Intelligent control method for electric actuator

Abstract

The application relates to the technical field of actuator calibration, in particular to an intelligent control method for an electric actuator, which comprises the following steps: starting a calibration mode of the actuator, controlling the actuator to run from zero range to full range, acquiring actual displacements corresponding to different ranges and theoretical displacements, determining correction errors of each range, obtaining an error change rate, screening the ranges, acquiring abnormal points and normal points, collecting multiple actuators of the same type, calculating relative change deviations of the abnormal points, constructing a compensation mapping model according to a mapping relationship between the relative change deviations of the abnormal points and the error change rates of all the actuators, predicting the relative change deviations of the ranges for a target actuator to be calibrated, generating a compensation function, calibrating a range corresponding to a target displacement, and generating a calibrated PWM control instruction to drive the target actuator. The application can actively offset nonlinear errors of the actuator, and realizes accurate positioning control in a full-range range.

Description

Technical Field

[0001] This application relates to the field of actuator calibration technology, specifically to an intelligent control method for electric actuators. Background Technology

[0002] Electric actuators, as key terminal operating components in industrial automation systems, are widely used in applications requiring precise position control, such as valves, dampers, and robotic arms. By receiving PWM control commands, the actuator drives the controlled object to complete proportional opening and closing processes. Its control accuracy directly determines the final positioning accuracy of loads such as valves and dampers.

[0003] Ideally, there is a strict linear relationship between the input control signal and the output displacement of the actuator. However, in actual operation, due to mechanical factors such as gear backlash, lead screw wear, and local lubrication failure, the actuator's transmission mechanism will exhibit complex nonlinear characteristics. This nonlinearity is particularly manifested in local fault phenomena such as jamming or loosening at specific positions. The actuator itself cannot identify and compensate for this local nonlinear error, resulting in a decrease in its control accuracy across the entire range. At the same time, since the actuator's movement is continuous, the displacement error generated after its transmission mechanism passes through the fault point will accumulate and be transmitted to all subsequent target positions, forming a cumulative error effect, ultimately resulting in low overall control accuracy of the actuator. Summary of the Invention

[0004] To address the aforementioned technical problems, an intelligent control method for electric actuators is provided to resolve existing issues.

[0005] The solution to the technical problem of this application is to provide an intelligent control method for an electric actuator, including the following steps:

[0006] Start the actuator's calibration mode, control the actuator to run from zero stroke to full stroke, and obtain the actual displacement and theoretical displacement corresponding to different strokes;

[0007] Based on the difference between actual and theoretical displacement, we evaluate its performance within the allowable control accuracy range and determine the correction error for each range; we analyze the rate of change of the correction error for different ranges to obtain the error change rate, screen the ranges, and identify outlier and normal points.

[0008] Collect multiple actuators of the same model, analyze the deviation of the actual displacement rate of a single actuator from the normal point at the abnormal point, and calculate the relative change deviation at each abnormal point; based on the mapping relationship between the relative change deviation of the abnormal point and the error change rate under all actuators, construct a compensation mapping model.

[0009] For the target actuator to be calibrated, based on the error change rate of the outlier, the relative change deviation of the range corresponding to the outlier is predicted using a compensation mapping model, and the compensation coefficient of each range is determined. Based on the mapping relationship between the compensation coefficient and the theoretical displacement, a compensation function is generated. Combined with the linear change of the range with respect to the theoretical displacement, the range corresponding to the target displacement is calibrated, and a calibrated PWM control command is generated to drive the target actuator.

[0010] Preferably, the correction error for each range includes:

[0011] The difference between the actual displacement and the theoretical displacement corresponding to each range is taken as the deviation, and the difference between the absolute value of the deviation and the preset control accuracy is taken as the control error.

[0012] If the control error is greater than 0, the correction error corresponding to each range is the control error, and its sign is the same as the deviation. Otherwise, the correction error is assigned a value of 0.

[0013] Preferably, obtaining the error change rate includes: performing curve fitting on the correction error of all ranges, and calculating the derivative value of the fitted curve at each range as the error change rate of each range.

[0014] Preferably, the acquisition of abnormal points and normal points includes: marking the range where the error change rate is equal to 0 as a normal point, and vice versa, marking it as an abnormal point.

[0015] Preferably, the calculation of the relative change deviation of each outlier point includes:

[0016] Each range and its corresponding actual displacement are combined into a two-dimensional array. Curve fitting is performed on all two-dimensional arrays to obtain the fitting function of the actual displacement with respect to the range. The derivative of the function is then calculated, and the derivative at each range is taken as the rate of change of displacement for each range.

[0017] The relative change deviation is the difference between the displacement change rate of the range corresponding to each abnormal point and the average displacement change rate of the range corresponding to all normal points.

[0018] Preferably, the method for constructing the compensation mapping model is as follows: obtaining the static features of each actuator; taking each abnormal point of each actuator as a sample, and forming a training set by combining the error change rate, relative change deviation and the static features of the actuator to which each sample belongs, wherein the relative change deviation is the label of the sample, and training the neural network model as the compensation mapping model.

[0019] Preferably, determining the compensation coefficient for each range includes: assigning a value of 0 to the relative change deviation of the target actuator at each normal point corresponding to the range, then the compensation coefficient of the target actuator at each abnormal point corresponding to the range is the predicted relative change deviation, the sign of which is opposite to the error change rate, while the compensation coefficient for each normal point corresponding to the range is 0.

[0020] Preferably, the generation compensation function includes:

[0021] Piecewise fitting is performed on the compensation coefficients of the target actuator to be calibrated across all ranges to obtain the fitting function;

[0022] The linear functions of different ranges with respect to the theoretical displacement are denoted as the theoretical mapping function. Using the theoretical mapping function, the fitting function of the compensation coefficient with respect to the range is transformed into a function of the compensation coefficient with respect to the theoretical displacement, which is then used as the compensation function.

[0023] Preferably, the segmented fitting specifically involves: performing local curve fitting on the compensation coefficients within the range intervals formed by continuous normal points and continuous abnormal points, respectively.

[0024] Preferably, the calibration of the range corresponding to the target displacement includes: integrating the derivative of the theoretical mapping function and the sum of the compensation function over the period from the minimum theoretical displacement to the target displacement, and using this as the calibrated range corresponding to the target displacement.

[0025] This application has at least the following beneficial effects:

[0026] This application calculates the correction error for each range by checking whether the deviation between the actual and theoretical displacement is within the allowable range. Its advantage lies in identifying and locating ranges exceeding the allowable control accuracy, facilitating subsequent screening of abnormal locations where faults occur. Obtaining the error change rate allows for range screening, identifying abnormal and normal points. Its advantage lies in accurately capturing the specific location of momentary jamming or loosening of the transmission mechanism and effectively eliminating interference from accumulated errors, improving the accuracy of abnormal location identification. Calculating the relative change deviation of each abnormal point considers the deviation of the displacement change rate from the normal point, quantifying the additional control required to overcome specific mechanical defects at the abnormal point. Constructing a compensation mapping model allows learning the mapping relationship between the error change rate and the relative change deviation from a large number of actuators of the same model. It can cover the entire measurement range, forming a continuous mapping relationship; predict the relative change deviation of the range, determine the compensation coefficient of each range, and generate a compensation function. Its beneficial effect is that by using the magnitude of the relative change deviation and the sign of the error change rate, it can clearly determine whether the PWM control quantity needs to be increased or decreased for each range, so as to accurately quantify the compensation magnitude, and thus form a continuous compensation function from discrete points, realizing the continuity and smoothness of control commands; calibrate the range corresponding to the target displacement, and generate calibrated PWM control commands to drive the target actuator. Its beneficial effect is that by embedding the compensation function into integral calculation, it actively calculates the control quantity that needs to be compensated based on the target displacement, rather than passively correcting after the error occurs. This feedforward compensation mechanism enables the actuator to actively cancel nonlinear errors, realize precise positioning control across the entire range, and significantly improve the control accuracy of the actuator. Attached Figure Description

[0027] The following detailed description of an intelligent control method for an electric actuator according to the present application is provided in conjunction with the accompanying drawings.

[0028] Figure 1 A flowchart illustrating the steps of an intelligent control method for an electric actuator provided in this application embodiment;

[0029] Figure 2 A flowchart illustrating the steps of the method for obtaining the compensation function provided in this application embodiment. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of an intelligent control method for an electric actuator, in conjunction with the accompanying drawings and embodiments, is provided. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0031] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0032] Please see Figure 1 The diagram illustrates a flowchart of a smart control method for an electric actuator according to an embodiment of this application. The method includes the following steps:

[0033] Step 1: Start the actuator calibration mode, control the actuator to run from zero stroke to full stroke, and obtain the actual displacement and theoretical displacement corresponding to different strokes.

[0034] Electric actuators are key terminal actuators in industrial automation systems. They receive control commands, such as PWM frequency ratios, and convert them into precise linear or angular displacement outputs, thereby driving transmission mechanisms such as valves, dampers, and baffles to achieve continuous control of process parameters such as flow rate, pressure, and temperature.

[0035] Due to factors such as motor speed drift, transmission mechanism wear, gear backlash, and load variations, errors can accumulate in the mechanical and electronic components of electric actuators, leading to inaccurate control of the transmission mechanism and consequently, inaccurate control of the final actuators such as valves and baffles. Therefore, electric actuators need to be calibrated.

[0036] Traditional calibration methods calibrate actuators by analyzing the error between the actual and theoretical displacement at a specific range position and, based on the linear relationship between range and displacement, calculating the displacement error at other range positions and compensating accordingly. For example, actuator calibration is typically performed by analyzing the error between the actual and theoretical positions at zero or full range. However, if the actuator's transmission mechanism is aged in some areas, this linear relationship will be disrupted. For instance, if the actuator's transmission mechanism is aged and causes jamming within the range of 20% to 30%, resulting in local nonlinearity, it may still function normally within the remaining range. Once the actuator operates beyond 30% of its range, it will pass through these aged or jammed positions, accumulating errors and leading to inaccurate actuator calibration.

[0037] Electric actuators are divided into linear actuators and angle actuators. In this embodiment, an angle actuator is used as an example. Through PWM frequency ratio, the motor is driven and, via a gear transmission mechanism, outputs a precise angular displacement, thereby controlling the valve opening. Therefore, the subsequent displacement refers to the angular displacement.

[0038] In this embodiment, the angular displacement range of the electric actuator is: Therefore, ideally, the actuator's PWM frequency ratio, i.e., its range, has a strictly linear relationship with the theoretical displacement of the output shaft. When the PWM frequency ratio is 0, its range is defined as zero, and the theoretical displacement of the output shaft is 0°, at which point the valve baffle is in the fully closed position. When the PWM frequency ratio is 1, its range is defined as full range, and the theoretical displacement of the output shaft is... At this point, the valve baffle should be fully open.

[0039] Therefore, for any PWM frequency ratio ,in, The corresponding theoretical displacement The linear equation is:

[0040]

[0041] in, It can represent the PWM frequency ratio, or it can be called the range.

[0042] Based on the above analysis, the calibration mode of the electric actuator is activated, and the electric actuator is controlled to run from zero stroke to full stroke, and the actual displacement corresponding to each stroke is recorded.

[0043] In this embodiment, the interval between two adjacent ranges is 0.05. As for other implementation methods, the implementer can set it according to the actual situation.

[0044] Therefore, based on the linear relationship between the PWM frequency ratio and the theoretical displacement, the theoretical displacement corresponding to each range is obtained;

[0045] Thus, the actual and theoretical displacements of the electric actuator for each range are obtained.

[0046] Step 2: Based on the difference between the actual displacement and the theoretical displacement, assess its performance within the allowable control accuracy range and determine the correction error for each range; analyze the rate of change of the correction error for different ranges to obtain the error change rate, screen the ranges, and obtain abnormal points and normal points.

[0047] Furthermore, the deviation between the actual displacement and the theoretical displacement is analyzed to determine whether it is within the allowable range, in order to identify and quantify significant errors exceeding the inherent accuracy of the actuator caused by aging or damage to the mechanical transmission mechanism. Specifically:

[0048] The difference between the actual displacement and the theoretical displacement corresponding to each range is taken as the deviation.

[0049] The difference between the absolute value of the deviation and the preset control precision is taken as the control error;

[0050] If the control error is greater than 0, the correction error corresponding to each range is the control error, and its sign is the same as the deviation. Otherwise, the correction error is assigned a value of 0.

[0051] In this embodiment, the formula for calculating the correction error corresponding to each range is:

[0052]

[0053] in, Indicates the correction error. Indicates the deviation amount. This indicates the preset control precision. Represents a symbolic function. express Activation function Indicates control error;

[0054] Among them, when ,but ;when ,but ;when ,but ;and Activation function means that, given an element , The activation function is defined as the maximum value between an element and 0; the specific function is as follows:

[0055]

[0056] Among them, when element When less than or equal to 0, The activation function outputs 0 when the element When greater than 0, The activation function output is .

[0057] It should be noted that the preset control precision is set according to the product specifications of the electric actuator. In this embodiment, the control precision of the actuator is [insert value here]. Therefore, the preset control precision value is As another implementation method, the implementer can set it according to the actual situation.

[0058] It should be noted that the larger the deviation, the greater the difference between the actual displacement and the ideal displacement of the actuator, reflecting a larger overall control error of the actuator; the larger the control error, the more serious the displacement error of the actuator exceeds the normal allowable range, and the higher the risk of control failure; if the displacement error of the actuator is within the allowable range, it means that the actuator is in normal working condition and does not need to be calibrated for this range; if the displacement error of the actuator exceeds the allowable range, it means that there may be transmission loosening or aging blockage in the actuator at this range position, and the actuator needs to be calibrated.

[0059] Secondly, if the actuator transmission mechanism is in a normal state at a certain range position, the correction error corresponding to that range should remain stable. Even if there is a cumulative error at that range position, as long as the correction error does not fluctuate drastically with changes in range, it can be concluded that the transmission mechanism is not affected by nonlinear mechanical resistance when passing through that range position. Based on this analysis, by analyzing the changes in correction error under each range, abnormal and normal ranges are identified, specifically:

[0060] Curve fitting is performed on the correction error for all ranges, and the derivative of the fitted curve at each range is calculated as the error change rate.

[0061] In this embodiment, the least squares method is used for curve fitting. The least squares method is a well-known technique and will not be described in detail here. Secondly, the calculation of the derivative is a well-known technique and will not be described in detail here.

[0062] The range where the error rate of change is equal to 0 is marked as a normal point, and the range where the error rate of change is equal to 0 is marked as an abnormal point.

[0063] It should be noted that an error change rate of 0 indicates that the correction error at that range position is relatively stable, reflecting that no calibration is needed for that range position; an error change rate greater than 0 indicates that the correction error at that range position is on an upward trend, and the actual displacement of the actuator at that range position is deviating ahead of the theoretical displacement in a manner exceeding the accuracy range, reflecting that the transmission mechanism is loose at that range position; an error change rate less than 0 indicates that the correction error at that range position is on a downward trend, and the actual displacement of the actuator at that range position is deviating behind the theoretical displacement in a manner exceeding the control accuracy range, reflecting that the transmission mechanism is experiencing jamming due to aging and blockage at that range position.

[0064] The range corresponding to the normal point does not require displacement calibration, while the range corresponding to the abnormal point does require displacement calibration.

[0065] At this point, we have obtained the outliers and normal points.

[0066] Step 3: Collect multiple actuators of the same model, analyze the deviation of the actual displacement rate of a single actuator from the normal point at the abnormal point, and calculate the relative change deviation of each abnormal point; based on the mapping relationship between the relative change deviation of the abnormal points of all actuators and the error change rate, construct a compensation mapping model.

[0067] Furthermore, based on the ideal state, the range of the electric actuator is a linear function of the theoretical displacement, specifically expressed as:

[0068]

[0069] in, The dependent variable is the range. This represents the theoretical displacement at full stroke, i.e., the maximum theoretical displacement. This is the theoretical displacement corresponding to zero stroke, i.e., the minimum theoretical displacement. is the independent variable, representing the theoretical displacement;

[0070] Since full range represents 1 and zero range represents 0, therefore, This represents a linear function.

[0071] In this embodiment, equal , It is 0°, therefore, .

[0072] The linear function is denoted as the theoretical mapping function, and its derivative is obtained by taking the derivative of the theoretical mapping function.

[0073] It should be noted that, since the derivative of the theoretical mapping function is... It is a constant value.

[0074] Accordingly, based on the actual displacement corresponding to each range, a practical mapping function of the range with respect to the actual displacement is constructed, specifically as follows:

[0075] Each range and its corresponding actual displacement are combined into a two-dimensional array. Curve fitting is performed on all two-dimensional arrays to obtain the fitting function of the range with respect to the actual displacement as the actual mapping function.

[0076] In the actual mapping function, the range is the dependent variable and the actual displacement is the independent variable.

[0077] In this embodiment, the least squares method is used for curve fitting. The least squares method is a well-known technique and will not be described in detail here.

[0078] The derivative of the inverse function of the actual mapping function is taken, and the derivative at each range is taken as the displacement change rate of each range.

[0079] It should be noted that the inverse function is a well-known technique and will not be elaborated here. The inverse function represents the actual displacement as a function of the range, where the independent variable of the inverse function is the range and the dependent variable is the actual displacement. Alternatively, curve fitting can be performed on all two-dimensional arrays to obtain the fitting function of the actual displacement as a function of the range. The derivative of this function can be directly calculated to obtain the rate of change of displacement at each range, without the need to calculate the inverse function.

[0080] Calculate the average displacement change rate for all normal points corresponding to the range;

[0081] It should be noted that if the actuator is in an ideal state, the rate of displacement change is constant, and therefore the average value is also constant.

[0082] The difference between the displacement change rate of each abnormal point and the average value is taken as the relative change deviation of each abnormal point.

[0083] In this implementation, the absolute value of the difference between the displacement change rate of each abnormal point and the average value is taken as the relative change deviation of each abnormal point, while the relative change deviation of each normal point is assigned a value of 0.

[0084] It should be noted that the average value reflects the average level of the actual displacement change rate of the actuator within its healthy and normal range. The greater the relative change deviation, the further the instantaneous working state of the actuator transmission mechanism deviates from its normal healthy state at the abnormal point, reflecting the severity of the disturbance to the mechanical transmission mechanism at that abnormal point.

[0085] Data from multiple actuators of the same model were collected. Following the above process, the error change rate and relative deviation of each actuator at each abnormal point were calculated. The service life, average operating temperature, average operating humidity and aging index of each actuator were obtained as static characteristics of each actuator.

[0086] It should be noted that the aging index is measured by the number of failures.

[0087] Each outlier of each actuator is treated as a sample. The error change rate, relative change deviation and the static characteristics of the actuator to which each sample belongs are combined to form a training set, where the relative change deviation is the label of the sample.

[0088] The neural network model is trained based on the training set and used as a compensation mapping model;

[0089] In this embodiment, a Long Short-Term Memory (LSTM) network model is used for training. The LSM network model is a well-known technology and will not be described in detail here. As other implementation methods, implementers may use other methods of the prior art, such as recurrent neural network models. This embodiment does not impose any special restrictions on this.

[0090] It should be noted that, due to the different aging characteristics and varying service life of actuators, different actuators may correspond to different relative deviations even with the same error change rate. By using static features as input, the adaptability and accuracy of the model under different operating conditions can be improved.

[0091] Thus, the compensation mapping model is obtained.

[0092] Step 4: For the target actuator to be calibrated, based on the error change rate of each range, the relative change deviation of the range is predicted using the compensation mapping model, and the compensation coefficient of each range is determined. Based on the mapping relationship between the compensation coefficient and the theoretical displacement, a compensation function is generated. Combined with the linear change of the range with respect to the theoretical displacement, the range corresponding to the target displacement is calibrated, and the calibrated PWM control command is generated to drive the target actuator.

[0093] The flowchart of the method for obtaining the compensation function provided in this application embodiment is as follows: Figure 2 As shown.

[0094] Furthermore, for each range of the target actuator to be calibrated, corresponding to the actual displacement and theoretical displacement, the error change rate of the target actuator in each range is obtained;

[0095] The static characteristics of the target actuator to be calibrated and the error change rate of the corresponding range at each abnormal point are used as inputs to the compensation mapping model to obtain the predicted relative change deviation; wherein, the relative change deviation of the target actuator at each normal point is assigned to 0.

[0096] The compensation coefficient of the target actuator to be calibrated at each abnormal point is the predicted relative change deviation, and its sign is opposite to the error change rate, while the compensation coefficient of each normal point is 0.

[0097] It should be noted that if the error change rate is less than 0, it indicates that there is a jamming phenomenon caused by aging and blockage in the transmission mechanism at this range position. Therefore, in order for the transmission mechanism to reach the corresponding displacement, the range needs to be increased, that is, the PWM frequency ratio needs to be increased. At this time, the compensation coefficient is a positive value.

[0098] Piecewise fitting is performed on the compensation coefficients of the target actuator to be calibrated across all ranges to obtain the fitting function;

[0099] In this embodiment, cubic spline interpolation is used for fitting. Cubic spline interpolation is a well-known technique and will not be described in detail here. The specific process of piecewise fitting is as follows: the range contained in the range corresponding to all consecutive abnormal points is taken as a range interval, and the range contained in the range corresponding to all consecutive normal points is taken as a range interval. Therefore, the entire range is divided into multiple range intervals, and local curve fitting is performed on the compensation coefficient in each range interval.

[0100] Based on the linear relationship between the range and the theoretical displacement, the fitting function is transformed into a function of the compensation coefficient with respect to the theoretical displacement, which serves as the compensation function.

[0101] It should be noted that the fitted function is assumed to be... , where is the compensation coefficient as a function of the measurement range, since the linear relationship between the measurement range and the theoretical displacement is... Therefore, the fitting function It can be transformed into a discussion of theoretical displacement. The compensation function will fit the function. The range in Change to Thus, the compensation coefficient with respect to the theoretical displacement is obtained. The compensation function.

[0102] The actuator stores a compensation function. In the actuator's normal operating mode, when it is necessary to control the actuator to move to a target displacement, the range of the electric actuator is calibrated using the compensation function. Specifically:

[0103] The derivative of the theoretical mapping function and the sum of the compensation function are integrated over the period from the minimum theoretical displacement to the target displacement, and this integral is used as the calibrated range corresponding to the target displacement.

[0104] The specific formula is as follows:

[0105]

[0106] in, The target displacement is The corresponding calibrated range The derivative function, i.e. , For the compensation function, This is the theoretical displacement corresponding to zero stroke, i.e., the minimum theoretical displacement. For the target displacement, Represents the integral variable;

[0107] The PWM frequency ratio corresponding to the calibrated range is used as the final control command to drive and control the electric actuator.

[0108] It should be noted that, It is a constant value; Indicates in To compensate for mechanical transmission errors at a certain displacement position, additional PWM control quantity, i.e., range, needs to be added or reduced on the ideal reference. When the result of the compensation function at a certain displacement position is greater than 0, it means that PWM needs to be added at that point to overcome the jamming phenomenon. When the result of the compensation function at a certain displacement position is less than 0, it means that PWM needs to be added at that point to overcome the jamming phenomenon. When the result of the compensation function at a certain displacement position is equal to 0, it means that the transmission at that displacement position is normal and no compensation is needed.

[0109] It should be noted that target displacement refers to the position that the target actuator is required to reach under normal operating conditions.

[0110] Through the above process, the electric actuator is calibrated and controlled in real time, ensuring that the actuator's response to each target displacement command remains highly accurate even when there are nonlinear errors in the transmission mechanism.

[0111] It should be understood that, although Figure 1 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0112] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0113] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application. Therefore, any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of this application, without departing from the content of the technical solution of this application, shall fall within the protection scope of the technical solution of this application.

Claims

1. A method for intelligent control of an electric actuator, characterized in that, The method includes the following steps: Start the actuator's calibration mode, control the actuator to run from zero stroke to full stroke, and obtain the actual displacement and theoretical displacement corresponding to different strokes; Based on the difference between actual and theoretical displacement, we evaluate its performance within the allowable control accuracy range and determine the correction error for each range; we analyze the rate of change of the correction error for different ranges to obtain the error change rate, screen the ranges, and identify outlier and normal points. Collect multiple actuators of the same model, analyze the deviation of the actual displacement rate of a single actuator from the normal point at the abnormal point, and calculate the relative change deviation at each abnormal point; based on the mapping relationship between the relative change deviation of the abnormal point and the error change rate under all actuators, construct a compensation mapping model. For the target actuator to be calibrated, based on the error change rate of the outlier, the relative change deviation of the range corresponding to the outlier is predicted using a compensation mapping model, and the compensation coefficient of each range is determined. Based on the mapping relationship between the compensation coefficient and the theoretical displacement, a compensation function is generated. Combined with the linear change of the range with respect to the theoretical displacement, the range corresponding to the target displacement is calibrated, and a calibrated PWM control command is generated to drive the target actuator.

2. The intelligent control method for an electric actuator as described in claim 1, characterized in that, The correction error for each range includes: The difference between the actual displacement and the theoretical displacement corresponding to each range is taken as the deviation, and the difference between the absolute value of the deviation and the preset control accuracy is taken as the control error. If the control error is greater than 0, the correction error corresponding to each range is the control error, and its sign is the same as the deviation. Otherwise, the correction error is assigned a value of 0.

3. The intelligent control method for an electric actuator as described in claim 1, characterized in that, The method for obtaining the error change rate includes: performing curve fitting on the correction error of all ranges, calculating the derivative value of the fitted curve at each range, and using it as the error change rate for each range.

4. The intelligent control method for an electric actuator as described in claim 1, characterized in that, The process of obtaining abnormal and normal points includes: marking the range where the error change rate is equal to 0 as a normal point, and vice versa, marking it as an abnormal point.

5. The intelligent control method for an electric actuator as described in claim 1, characterized in that, The calculation of the relative change deviation of each outlier point includes: Each range and its corresponding actual displacement are combined into a two-dimensional array. Curve fitting is performed on all two-dimensional arrays to obtain the fitting function of the actual displacement with respect to the range. The derivative of the function is then calculated, and the derivative at each range is taken as the rate of change of displacement for each range. The relative change deviation is the difference between the displacement change rate of the range corresponding to each abnormal point and the average displacement change rate of the range corresponding to all normal points.

6. The intelligent control method for an electric actuator as described in claim 1, characterized in that, The method for constructing the compensation mapping model is as follows: obtain the static features of each executor; Each abnormal point of each actuator is treated as a sample. The error change rate, relative change deviation and the static characteristics of the actuator to which each sample belongs are used to form a training set, where the relative change deviation is the label of the sample. The neural network model is then trained as a compensation mapping model.

7. The intelligent control method for an electric actuator as described in claim 1, characterized in that, The determination of the compensation coefficient for each range includes: if the relative change deviation of the target actuator to be calibrated at each normal point corresponding to the range is assigned to 0, then the compensation coefficient of the target actuator to be calibrated at each abnormal point corresponding to the range is the predicted relative change deviation, the sign of which is opposite to the error change rate, while the compensation coefficient for each normal point corresponding to the range is 0.

8. The intelligent control method for an electric actuator as described in claim 1, characterized in that, The generation compensation function includes: Piecewise fitting is performed on the compensation coefficients of the target actuator to be calibrated across all ranges to obtain the fitting function of the compensation coefficients with respect to the range. The linear functions of different ranges with respect to the theoretical displacement are denoted as the theoretical mapping function. Using the theoretical mapping function, the fitting function of the compensation coefficient with respect to the range is transformed into a function of the compensation coefficient with respect to the theoretical displacement, which is then used as the compensation function.

9. The intelligent control method for an electric actuator as described in claim 8, characterized in that, The segmented fitting specifically involves: performing local curve fitting on the compensation coefficients within the range intervals formed by continuous normal points and continuous abnormal points, respectively.

10. The intelligent control method for an electric actuator as described in claim 8, characterized in that, The calibration of the range corresponding to the target displacement includes: integrating the derivative of the theoretical mapping function and the sum of the compensation function over the period from the minimum theoretical displacement to the target displacement, and using this as the calibrated range corresponding to the target displacement.

Images