Encoder-based stroke control method for coal mine safety electric actuators
By using encoder-based bidirectional pulse classification and current monitoring, the mechanical backlash estimate is updated in real time, solving the positioning deviation problem caused by mechanical backlash wear and achieving high-precision electric actuator stroke control.
Patent Information
- Application Number
- CN202611161615.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-03
- Publication Date
- 2026-11-17
AI Technical Summary
In the prior art, the width of the mechanical clearance changes due to wear over time, causing the fixed clearance compensation parameter to fail, which affects the accuracy of the encoder count value and the reliability of the clearance crossing time identification.
An encoder-based calculation method is adopted, which updates the mechanical clearance estimate in real time by bidirectionally classifying encoder pulses and monitoring motor drive current. The method uses an exponentially weighted moving average to track the clearance width change and adaptively adjusts the clearance through current threshold.
It enables real-time tracking of the valve stem's actual position, eliminates the empty stroke deviation introduced by the clearance, improves positioning accuracy, and enables the compensation parameters to track the wear changes of the mechanical clearance, ensuring reliable determination of the clearance crossing time.
Smart Images

Figure CN123113447A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of electric actuator control technology, and in particular to a stroke control method for coal safety electric actuators based on encoder calculation. Background Technology
[0002] In electric valve actuators, a motor drives the valve stem to reciprocate via a gear transmission chain. There is mechanical backlash between the gears; when the motor reverses direction, a portion of the pulses accumulated by the encoder corresponds to the idle stroke within the backlash range, resulting in no effective displacement of the valve stem.
[0003] Existing technologies typically use fixed gap compensation parameters to correct encoder count values in order to eliminate positioning deviations after commutation.
[0004] However, the above method has the following problems: First, the width of the mechanical clearance changes slowly due to wear over time, and the fixed clearance compensation parameters will gradually become ineffective, resulting in a decrease in compensation accuracy; Second, the electrical characteristics of the motor also drift over time. If the current threshold used to determine the completion of clearance crossing remains unchanged, the judgment criteria will gradually deviate from the actual working conditions, affecting the reliability of the clearance crossing moment identification. Summary of the Invention
[0005] The purpose of this invention is to propose a stroke control method for coal safety electric actuators based on encoder calculation in order to solve the above-mentioned problems.
[0006] To achieve the above objectives, the present invention adopts the following technical solution:
[0007] A stroke control method for coal safety electric actuators based on encoder calculations includes:
[0008] The encoder pulse signal is collected. When the motion direction is detected to be reversed, the Boolean flag in the gap interval state variable is set to true, the consumed gap depth is set to zero, the motion direction mark is recorded, and the motor drive current time series is collected simultaneously.
[0009] When the Boolean flag is true, subsequent encoder pulses are classified bidirectionally. If the pulse direction is consistent with the motion direction mark, the consumed gap depth is increased by one pulse equivalent; otherwise, it is decreased by one pulse equivalent.
[0010] Monitor the motor drive current. When the current change exceeds the gap crossing current threshold, it is determined that the gap crossing is complete. Set the Boolean flag to false and update the mechanical gap estimate with the gap crossing amount using the exponential weighted moving average method.
[0011] In response to the target stroke command, the encoder target pulse number is calculated based on the stroke difference between the actual estimated position of the valve stem and the target position, as well as the compensation amount determined based on the direction of motion, and the motor is driven to perform positioning.
[0012] Preferably, the method of updating the mechanical clearance estimate using the current clearance crossing amount using the exponentially weighted moving average method specifically involves:
[0013] Let the first The amount of passage through the gap is Updated mechanical clearance estimates Calculated by the following formula:
[0014] ;
[0015] in, This is the estimated mechanical clearance value from the last update. For smoothing coefficients; , , All units are encoder pulse equivalents.
[0016] Preferably, the actual estimated position of the valve stem is equal to the encoder's cumulative count value minus the invalid stroke offset corresponding to the consumed clearance depth;
[0017] The compensation amount determined based on the direction of motion includes: if the required direction of motion is consistent with the direction of motion marker within the current gap interval, the compensation amount is the difference between the estimated mechanical gap value and the consumed gap depth; if the required direction of motion is in the opposite direction, the compensation amount is the value of the consumed gap depth.
[0018] Preferably, the bidirectional classification process further includes: when the consumed gap depth decreases to zero, re-recording the current flip direction as a new motion direction marker, and resetting the consumed gap depth to zero to start a new round of gap crossing tracking.
[0019] Preferably, when the Boolean flag is false, the consumed gap depth is zero. In this case, the compensation amount in the same direction is equal to the estimated mechanical gap value, and the compensation amount in the opposite direction is zero.
[0020] Preferably, the method further includes:
[0021] A four-dimensional feature parameter vector is extracted from the current time series of this gap crossing process. The four-dimensional feature parameter vector includes the mean no-load current in the gap interval, the peak value of the current change at the time of crossing completion, the current change rise slope, and the current transition time from no-load to load. The four-dimensional feature parameter vector is appended with a crossing timestamp and stored in the gap crossing current feature history sequence.
[0022] Preferably, the method further includes:
[0023] Once the number of records in the historical sequence of the gap current characteristics reaches the minimum number of regression samples, linear regression fitting is performed on the four feature parameters in the four-dimensional feature parameter vector to obtain the drift direction and drift rate of each feature parameter with the cumulative running time; wherein, the first The linear model for the characteristic parameters is:
[0024] ;
[0025] in, For cumulative running time, For the first The drift rate of each characteristic parameter, whose sign indicates the drift direction. The intercept is... and It is obtained by solving the least squares method.
[0026] Preferably, the method further includes:
[0027] Based on the linear models corresponding to the mean no-load current and the peak current surge, linear extrapolation is performed at the time point of the next expected operating period to obtain the predicted no-load current values. and mutation peak prediction value Updated gap crossing current threshold Calculated by the following formula:
[0028] ;
[0029] in, The preset safety factor is a dimensionless positive real number.
[0030] Preferably, the method further includes:
[0031] For each newly added four-dimensional feature parameter vector of the gap crossing current feature history sequence, calculate the residuals of its four feature parameter values relative to the corresponding linear regression trend line; when the residual of any feature parameter exceeds three times the standard deviation of the feature parameter residual sequence, mark the record as an abnormal record and remove it from the gap crossing current feature history sequence; after removing the abnormal record, re-perform linear regression fitting on the remaining records to update the drift direction and drift rate.
[0032] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:
[0033] 1. This invention performs bidirectional classification and counting of encoder pulses after commutation using gap interval state variables, tracks the consumed gap depth in real time, and can obtain the deviation between the actual estimated position of the valve stem and the cumulative count value of the encoder at any time. Based on the relationship between the target movement direction and the current gap crossing state, compensation amounts are calculated separately for both unidirectional and reverse scenarios, thereby eliminating the empty travel deviation introduced by the gap when the positioning command is executed.
[0034] 2. This invention determines the physical moment when the gap crossing is completed by detecting sudden changes in current, and continuously updates the mechanical gap estimate using an exponentially weighted moving average method, so that the compensation parameters can track the slow changes in gap width caused by wear, and avoid the fixed compensation parameters from gradually becoming ineffective over time. Attached Figure Description
[0035] Further details, features, and advantages of this application are disclosed in the following description of exemplary embodiments in conjunction with the accompanying drawings, in which:
[0036] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation
[0037] Several embodiments of this application will now be described in more detail with reference to the accompanying drawings to enable those skilled in the art to implement this application. This application may be embodied in many different forms and for various purposes and should not be limited to the embodiments set forth herein. These embodiments are provided to make this application thorough and complete, and to fully convey the scope of this application to those skilled in the art. The embodiments described do not limit this application.
[0038] Unless otherwise defined, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It will be further understood that terms such as those defined in commonly used dictionaries shall be interpreted as having a meaning consistent with their meaning in the relevant field and / or the context of this specification, and shall not be interpreted in an idealized or overly formal sense unless expressly defined herein.
[0039] Example 1
[0040] Its specific implementation method is combined with the appendix Figure 1 Please provide a detailed explanation.
[0041] In this embodiment, it includes:
[0042] Step 1: Acquire encoder pulse signals. When a reversal of the motion direction is detected, initialize the state variables of the gap interval and synchronously acquire the time series of the motor drive current.
[0043] The controller continuously acquires encoder pulse signals and monitors the direction of motion. When a reversal of the direction of motion is detected, the Boolean flag in the gap interval state variable is set to true, the consumed gap depth is reset to zero, and the motion direction marker when entering the gap interval is recorded. Simultaneously, the time series of the motor drive current after the reversal is acquired at a high sampling rate.
[0044] It should be noted that the gap interval state variable contains three components: a Boolean flag, the consumed gap depth, and a motion direction marker. A true Boolean flag indicates that the motor is currently within the gap interval, meaning the motor has reversed direction but the gear pair has not yet completed meshing. The consumed gap depth, expressed in encoder pulse equivalents, represents the number of valid cross-pass pulses accumulated by the encoder after the commutation. The motion direction marker records the commutation direction when entering the gap interval, used for classifying the direction of subsequent pulses.
[0045] Step 2: Perform bidirectional classification processing on subsequent encoder pulses within the gap interval to obtain the consumed gap depth.
[0046] When the Boolean flag for the gap interval is true, each subsequent encoder pulse undergoes bidirectional classification. If the pulse direction matches the motion direction marker when entering the gap interval, the consumed gap depth is incremented by one pulse equivalent. If the pulse direction is opposite, the consumed gap depth is decremented by one pulse equivalent.
[0047] It should be noted that when the consumed gap depth decreases to zero, it indicates that the motor has reversed again within the gap interval and returned to the gap start position. At this time, the current reversal direction is re-recorded as the new motion direction marker, and the consumed gap depth is reset to zero, starting a new round of gap crossing tracking. This process ensures that when multiple reversals occur within the gap interval, the consumed gap depth always reflects the net crossing amount relative to the most recent gap start position.
[0048] Step 3: Monitor the motor drive current. When the sudden change in current exceeds the gap crossing current threshold, determine that the gap crossing is complete and update the mechanical gap estimate.
[0049] The motor drive current is continuously monitored within the gap interval. When the sudden change in current exceeds the currently valid gap crossing current threshold, the gap crossing is considered complete. The gap interval Boolean flag is set to false, and the gap depth consumed at this time is recorded as the gap crossing amount. Using the exponentially weighted moving average method, with the gap crossing amount as input, the updated mechanical gap estimate is output.
[0050] It should be noted that the determination of the completion of gap crossing is based on the magnitude of the current surge. Within the gap range, the motor-driven gear moves within the unloaded gap space, and the drive current is at an unloaded level. When the gear pair re-engages, a load is suddenly applied to the motor, causing a step increase in the drive current. This current surge marks the physical completion moment of gap crossing. Using this as the determination criterion allows for the precise timing of gap crossing completion, rather than relying solely on the estimation of encoder pulse count.
[0051] The exponentially weighted moving average method assigns higher weight to recent clearance crossings, enabling the mechanical clearance estimate to track the slow changing trend of clearance width due to wear, while also smoothing out noise from individual measurements. Let the... The amount of passage through the gap is Updated mechanical clearance estimates Calculated by the following formula:
[0052] ;
[0053] in, This is the sequence number of the current gap crossing. This is the estimated mechanical clearance value from the last update. For smoothing coefficients, The larger the value, the more sensitive it is to recent crossing volumes. The smaller the value, the stronger the smoothing effect. , , The units are all encoder pulse equivalents, and the dimensions are consistent.
[0054] Step 4: Responding to the target stroke command, calculate the compensated encoder target pulse count based on the actual estimated position of the valve stem and the estimated mechanical backlash, and drive the motor to perform positioning.
[0055] Upon receiving a target stroke command, the stroke difference between the target position and the actual estimated position of the valve stem is calculated. The actual estimated position of the valve stem is equal to the encoder's cumulative count value minus the invalid stroke offset corresponding to the consumed clearance depth.
[0056] Determine the required direction of motion to reach the target position. If the required direction of motion matches the direction marked within the current clearance interval, the required compensation is the difference between the estimated mechanical clearance and the already consumed clearance depth. If the required direction of motion is in the opposite direction, the required compensation is the value of the already consumed clearance depth. Add the travel difference to the corresponding direction's compensation, convert it into the encoder's target pulse count, and drive the motor to perform positioning.
[0057] It should be noted that the above compensation calculation distinguishes between two scenarios. When the required motion direction is consistent with the motion direction within the gap interval, the depth consumed in the gap remains valid in subsequent motions in the same direction. Only the remaining uncrossed gap portion needs to be compensated, i.e., the difference between the estimated mechanical gap value and the consumed gap depth. When the required motion direction is reversed, the encoder pulse corresponding to the consumed gap depth will be retracted in the initial stage of the reverse motion. This part of the stroke does not produce an effective reverse displacement of the valve stem, therefore the consumed gap depth must be included in the compensation amount. The above stroke difference, compensation amount, and encoder target pulse number are all in encoder pulse equivalent units, with consistent dimensions, and can be directly superimposed.
[0058] It should be noted that when determining the required direction of motion in step 4, if the Boolean flag of the current gap interval is false, that is, it is not currently within the gap interval, then the consumed gap depth is zero. In this case, the compensation amount in the same direction is equal to the estimated mechanical gap value, and the compensation amount in the opposite direction is zero. The compensation logic degenerates into standard full gap compensation or no compensation, which is consistent with the general case within the gap interval in terms of calculation form.
[0059] In this embodiment of the application, in order to enable the gap through current threshold to adapt to the drift changes in the electrical characteristics of the motor over time, the following steps are also included:
[0060] Step 5: Extract a four-dimensional feature parameter vector from the current time series of the gap crossing process, add a timestamp, and store it in the gap crossing current feature history sequence.
[0061] Four characteristic parameters were extracted from the current time series of this gap crossing process: the average no-load current within the gap interval, the peak value of the current jump at the moment of crossing completion, the rising slope of the current jump, and the current transition time from no-load to load. The four-dimensional characteristic parameter vector was appended with a crossing timestamp and stored in the gap crossing current characteristic history sequence.
[0062] It should be noted that the mean no-load current is taken as the arithmetic mean of the sampling segment before the current abrupt change occurs within the gap interval. The peak value of the current abrupt change is the first local maximum after the current step rise. The slope of the current abrupt change rise is the rate of change of current during the time interval from the no-load level to the peak value of the abrupt change, i.e., the difference between the peak value of the abrupt change and the mean no-load current divided by the corresponding rise time. The current transition time is the duration during which the current moves from the no-load level to the load level and stabilizes.
[0063] It should be noted that the four feature parameters have different dimensions: the mean no-load current and the peak current surge are measured in amperes, the current surge rise slope is measured in amperes per second, and the current transition time is measured in seconds. In step 6, linear regression fitting is performed independently for each feature parameter. Since the feature parameters do not participate in cross-dimensional mixing operations, there is no need to perform uniform normalization on the four-dimensional feature parameter vector.
[0064] Step 6: Perform linear regression fitting on each feature parameter in the historical sequence of gap current characteristics to obtain the drift direction and drift rate of each feature parameter.
[0065] Once the number of records in the historical sequence of gap crossing current characteristics reaches the minimum number of regression samples, linear regression fitting is performed on each of the four feature parameters. The input to the linear regression is the crossing timestamp of each record and the corresponding feature parameter value, and the output is the drift direction and drift rate of each feature parameter as a function of the cumulative running time.
[0066] It should be noted that, for the first Let the nth feature parameter be the nth feature parameter. The time stamp of the historical record is The corresponding feature parameter value is Linear regression fitting yields the following linear model for this feature parameter with respect to time:
[0067] ;
[0068] in, The index of the feature parameter. This refers to the sequence number in the historical records. For cumulative running time, For the first The drift rate of each characteristic parameter, whose sign indicates the drift direction. This is the intercept. and It is obtained by solving all valid records in the characteristic history sequence of gap current using the least squares method.
[0069] Step 7: Calculate the updated gap ride-through current threshold by linearly extrapolating the mean no-load current and the peak current spike based on the drift rate.
[0070] Based on the linear regression equations for the mean no-load current and the peak current surge, linear extrapolation is performed at the next expected operating time point to obtain the predicted no-load current and the predicted peak current surge. The difference between the predicted peak current surge and the predicted no-load current surge is multiplied by a preset safety factor to obtain the updated gap crossing current threshold.
[0071] It should be noted that the updated gap through current threshold The calculation method is as follows:
[0072] ;
[0073] in, This represents the predicted peak value of mutations during the next expected operating period, in amperes. This is the predicted no-load current for the next expected operating period, in amperes. The preset safety factor is a dimensionless positive real number. and The values are obtained by extrapolating the linear regression equations corresponding to the peak current surge and the mean no-load current at the time point of the next expected operating period, respectively. That is, the target time point is substituted into the linear model of the corresponding characteristic parameters in step 6. Calculated. and The units are the same, the difference is in amperes, multiplied by a dimensionless coefficient. The result after The unit remains ampere, consistent with the dimension of the current surge amplitude used for comparison in step 3.
[0074] It should be noted that among the four feature parameters extracted in step 5, the current surge rise slope and current transition time are involved in the linear regression fitting in step 6. Their drift trend serves as an auxiliary monitoring quantity for the overall change of the motor's electrical characteristics, providing a more complete basis for anomaly judgment in the residual detection in step 8. The gap crossing current threshold calculation in step 7 only uses the mean no-load current and the peak value of the current surge. This is because the physical meaning of the gap crossing current threshold is the judgment benchmark for the current surge amplitude, which is directly determined by the difference between the surge peak and the mean no-load current. The rise slope and transition time do not directly participate in the amplitude calculation.
[0075] In this embodiment of the application, in order to eliminate the influence of outlier data caused by transient interference or abnormal operating conditions on the accuracy of linear regression fitting, the following steps are also included in step 6:
[0076] Step 8: Perform residual detection on the newly added gap ride-through current characteristic parameter vector and remove abnormal records.
[0077] For each newly added gap crossing current characteristic history sequence, the residuals of the four characteristic parameter vectors relative to the corresponding linear regression trend line are calculated. When the residual of any characteristic parameter exceeds three times the standard deviation of the residual sequence for that characteristic parameter, the record is marked as an outlier and removed from the gap crossing current characteristic history sequence. After removing outliers, linear regression fitting is performed again on the remaining records to update the drift direction and drift rate.
[0078] Example 2
[0079] The present invention also proposes a mechanical backlash compensation positioning control system, comprising:
[0080] The gap interval state initialization module is used to acquire encoder pulse signals. When a reversal of the motion direction is detected, the gap interval state variables are initialized and the motor drive current time series is acquired synchronously.
[0081] The gap depth tracking module is used to perform bidirectional classification processing on subsequent encoder pulses within the gap interval to obtain the consumed gap depth.
[0082] The gap crossing determination module is used to monitor the motor drive current. When the current change exceeds the gap crossing current threshold, the gap crossing is determined to be complete, and the mechanical gap estimate is updated.
[0083] The compensation positioning execution module is used to respond to the target stroke command, calculate the compensated encoder target pulse number based on the actual estimated position of the valve stem and the estimated mechanical clearance, and drive the motor to perform positioning.
[0084] The current feature extraction module is used to extract a four-dimensional feature parameter vector from the current time series of the gap crossing process, and add a timestamp to store it in the gap crossing current feature history sequence.
[0085] The trend regression analysis module is used to perform linear regression fitting on each characteristic parameter in the historical sequence of gap crossing current characteristics to obtain the drift direction and drift rate of each characteristic parameter.
[0086] The threshold adaptive update module is used to perform linear extrapolation of the mean no-load current and the peak value of current mutation based on the drift rate, and calculate the updated gap crossing current threshold.
[0087] The abnormal record removal module is used to perform residual detection on newly added gap through current characteristic parameter vectors and remove abnormal records.
[0088] For example, the mechanical backlash compensation positioning control system proposed in this invention may also include the various features and combinations thereof in the method embodiments, which will not be elaborated here.
[0089] In the steam pipeline system of a chemical plant, an electric butterfly valve actuator is responsible for regulating the flow rate of the main steam pipeline. This actuator is equipped with an incremental encoder (512 pulses per revolution), a current sampling circuit (sampling rate of 2kHz), and an embedded controller. The valve has been in operation for approximately 18 months, and the gear pair shows some wear. What is the estimated mechanical clearance under initial conditions? For 38 pulse equivalents, the gap through current threshold The current encoder count for the valve stem is 2150 pulses, corresponding to a current valve stem position of 0.43A. The control system receives a new target stroke command, requiring the valve stem to move in the opening direction to the target position (corresponding to an encoder count of 2370 pulses). Before execution, the valve has just completed a positioning action in the closing direction and is currently within the gap range, having consumed 12 pulse equivalents of gap depth. The movement direction is marked as the closing direction.
[0090] The gap interval state initialization module continuously acquires encoder pulse signals. After the previous closing direction positioning action ends, the controller detects that the movement direction has flipped from the closing direction to the opening direction. At this time, the Boolean flag of the gap interval state variable is set to true, the consumed gap depth is reset to zero, and the movement direction marker is updated to the opening direction. Simultaneously, the time series of the motor drive current after the flip is acquired at a sampling rate of 2kHz, providing data for the gap crossing determination in the subsequent step 3.
[0091] Table 1 Initialization results of state variables during the gap interval at the moment of direction reversal
[0092] State variable components Before flipping Initialization state after flipping Boolean symbol Fake real Consumed gap depth (pulse) 12 0 Movement direction markers Close direction Open direction
[0093] When the Boolean flag is true, the gap depth tracking module performs bidirectional classification processing on each encoder pulse arriving after commutation. After flipping, the motor moves in the opening direction, and the encoder continuously generates opening direction pulses, each pulse increasing the consumed gap depth by one pulse equivalent. During this gap crossing process, the motor maintains unidirectional movement, with no reverse pulses inserted, and the consumed gap depth monotonically increases from 0 until the gap crossing completion judgment is triggered.
[0094] Table 2. Encoder pulse bidirectional classification processing procedure within the gap interval (partial time frames)
[0095] Pulse sequence number Pulse direction Is it consistent with the direction of motion markings? Consumed gap depth (pulse) The first one Open direction yes 1 The 8th Open direction yes 8 The 21st Open direction yes 21 The 35th Open direction yes 35
[0096] The gap crossing detection module continuously monitors the motor drive current. Within the gap range, the gear pair is not yet engaged, the motor is in an unloaded state, and the drive current stabilizes at the unload level of approximately 0.18A. When the accumulated gap depth reaches 37 pulses, the sampling circuit detects a step increase in current, with the current surge reaching 0.51A, exceeding the current effective threshold. A, the gap crossing is determined to be complete. The Boolean flag is set to false, and the gap crossing distance is... One pulse. The mechanical clearance estimate is updated using the exponentially weighted moving average formula:
[0097]
[0098] in For smoothing coefficients, Each pulse represents the gap crossing distance for this operation. Each pulse represents the previous estimated mechanical clearance value, calculated as follows: Each pulse is updated, rounded down, and then used in subsequent calculations.
[0099] Table 3. Results of Clearance Crossing Judgment and Mechanical Clearance Estimation Update
[0100]
[0101] The compensation positioning execution module responds to the target stroke command. At this point, gap crossing is complete, the Boolean flag is false, and the consumed gap depth is zero. The encoder count corresponding to the target position is 2370 pulses, and the current cumulative encoder count is 2150 pulses. The actual estimated position of the valve stem is equal to the cumulative encoder count minus the invalid stroke offset corresponding to the consumed gap depth, i.e. One pulse, with a stroke difference of 220 pulses. The required movement direction is the opening direction. Since the current Boolean flag is false, the consumed clearance depth is zero. In the same direction case, the compensation amount is equal to the estimated mechanical clearance. The encoder target pulse count is the travel difference plus compensation, i.e., [number of pulses]. One pulse, the controller drives the motor to perform 258 pulses of directional movement.
[0102] Table 4 Summary of data from the compensation positioning execution process
[0103] Parameters Numerical value (pulse) Current encoder cumulative count value 2150 Consumed gap depth 0 Valve stem actual estimated position 2150 Target position corresponding encoder count value 2370 Travel difference 220 Required compensation amount 38 Encoder target pulse count 258
[0104] The current feature extraction module extracts four-dimensional feature parameters from the current time series of this gap crossing process. The mean no-load current is taken from the sampling segment before the current abrupt change, the peak value of the current abrupt change is the first local maximum after the step rise, the rise slope is obtained by dividing the difference between the peak value of the abrupt change and the mean no-load current by the rise time, and the current transition time is the duration from the current leaving the no-load level to stabilizing at the load level. The four-dimensional feature parameter vector is appended with the timestamp of this crossing and stored in the historical sequence.
[0105] Table 5 Characteristic parameter vector of this gap crossing current
[0106]
[0107] After the number of historical records of the gap crossing current characteristic reaches the minimum number of regression samples, the trend regression analysis module performs linear regression fitting on each of the four characteristic parameters. Taking the mean no-load current as an example, the regression input is the crossing timestamp of each record. and corresponding The value is obtained by solving the linear model using the least squares method. ,in This represents the drift rate; its sign reflects the drift direction, and its value reflects the drift rate. The intercept is used. The four feature parameters are fitted independently and do not interfere with each other.
[0108] Table 6 shows the linear regression fitting results for the four feature parameters.
[0109]
[0110] The threshold adaptive update module uses the linear regression equations corresponding to the mean no-load current and the peak current surge to perform linear extrapolation at the next expected operating time point (4500h of year A) to obtain the predicted no-load current value. and mutation peak prediction value The updated gap crossing current threshold is calculated using the following formula:
[0111] ;
[0112] in This is the predicted peak value for mutations. The no-load current is the predicted value, which is obtained by substituting the extrapolated value of the corresponding linear model at the target time point in step 6. Preset safety factor; The unit is ampere, which is consistent with the dimension of the current change amplitude in step 3.
[0113] Table 7. Adaptive Update Results of Gap Crosscurrent Threshold
[0114]
[0115] The anomaly record removal module calculates the residuals for the four-dimensional feature parameter vectors of each newly added historical sequence. Taking the newly added record as an example, it calculates the residuals of its four feature parameter values relative to the respective trend lines in step 6 and compares them with three times the standard deviation of the corresponding residual sequence. In this case, none of the parameter residuals exceed three times the standard deviation, so the record is retained. If an abnormal operating condition occurs (such as transient fluctuations in steam pressure causing abnormal current), the residual of a certain feature parameter in the corresponding record will exceed a threshold. This record is then marked and removed, and linear regression fitting is re-executed to update the record. and .
[0116] Table 8 Residual detection results of newly added records
[0117] Feature parameters Measured value Trendline Prediction Value residual Three standard deviations of the residual series Is it abnormal? Average no-load current (A) 0.18 0.18 0.00 0.03 no Peak value of sudden current change (A) 0.69 0.71 -0.02 0.05 no Rising slope (A / s) 8.20 8.16 0.04 0.62 no Current transition time (s) 0.062 0.059 0.003 0.011 no
[0118] The entire implementation process presents a clear data flow logic. Step 1 initializes the gap interval state variables after detecting a direction reversal, providing initial conditions for the pulse bidirectional classification and counting in Step 2. Step 2 updates the consumed gap depth in real time, and this value is recorded as the crossing amount when the crossing determination is completed in Step 3. The mechanical clearance estimate is updated using an exponentially weighted moving average formula. Step 4 directly uses the consumed gap depth from Step 2 and the depth from Step 3. Calculate the compensated encoder target pulse count to drive the motor for precise positioning. Step 5 extracts a four-dimensional feature vector from the current time series in Step 3 and inputs it into the linear regression fitting in Step 6. The regression result in Step 6 simultaneously serves the threshold extrapolation calculation in Step 7 and the residual detection in Step 8. Step 7 updates the gap crossing current threshold. Feedback to step 3 allows subsequent crossing judgment criteria to adaptively adjust as the motor's electrical characteristics drift. Abnormal record removal in step 8 ensures the regression fitting accuracy of step 6, thereby maintaining the reliability of threshold updates in step 7. Data from each step flows between modules in units of encoder pulse equivalent or amperes, maintaining consistent dimensions and forming a complete adaptive compensation closed loop.
[0119] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.
[0120] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.
[0121] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
Claims
1. A stroke control method for coal mine safety electric actuators based on encoder calculation, characterized in that, include: The encoder pulse signal is collected. When the motion direction is detected to be reversed, the Boolean flag in the gap interval state variable is set to true, the consumed gap depth is set to zero, the motion direction mark is recorded, and the motor drive current time series is collected simultaneously. When the Boolean flag is true, subsequent encoder pulses are classified bidirectionally. If the pulse direction is consistent with the motion direction mark, the consumed gap depth is increased by one pulse equivalent; otherwise, it is decreased by one pulse equivalent. Monitor the motor drive current. When the current change exceeds the gap crossing current threshold, it is determined that the gap crossing is complete. Set the Boolean flag to false and update the mechanical gap estimate with the gap crossing amount using the exponential weighted moving average method. In response to the target stroke command, the encoder target pulse number is calculated based on the stroke difference between the actual estimated position of the valve stem and the target position, as well as the compensation amount determined based on the direction of motion, and the motor is driven to perform positioning.
2. The stroke control method for coal safety electric actuators based on encoder calculation according to claim 1, characterized in that, The mechanical clearance estimate is updated using the current clearance crossing amount using an exponentially weighted moving average method, specifically as follows: Let the first The amount of passage through the gap is Updated mechanical clearance estimates Calculated by the following formula: ; in, This is the estimated mechanical clearance value from the last update. For smoothing coefficients; , , All units are encoder pulse equivalents.
3. The stroke control method for coal safety electric actuators based on encoder calculation according to claim 1, characterized in that, The actual estimated position of the valve stem is equal to the encoder's cumulative count value minus the invalid stroke offset corresponding to the consumed clearance depth; The compensation amount determined based on the direction of motion includes: if the required direction of motion is consistent with the direction of motion marker within the current gap interval, then the compensation amount is the difference between the estimated mechanical gap value and the consumed gap depth. If the required direction of movement is reversed, the compensation amount is the value of the consumed gap depth.
4. The stroke control method for coal safety electric actuators based on encoder calculation according to claim 1, characterized in that, The bidirectional classification process also includes: when the consumed gap depth decreases to zero, the current flip direction is re-recorded as a new motion direction marker, and the consumed gap depth is reset to zero to start a new round of gap crossing tracking.
5. The stroke control method for a coal safety electric actuator based on encoder calculation according to claim 3, characterized in that, When the Boolean flag is false, the consumed gap depth is zero. In this case, the compensation amount in the same direction is equal to the estimated mechanical gap value, and the compensation amount in the opposite direction is zero.
6. The stroke control method for a coal safety electric actuator based on encoder calculation according to claim 1, characterized in that, Also includes: A four-dimensional feature parameter vector is extracted from the current time series of this gap crossing process. The four-dimensional feature parameter vector includes the mean no-load current in the gap interval, the peak value of the current change at the time of crossing completion, the current change rise slope, and the current transition time from no-load to load. The four-dimensional feature parameter vector is appended with a crossing timestamp and stored in the gap crossing current feature history sequence.
7. The stroke control method for a coal safety electric actuator based on encoder calculation according to claim 6, characterized in that, Also includes: Once the number of records in the historical sequence of the gap current characteristics reaches the minimum number of regression samples, linear regression fitting is performed on the four feature parameters in the four-dimensional feature parameter vector to obtain the drift direction and drift rate of each feature parameter with the cumulative running time; wherein, the first The linear model for the characteristic parameters is: ; in, For cumulative running time, For the first The drift rate of each characteristic parameter, whose sign indicates the drift direction. The intercept is... and It is obtained by solving the least squares method.
8. The stroke control method for a coal safety electric actuator based on encoder calculation according to claim 7, characterized in that, Also includes: Based on the linear models corresponding to the mean no-load current and the peak current surge, linear extrapolation is performed at the time point of the next expected operating period to obtain the predicted no-load current values. and mutation peak prediction value Updated gap crossing current threshold Calculated by the following formula: ; in, The preset safety factor is a dimensionless positive real number.
9. The stroke control method for a coal safety electric actuator based on encoder calculation according to claim 8, characterized in that, Also includes: For each newly added four-dimensional feature parameter vector of the gap crossing current feature history sequence, calculate the residuals of its four feature parameter values relative to the corresponding linear regression trend line; when the residual of any feature parameter exceeds three times the standard deviation of the feature parameter residual sequence, mark the record as an abnormal record and remove it from the gap crossing current feature history sequence; after removing the abnormal record, re-perform linear regression fitting on the remaining records to update the drift direction and drift rate.
