbance speed signal in the disturbance signal group as a disturbance feature label to obtain a disturbance label group;

[0114] receiving an operator's parameter integration operation on the displayed speed regulation parameters in the control parameter set;

[0115] adjusting the tracking identifier corresponding to the parameter integration operation to the same identifier, and determining each speed regulation parameter in the control parameter set as a reference speed signal to obtain a reference signal group;

[0116] determining the tracking identifier corresponding to each reference speed signal in the reference signal group as a reference feature label to obtain a reference label group;

[0117] combining the reference signal group, the reference label group, the disturbance signal group and the disturbance label group to obtain the training dataset.

[0118] It should be noted that the application further illustrates the generation method of the training data set, which constructs the training data set through the filtering removal operation and the parameter integration operation of the operator on the control parameter set. This method can make full use of the experience of the operator, classify and label the abnormal disturbance signals and normal speed regulation signals in the actual operation, so as to generate a high-quality training data set. The construction of the training data set is the basis for the training of the sliding mode control model, and the data set generated in this way can better reflect the working conditions of the motor in the actual operation, and provide strong support for the training and optimization of the model.

[0119] Specifically, the generation of the training data set involves multiple key concepts. The filtering removal operation refers to the process of excluding abnormal disturbance signals from the control parameter set by the operator. These abnormal disturbance signals are usually caused by interference or failure in the operation of the motor. Disturbance speed regulation signals refer to these excluded abnormal signals, which are labeled as disturbance feature labels to represent the abnormal characteristics of the signals. The parameter integration operation refers to the process of integrating normal speed regulation signals in the control parameter set by the operator. These normal speed regulation signals are labeled as reference feature labels to represent the normal characteristics of the signals. Reference speed regulation signals refer to the normal speed regulation signals after integration, which together with the reference feature labels constitute the reference signal group in the training data set. In this way, the training data set can contain speed regulation signals under both normal and abnormal working conditions, providing comprehensive data support for model training.

[0120] Preferably, the generation step of the training data set can be further refined. First, the operator observes the speed regulation parameters in the control parameter set through the monitoring system, identifies and performs the filtering removal operation to remove abnormal disturbance signals from the data set and label them as disturbance feature labels. Next, the operator performs the parameter integration operation on the remaining normal speed regulation signals, labeling them as reference feature labels. Then, the reference speed regulation signals and the disturbance speed regulation signals are combined into the reference signal group and the disturbance signal group respectively, and the corresponding feature labels are combined into the reference label group and the disturbance label group. Finally, these signal groups and label groups are combined to form a complete training data set. For example, the operator can set a threshold to identify abnormal signals, and when the speed regulation parameter exceeds the threshold, it is labeled as a disturbance signal; for normal signals, statistical analysis is performed for integration to ensure the representativeness and consistency of the reference signals. The training data set generated in this way can more accurately reflect various working conditions of the motor in operation, providing a high-quality data basis for the training of the sliding mode control model.

[0121] In some embodiments, the state quantity extraction is used to extract multiple operating characteristics of the motor, including at least one of the following: current characteristics, torque characteristics, and speed characteristics, for multi-parameter fusion speed regulation control.

[0122] It should be noted that the application emphasizes the function of the sliding mode control model in state quantity extraction, especially its ability to extract various operating characteristics of the motor, including current characteristics, torque characteristics, and speed characteristics, to achieve multi-parameter fusion speed control. State quantity extraction refers to extracting key information that can represent the operating state of the motor from motor operating state data, which is crucial for achieving precise speed control. Through multi-parameter fusion, the sliding mode control model can comprehensively consider multiple key factors in motor operation, thereby improving the precision and adaptability of speed control, enabling it to better cope with complex working conditions.

[0123] Specifically, state quantity extraction refers to extracting key parameters that can reflect the operating state of the motor from motor operating state data. Current characteristics refer to information such as the size and trend of the current during motor operation, such as the peak value, average value, and fluctuation range of the current. Torque characteristics refer to the size and variation of the motor output torque, such as torque stability and dynamic response characteristics. Speed characteristics refer to the numerical value and trend of the motor speed, such as speed stability, acceleration and deceleration performance, etc. These characteristic parameters can comprehensively reflect the operating state of the motor, providing an important basis for speed control. Multi-parameter fusion speed control refers to comprehensively considering multiple characteristic parameters and analyzing and processing these parameters through algorithms to achieve precise control of motor speed. This method can better adapt to the operating requirements of the motor under different working conditions, improving the flexibility and stability of speed control.

[0124] Preferably, the specific implementation steps of state quantity extraction can be further refined. First, collect motor operating state data through sensors, including real-time values of current, torque, and speed parameters. Then, preprocess the collected data, such as removing noise, filling missing values, etc., to improve data quality. Next, use signal processing algorithms to extract current characteristics, torque characteristics, and speed characteristics. For example, current characteristics can be extracted by calculating the mean and standard deviation of the current signal, torque characteristics can be extracted by analyzing the rate of change of the torque signal, and speed characteristics can be extracted by calculating the frequency and amplitude of the speed signal. Finally, input these characteristic parameters into the sliding mode control model, which generates speed control instructions by comprehensively analyzing these parameters, achieving multi-parameter fusion speed control. For example, when an abnormal increase in current is detected, the model can determine that the motor may be in an overload state and adjust the speed parameters accordingly to reduce the motor load; when the speed fluctuates greatly, the model can stabilize the motor operation by adjusting the speed characteristic parameters. In this way, the sliding mode control model can achieve comprehensive monitoring and precise control of the motor operating state, improving the performance and reliability of motor speed control.

[0125] The above various embodiments of the present application have the following beneficial effects: the present application can improve the control performance of the electric fork truck motor speed regulation system, and through intelligent selection and noise suppression processing of the control cycle signal, the accuracy of the control parameters can be ensured. The cooperative working mechanism of the main controller module adopting the sliding mode control model and the central processor can realize accurate extraction and dynamic compensation of the speed regulation characteristics, thereby optimizing the generation process of the speed regulation instruction. Combined with the state parameter detection and dynamic disturbance component tracking compensation mechanism, the anti-interference ability of the system under complex working conditions can be enhanced, and through the signal-to-noise ratio analysis and load characteristic identification, the speed regulation demand under different load conditions can be adapted.

[0126] Through model training of the preset parameter setting method, an intelligent control system with global and local regulation capabilities can be established, so that the speed regulation process has both steady-state accuracy and transient response performance. The training data generation mode based on operator interaction can ensure that the system learns the optimal speed regulation parameter characteristics. The multi-parameter fusion speed regulation strategy can integrate key characteristics such as current, torque and speed to realize high-precision control of the motor drive device. This composite control method can simultaneously consider response speed, control accuracy and running stability.

[0127] Further, the storage medium of the embodiments of the present application stores program instructions capable of implementing all the above methods, wherein the program instructions can be stored in the above storage medium in the form of a software product, including a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor to execute all or part of the steps of the methods described in the embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media capable of storing program codes, or a computer, a server, a mobile phone, a tablet, and the like.

[0128] The above description is only some of the preferred embodiments of the present application and a description of the principles of the technology used. Those skilled in the art should understand that the scope of the application involved in the embodiments of the present application is not limited to the technical solutions formed by the specific combinations of the above technical features, and should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above inventive concept. For example, the above features and technical features with similar functions disclosed in the embodiments of the present application (but not limited to) are replaced with each other to form a technical solution.

Claims

1. A high-order sliding mode compound control method for speed regulation of an electric forklift motor, applied to a motor drive device, characterized in that, The method comprises: selecting a control cycle signal in the motor operating state data, and generating a control parameter sample from the control cycle signal; sending the control parameter sample to a main controller module in a pre-constructed sliding mode control model for state quantity extraction to obtain motor speed regulation feature data, wherein the main controller module of the sliding mode control model is pre-deployed in a local controller; wherein the sliding mode control model is a high-order sliding mode composite control model suitable for complex working conditions, and the fast and accurate control of motor speed regulation is realized through the design of a sliding surface and a parameter adjustment module; sending the motor speed regulation feature data to a central processor for dynamic compensation of the speed regulation parameters by a parameter adjustment module corresponding to the sliding mode control model in the central processor to obtain the regulation and control instruction of the motor speed regulation; in response to the regulation and control instruction returned by the central processor, outputting the instruction in the execution unit of the local controller; and the sliding mode control model is generated by a preset parameter setting method; wherein the preset parameter setting method comprises: obtaining a training data set, wherein the training samples in the training data set include reference speed regulation signals, disturbance speed regulation signal pairs, and corresponding reference feature labels and disturbance feature labels; selecting training samples from the training data set and inputting them into the main controller module of the initial sliding mode control model to generate reference feature matrices and disturbance feature matrices, wherein the parameters of the main controller module in the initial sliding mode control model remain fixed, the parameter adjustment module in the initial sliding mode control model includes a global adjustment unit and a local adjustment unit, and the parameter adjustment module corresponds to a single control loop; inputting the reference feature matrix into the parameter adjustment module of the initial sliding mode control model to output a reference prediction label, wherein the local adjustment unit in the parameter adjustment module is used for dynamic adjustment processing of transient features in the reference feature matrix; inputting the disturbance feature matrix into the parameter adjustment module of the initial sliding mode control model to output a disturbance prediction label, wherein the global adjustment unit in the parameter adjustment module is used for dynamic adjustment processing of steady-state features in the reference feature matrix; generating an adjustment error value according to the obtained reference prediction label, disturbance prediction label, reference feature label, and disturbance feature label; adjusting the control parameters in the parameter adjustment module according to the adjustment error value, wherein the adjustment of the parameters is realized by the gradient descent method.

2. The method of claim 1, wherein, Before sending the control parameter sample to the pre-constructed sliding mode control model, the following steps are performed: performing noise suppression processing on the control parameter sample to remove high-frequency interference signals therein.

3. The method of claim 1, wherein, The method further comprises: performing state parameter detection on each sampling cycle signal in the motor operating state data to generate a state evaluation result set; determining evaluation results in the state evaluation result set that satisfy a preset condition, and determining the sampling signals corresponding to the evaluation results that satisfy the preset condition as control cycle signals.

4. The method according to any of claims 1 or 3, characterized in that, The method further comprises: performing fluctuation detection on the motor operating state data to determine dynamic disturbance components in the state data; tracking the detected dynamic disturbance components to obtain a disturbance identification set; determining fluctuation ranges of the detected dynamic disturbance components in the operating state data to obtain a disturbance fluctuation boundary set; based on the disturbance identification set and the disturbance fluctuation boundary set, performing state parameter detection on the sampling signals corresponding to each detected dynamic disturbance component to generate a state evaluation result set.

5. The method of claim 3, wherein, The method further comprises: in response to determining that there is no evaluation result in the state evaluation result set that meets the preset condition, reselecting the control period of the motor operating state data collected by the sensor.

6. The method of claim 4, wherein, The method further comprises: based on the disturbance identification set and the disturbance fluctuation boundary set, performing state parameter detection on the sampling signals corresponding to each detected dynamic disturbance component to generate a state evaluation result set. According to the disturbance fluctuation boundary set, determine a signal signal-to-noise ratio set and an interference suppression ratio set of the sampling signals corresponding to each identification in the disturbance identification set; performing load characteristic identification on the sampling signals corresponding to each fluctuation boundary in the disturbance fluctuation boundary set to generate a working condition mode information set, wherein each working condition mode information in the working condition mode information set corresponds to an operating state, and each working condition mode information includes an operating condition label of the motor relative to the driving device in the control period signal, and the operating condition label includes at least one of the following: an idle load condition label, a half load condition label, and a full load condition label; 7. The method of claim 6, wherein, According to the signal signal-to-noise ratio set, the interference suppression ratio set and the working condition mode information set, determine a state evaluation result set corresponding to each disturbance identification. The method further comprises: determining the signal signal-to-noise ratio of the sampling signals corresponding to each identification in the disturbance identification set to obtain a signal signal-to-noise ratio set; 8. The method of claim 1, wherein, determining the interference suppression ratio of the sampling signals corresponding to each identification in the disturbance identification set to obtain an interference suppression ratio set. The training data set is generated by the following steps: receiving an operator's filter removal operation on the displayed speed regulation parameters in the control parameter set; determining the speed regulation parameter corresponding to the filter removal operation as a disturbance speed regulation signal to obtain a disturbance signal group, wherein the disturbance speed regulation signal is an abnormal disturbance signal excluded by the operator; determining the tracking identification corresponding to each disturbance speed regulation signal in the disturbance signal group as a disturbance feature label to obtain a disturbance label group; receiving an operator's parameter integration operation on the displayed speed regulation parameters in the control parameter set; adjusting the tracking identification corresponding to the parameter integration operation to the same identification, and determining each speed regulation parameter in the control parameter set as a reference speed regulation signal to obtain a reference signal group; determining the tracking identification corresponding to each reference speed regulation signal in the reference signal group as a reference feature label to obtain a reference label group; 9. The method of claim 1, wherein, combining the reference signal group, the reference label group, the disturbance signal group and the disturbance label group to obtain a training data set. The state quantity extraction is used to extract various operating characteristics of the motor, including at least one of the following: current characteristics, torque characteristics, and speed characteristics, to perform multi-parameter fusion speed regulation control.

Images