A multivariable adaptive control method for valve actuators

Abstract

The application relates to the technical field of automatic control, and discloses a multivariable adaptive control method for a valve actuator, which comprises the following steps: synchronously normalizing multi-source data to obtain a standard sequence; decomposing to obtain an opening deviation value, a deviation change rate and a pressure difference fluctuation characteristic value; according to a valve opening set value and the pressure difference fluctuation characteristic value, addressing a feedforward adjustment amount from a feedforward table, and cooperating with the opening deviation value to obtain a feedforward component in a nonlinear weight mode; a two-dimensional state point is formed by the opening deviation value and the change rate, and a projection distance of the two-dimensional state point on a preset compensation plane is calculated; according to the projection distance, a compensation coefficient is positioned by one-dimensional projection of a compensation curve, and the compensation coefficient is modulated and coupled with the opening deviation value to obtain a compensation component; a current period weight is obtained by correcting an initial weight according to an effect evaluation value of a previous period; according to the current period weight and a communication protocol, adaptive control instructions are obtained by protocol mapping and packaging of the feedforward and compensation components; and the application can improve the efficiency of multivariable adaptive control of the valve actuator.

Description

Technical Field

[0001] This invention relates to the field of automatic control technology, and in particular to a multivariable adaptive control method for valve actuators. Background Technology

[0002] Traditional valve actuator control schemes often employ a single-variable, fixed-parameter adjustment mode, which cannot synchronously process multi-source coupled signals such as valve opening and differential pressure. Data acquisition and processing suffer from asynchronous time scales and inconsistent amplitude scales, making it difficult to form a standardized control input sequence. This results in persistent deviations between the control response and actual operating conditions. Existing control strategies do not incorporate feedforward pre-adjustment based on differential pressure fluctuation characteristics, relying solely on opening deviations for lag correction. Furthermore, the generation of control components lacks a nonlinear coordination mechanism, failing to quickly offset control errors caused by external disturbances.

[0003] Current valve actuator compensation control relies on fixed parameter curves and does not perform adaptive positioning calculations based on the two-dimensional state of opening deviation and deviation change rate. This results in insufficient accuracy in compensation coefficient matching and signal distortion during compensation component modulation. Furthermore, the fixed configuration of control weights prevents dynamic correction based on the previous cycle's control effect, leading to poor compatibility between control command output and communication protocols. Overall, the control efficiency and regulation stability fail to meet the demands of high-precision operating conditions. Therefore, improving the efficiency of adaptive control command generation for valve actuators has become an urgent problem to be solved. Summary of the Invention

[0004] This invention provides a multivariable adaptive control method for valve actuators to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides a multivariable adaptive control method for a valve actuator, comprising:

[0006] I. Synchronously normalize the multi-source data of the valve actuator to obtain a standard sequence of the multi-source data;

[0007] II. Perform deviation fluctuation characteristic decomposition on the standard sequence to obtain the opening deviation value, opening deviation change rate and pressure difference fluctuation characteristic value of the standard sequence;

[0008] III. Based on the valve opening setpoint and the differential pressure fluctuation characteristic value of the multi-source data, perform multi-dimensional addressing matching on the feedforward table of the valve actuator to obtain the feedforward adjustment amount of the feedforward table, and perform nonlinear weight coordination between the feedforward adjustment amount and the opening deviation value to obtain the feedforward component of the valve actuator;

[0009] IV. Construct a two-dimensional state point from the opening deviation value and the opening deviation change rate, and calculate the vertical projection distance of the two-dimensional state point on the preset compensation plane to obtain the projection distance of the two-dimensional state point.

[0010] V. Based on the projection distance, perform one-dimensional coordinate projection positioning on the compensation curve of the valve actuator to obtain the compensation coefficient of the projection distance, and modulate and couple the compensation coefficient with the opening deviation value to obtain the compensation component of the valve actuator;

[0011] VI. Based on the historical performance evaluation value of the valve actuator in the previous cycle, the initial weight of the valve actuator is corrected according to performance orientation to obtain the current cycle weight of the valve actuator;

[0012] VII. Based on the current cycle weight and the preset communication protocol, the feedforward component and the compensation component are encapsulated by protocol mapping to obtain the adaptive control command of the valve actuator.

[0013] In a preferred embodiment, the step of synchronously normalizing the multi-source data of the valve actuator to obtain a standard sequence of the multi-source data includes:

[0014] Time-scale consistency correction is performed on the signal timestamps of the valve actuator in the multi-source data to obtain the time-aligned multi-source data of the valve actuator.

[0015] The values ​​of the data sources in the time-aligned multi-source data are mapped to a pre-stored normalized mapping table to obtain the standard values ​​of the data sources;

[0016] Based on the chronological order of the timestamps, the standard values ​​are sorted in ascending order of time to obtain the standard sequence of the multi-source data.

[0017] In a preferred embodiment, the step of performing deviation fluctuation characteristic decomposition on the standard sequence to obtain the aperture deviation value, aperture deviation change rate, and differential pressure fluctuation characteristic value of the standard sequence includes:

[0018] Based on the identification field of the data source, the standard sequence is decoupled from the signal channel to obtain the standard subsequence of the opening degree and the standard subsequence of the pressure difference.

[0019] The current sampling time standard value in the opening standard subsequence is compared with the valve opening setting value of the multi-source data to determine the amplitude difference, thereby locking the opening deviation value of the current sampling time of the standard sequence.

[0020] The standard value of the previous sampling time in the standard subsequence of valve opening is compared with the valve opening set value by time-series backtracking and difference identification to determine the opening deviation value of the standard sequence at the previous sampling time. The deviation slope of the current sampling time opening deviation value and the previous sampling time opening deviation value is captured to obtain the opening deviation change rate of the standard sequence.

[0021] The pressure difference fluctuation characteristic value of the standard sequence is obtained by offset detection between the current sampling time standard value and the initial sampling time standard value in the pressure difference standard subsequence.

[0022] In a preferred embodiment, the process of performing multi-dimensional addressing matching on the feedforward table of the valve actuator based on the valve opening setpoint and the differential pressure fluctuation characteristic value from the multi-source data to obtain the feedforward adjustment amount of the feedforward table includes:

[0023] The valve opening setting value of the multi-source data is taken as the first component, and the differential pressure fluctuation characteristic value is taken as the second component. The first component and the second component are bound in component order to obtain a two-dimensional coordinate pair of the first component and the second component.

[0024] The two-dimensional coordinate pairs are vectorized and assembled to obtain the two-dimensional addressing vector of the feedforward table in the valve actuator;

[0025] Based on the two-dimensional addressing vector, a joint interval assignment determination is performed on the row interval set and column interval set of the feedforward table, and the row interval into which the first component falls is taken as the target row interval, and the column interval into which the second component falls is taken as the target column interval.

[0026] Based on the spatial intersection of the target row interval and the target column interval, address decoding is triggered on the storage unit of the feedforward table to obtain the feedforward adjustment amount of the feedforward table.

[0027] In a preferred embodiment, the step of nonlinearly weighting the feedforward adjustment amount and the opening deviation value to obtain the feedforward component of the valve actuator includes:

[0028] Filter out the weighted segments corresponding to the magnitude of the opening deviation value;

[0029] The weight association table bound to the weight segment is decoded and output to obtain the collaborative weight value of the weight segment;

[0030] Based on the cooperative weight value, the feedforward adjustment amount is gain-scaled to obtain the scaled feedforward adjustment amount.

[0031] The scaled feedforward adjustment and the opening deviation value are superimposed by voltage to obtain the feedforward component of the valve actuator.

[0032] In a preferred embodiment, the step of constructing a two-dimensional state point from the opening deviation value and the opening deviation change rate, and calculating the vertical projection distance of the two-dimensional state point on a preset compensation plane to obtain the projection distance of the two-dimensional state point includes:

[0033] The opening deviation value is assigned to the first coordinate slot, and the opening deviation change rate is assigned to the second coordinate slot. The first coordinate slot and the second coordinate slot are then paired and anchored to obtain the two-dimensional state point, which includes the first coordinate value. Second coordinate value ;

[0034] By extracting coefficients from the planar expression of the preset compensation plane, the first normal coefficient of the preset compensation plane is obtained. Second normal coefficient and distance constant ;

[0035] Calculate the vertical projection distance from the two-dimensional state point to the preset compensation plane. The calculation formula is as follows:

[0036] ;

[0037] The vertical projection distance is... The first coordinate value of the two-dimensional state point is the opening deviation value. The second coordinate value of the two-dimensional state point is the opening deviation change rate. Let be the coefficient of the normal vector of the preset compensation plane in the first dimension. The coefficient of the normal vector of the preset compensation plane in the second dimension. The position constant of the preset compensation plane. The magnitude of the normal vector is used for distance normalization.

[0038] In a preferred embodiment, the step of performing one-dimensional coordinate projection positioning on the compensation curve of the valve actuator based on the projection distance to obtain the compensation coefficient of the projection distance includes:

[0039] The projection distance is projected onto the origin of a one-dimensional coordinate axis to obtain the normalized projection coordinate value of the projection distance;

[0040] Based on the normalized projection coordinate values, serial interval landing point matching is performed on the segmented boundary threshold sequence of the compensation curve, and the normalized projection coordinate values ​​and the threshold windows in the boundary threshold sequence are subjected to segment-by-segment membership adjudication to obtain the target segmented interval of the compensation curve.

[0041] Based on the ordinal number of the target segment interval, the coefficient storage slots bound to the ordinal number are selected to obtain the selected coefficient storage slots of the compensation curve.

[0042] The storage depth of the coefficient amplitude in the coefficient storage slot is captured, and the coefficient amplitude is extracted through the read channel activated by the gating switch to obtain the compensation coefficient of the projection distance.

[0043] In a preferred embodiment, modulating and coupling the compensation coefficient with the opening deviation value to obtain the compensation component of the valve actuator includes:

[0044] Using the compensation coefficient as the amplitude control quantity and the opening deviation value as the carrier signal, the carrier signal is weighted and mapped based on the amplitude control quantity to obtain the weighted mapped amplitude-modulated signal of the carrier signal.

[0045] The amplitude-modulated signal is subjected to absolute value rectification transformation to obtain the rectified envelope signal of the weighted mapped amplitude-modulated signal;

[0046] The rectified envelope signal is subjected to low-pass smoothing filtering to remove the carrier frequency component in the amplitude modulation signal, thereby obtaining the compensation component of the valve actuator.

[0047] In a preferred embodiment, the step of adjusting the initial weight of the valve actuator based on the historical performance evaluation value of the valve actuator in the previous cycle to obtain the current cycle weight of the valve actuator includes:

[0048] The previous cycle historical performance evaluation value of the valve actuator is used as the performance quantification index to classify the interval and obtain the correction level interval of the previous cycle.

[0049] Based on the performance correction parameter library of the valve actuator, the offset storage unit bound to the correction level interval is selected and collected by unit selection to obtain the weight offset of the correction level interval.

[0050] Symbolic weight injection is performed on the initial weight of the valve actuator and the weight offset to obtain the corrected intermediate weight of the initial weight;

[0051] The modified intermediate weights are subjected to boundary clamping constraints to obtain the current cycle weights of the valve actuator.

[0052] In a preferred embodiment, the step of performing protocol mapping and encapsulation on the feedforward component and the compensation component based on the current cycle weight and a preset communication protocol to obtain the adaptive control command for the valve actuator includes:

[0053] Based on the current cycle weight, the feedforward component and the compensation component are weighted and fused to obtain the composite control quantity of the feedforward component and the compensation component.

[0054] Based on the data field format of the preset communication protocol, the value range of the synthetic control quantity is mapped to the allowed value range of the protocol field to obtain the protocol payload of the synthetic control quantity.

[0055] Based on the frame structure definition of the preset communication protocol, the protocol payload is assembled into protocol frames to obtain a complete protocol frame of the protocol payload;

[0056] According to the bit order rules of the preset communication protocol, the complete protocol frame is serialized into a bit stream to obtain the adaptive control command of the valve actuator.

[0057] Compared with the prior art, the present invention has the following beneficial effects:

[0058] 1. This invention achieves standardized processing of valve actuator control signals through multi-source data synchronous normalization and deviation fluctuation feature decomposition, accurately extracting opening deviation, deviation change rate, and differential pressure fluctuation features, ensuring the consistency and stability of the control input sequence, and improving the efficiency of control signal analysis and feature extraction. Based on the opening setpoint and differential pressure fluctuation features, multi-dimensional addressing matching of the feedforward table is completed, and feedforward components are generated collaboratively with nonlinear weights, which can quickly form pre-adjustment control quantities, enhancing control response speed and disturbance suppression capability.

[0059] 2. This invention relies on two-dimensional state point projection calculation and adaptive positioning of compensation curves to accurately match compensation coefficients and complete compensation component modulation coupling, thereby improving control compensation accuracy and signal smoothness and avoiding control signal distortion. It dynamically adjusts control weights based on the control effect of the previous cycle, and completes control component encapsulation and output according to the communication protocol, achieving self-optimization of control parameters and efficient command adaptation, comprehensively improving the overall efficiency and regulation stability of the valve actuator's multi-variable adaptive control. Attached Figure Description

[0060] Figure 1 A flowchart illustrating a multivariable adaptive control method for a valve actuator according to an embodiment of the present invention;

[0061] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0062] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0063] This application provides a multivariable adaptive control method for a valve actuator. The executing entity of this multivariable adaptive control method for a valve actuator includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the multivariable adaptive control method for a valve actuator can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0064] Reference Figure 1 The diagram shown is a flowchart illustrating a multivariable adaptive control method for a valve actuator according to an embodiment of the present invention. In this embodiment, the multivariable adaptive control method for a valve actuator includes:

[0065] I. Synchronously normalize the multi-source data of the valve actuator to obtain a standard sequence of the multi-source data;

[0066] In this embodiment of the invention, the step of synchronously normalizing the multi-source data of the valve actuator to obtain a standard sequence of the multi-source data includes:

[0067] Time-scale consistency correction is performed on the signal timestamps of the valve actuator in the multi-source data to obtain the time-aligned multi-source data of the valve actuator.

[0068] The values ​​of the data sources in the time-aligned multi-source data are mapped to a pre-stored normalized mapping table to obtain the standard values ​​of the data sources;

[0069] Based on the chronological order of the timestamps, the standard values ​​are sorted in ascending order of time to obtain the standard sequence of the multi-source data.

[0070] To perform timescale consistency correction on the signal timestamps of the valve actuator in multi-source data, firstly, all signal timestamps corresponding to each data source in the multi-source data of the valve actuator are acquired. This multi-source data includes various relevant data such as the valve actuator's operating current data, valve opening data, ambient temperature data, and actuator vibration data. Each data point from each data source carries a unique signal timestamp, which is a real-time record of the data acquisition. Then, the data source with the highest timestamp accuracy among all data sources is selected as the time base. The timestamp accuracy of this time base is set to milliseconds, meaning that the timestamp of each data point is accurate to a specific number of milliseconds. Then, the timestamps of the other data sources are... The signal timestamps from the sources are calibrated according to this time reference. During the calibration process, if the accuracy of the signal timestamp of a certain data source is lower than the millisecond level, the corresponding milliseconds are added. The addition rule is to use the milliseconds of the middle moment within the data acquisition period as the standard. If there is a deviation in the signal timestamp of a certain data source, the deviation value is obtained by comparing it with the timestamp of the corresponding acquisition time in the time reference data source. The deviation value is then uniformly corrected to 0 to ensure that the timestamp accuracy of each data from all data sources reaches the millisecond level, and that all data from the same data source at the same acquisition time corresponds to the same timestamp. After the calibration is completed, all the data from the data sources obtained are the time-aligned multi-source data of the valve actuator.

[0071] The values ​​from the data sources in the time-aligned multi-source data are mapped to a pre-stored normalized mapping table. First, the construction method of the pre-stored normalized mapping table is clarified. This mapping table is pre-constructed based on the normal operating range of each data source of the valve actuator. Each data source corresponds to a unique mapping interval. For example, the normal operating range of valve opening data is 0% to 100%, and the corresponding normalized mapping interval is 0 to 1. The normal operating range of operating current data is 0.5A to 5A, and the corresponding normalized mapping interval is also 0 to 1. The normal operating range of each data source is determined by the factory parameters of the valve actuator, ensuring the rationality and accuracy of the mapping interval. Then, each value from each data source in the time-aligned multi-source data is extracted. Based on the data source type of the value, the corresponding mapping interval in the normalized mapping table is found. The value is then converted according to the upper and lower limits of the mapping interval. During the conversion process, it is ensured that the converted value strictly falls within the unified interval of 0 to 1, and each original value corresponds to a unique converted value, without omissions or repetitions. After completing the mapping conversion of all values, each converted value is the standard value of the data source.

[0072] Based on the chronological order of timestamps, the standard values ​​are sorted in ascending order of time. First, the timestamps corresponding to all standard values ​​are extracted. These timestamps are uniform millisecond-level timestamps after time stamp consistency correction. Each standard value is bound to a unique timestamp, and the timestamps are consecutive millisecond-level records with no missing or duplicate records. Then, all standard values ​​bound to timestamps are arranged in ascending order of timestamps. During the arrangement process, the numerical value of the timestamp is used as the sorting criterion. The smaller the value, the earlier the time, and the larger the value, the later the time. The arrangement ensures that each standard value corresponds to the correct timestamp and there is no mismatch between timestamp and standard value. At the same time, it ensures that the timestamps in the arranged standard value sequence are sequentially increasing, without reversal or omission. After the arrangement is completed, the ordered standard value sequence is the standard sequence of multi-source data.

[0073] The beneficial effects are as follows: By uniformly calibrating the timestamps of multi-source data to the same reference clock, timing deviations caused by inconsistent acquisition times of different signal sources can be eliminated, ensuring synchronous matching of all signals in the time dimension. This provides a consistent foundation of data for subsequent feature decomposition and control calculations, avoiding control errors caused by timing misalignments. Using a pre-defined normalization mapping table to uniformly transform the original values ​​can standardize data sources with different dimensions and numerical ranges into standard values ​​of the same dimension and numerical range, eliminating interference from numerical scale differences in control calculations and improving the consistency and comparability of data processing. Arranging the standard values ​​in ascending order according to their timestamps forms a continuous and time-ordered standard sequence, providing a stable and orderly data carrier for feature extraction such as opening degree and differential pressure. This ensures the accuracy and continuity of the feature decomposition process, providing reliable data support for subsequent control stages such as feedforward component generation and compensation component calculation. Overall, this improves the accuracy and efficiency of multi-source data processing, laying a data foundation for the stable implementation of adaptive control of valve actuators.

[0074] II. Perform deviation fluctuation characteristic decomposition on the standard sequence to obtain the opening deviation value, opening deviation change rate and pressure difference fluctuation characteristic value of the standard sequence;

[0075] In this embodiment of the invention, the step of performing deviation fluctuation characteristic decomposition on the standard sequence to obtain the aperture deviation value, aperture deviation change rate, and differential pressure fluctuation characteristic value of the standard sequence includes:

[0076] Based on the identification field of the data source, the standard sequence is decoupled from the signal channel to obtain the standard subsequence of the opening degree and the standard subsequence of the pressure difference.

[0077] The current sampling time standard value in the opening standard subsequence is compared with the valve opening setting value of the multi-source data to determine the amplitude difference, thereby locking the opening deviation value of the current sampling time of the standard sequence.

[0078] The standard value of the previous sampling time in the standard subsequence of valve opening is compared with the valve opening set value by time-series backtracking and difference identification to determine the opening deviation value of the standard sequence at the previous sampling time. The deviation slope of the current sampling time opening deviation value and the previous sampling time opening deviation value is captured to obtain the opening deviation change rate of the standard sequence.

[0079] The pressure difference fluctuation characteristic value of the standard sequence is obtained by offset detection between the current sampling time standard value and the initial sampling time standard value in the pressure difference standard subsequence.

[0080] Based on the identification field of the data source, the standard sequence is decoupled from the signal channel. First, it is determined that each data point in the standard sequence carries a corresponding data source identification field, which is a pre-defined unique character identifier. The identification field for the opening-related data source is set to "KD", and the identification field for the differential pressure-related data source is set to "YC". The identification field corresponds one-to-one with the data source type, with no duplication or confusion. Then, all data in the standard sequence are traversed, and the identification field carried by each data point is extracted one by one. The data is classified according to the type of the identification field. All data with the identification field "KD" are filtered out and integrated to form an independent subsequence, which is the opening standard subsequence of the standard sequence. At the same time, all data with the identification field "YC" are filtered out and integrated to form another independent subsequence, which is the differential pressure standard subsequence of the standard sequence. During the classification process, it is ensured that all data are accurately filtered into the corresponding subsequence, with no omissions or misclassifications. After the signal channel decoupling is completed, two products, the opening standard subsequence and the differential pressure standard subsequence, are obtained simultaneously.

[0081] The amplitude difference is compared between the standard value of the current sampling time in the standard subsequence of valve opening and the valve opening setpoint of the multi-source data to lock the opening deviation value of the standard sequence at the current sampling time. First, the current sampling time is determined, which is the timestamp corresponding to the latest data in the standard subsequence of valve opening. This timestamp has the same accuracy as the timestamp after the previous timescale consistency correction, both being in the millisecond range. At the same time, the standard value corresponding to the current sampling time is extracted, i.e., the standard value of the current sampling time. Then, the preset valve opening setpoint is extracted from the multi-source data. This setpoint is the valve actuator. During normal operation, the preset fixed opening standard value is set and stored in advance by the staff according to the actual work needs, and the dimension of this setting value is consistent with the standard value in the opening standard subsequence. The amplitude discrimination process is to calculate the difference between the standard value at the current sampling time and the valve opening setting value. In the calculation, the valve opening setting value is subtracted from the standard value at the current sampling time, and the difference obtained is the opening deviation value at the current sampling time of the standard sequence. After completing this calculation, the opening deviation value at the current sampling time can be locked to ensure that the deviation value corresponds one-to-one with the standard value and the valve opening setting value at the current sampling time, and there is no calculation error.

[0082] The valve opening standard subsequence is compared with the previous sampling time standard value and the valve opening set value using time-series backtracking to determine the opening deviation value of the standard sequence at the previous sampling time. The current sampling time opening deviation value and the previous sampling time opening deviation value are then compared using deviation slope capture to obtain the opening deviation change rate of the standard sequence. First, the previous sampling time is determined; the previous sampling time is the sampling time preceding the current sampling time, and its corresponding timestamp is one fixed sampling period smaller than the current sampling time's timestamp. This sampling period is a pre-set 100 milliseconds, meaning the time interval between the previous and current sampling times is fixed at 100 milliseconds. Then, the previous sampling time is extracted from the standard opening standard subsequence. The standard value corresponding to the sampling time is the standard value of the previous sampling time. The time-series backtracking difference detection process calculates the difference between the standard value of the previous sampling time and the valve opening setting value. The calculation is done by subtracting the valve opening setting value from the standard value of the previous sampling time. The difference obtained is the opening deviation value of the standard sequence at the previous sampling time. The deviation slope capture process is to subtract the opening deviation value of the previous sampling time from the opening deviation value of the current sampling time. The difference obtained is then divided by the sampling period of 100 milliseconds. The result is the opening deviation change rate of the standard sequence. This change rate can reflect the change of the opening deviation value over time. During the calculation process, it is ensured that the difference calculation is accurate at each step and the sampling period remains unchanged. Finally, a unique opening deviation change rate is obtained.

[0083] The pressure difference fluctuation characteristic value of the standard sequence is obtained by offset detection of the current sampling time standard value and the initial sampling time standard value in the pressure difference standard subsequence. First, the current sampling time and the initial sampling time in the pressure difference standard subsequence are determined. The current sampling time is the timestamp corresponding to the latest data in the pressure difference standard subsequence, which is consistent with the current sampling time of the opening standard subsequence. The initial sampling time is the timestamp corresponding to the first data in the pressure difference standard subsequence, that is, the time when the pressure difference data started to be collected. Then, the standard values ​​corresponding to the current sampling time and the initial sampling time in the pressure difference standard subsequence are extracted respectively to ensure that both standard values ​​are pressure difference-related standard values ​​and consistent with the data source type of the pressure difference standard subsequence. The offset detection process is to calculate the difference between the current sampling time standard value and the initial sampling time standard value. The calculation is done by subtracting the initial sampling time standard value from the current sampling time standard value. The difference obtained is the pressure difference fluctuation characteristic value of the standard sequence. This characteristic value can reflect the offset between the current pressure difference and the initial pressure difference. During the calculation process, it is ensured that the extraction of the two standard values ​​is accurate and the difference calculation is error-free, and finally a unique pressure difference fluctuation characteristic value is obtained.

[0084] The beneficial effects include decoupling the signal channels of the standard sequence through the data source identification field, which can accurately separate the two core signals of opening degree and differential pressure, forming independent and regular standard subsequences of opening degree and differential pressure, avoiding feature extraction distortion caused by signal aliasing, and providing a clean signal foundation for subsequent deviation and fluctuation analysis. Directly comparing the current standard value of opening degree with the opening degree setpoint and locking a fixed difference can accurately obtain the opening degree deviation value at the current sampling time, providing accurate deviation basis for real-time control and adjustment. By obtaining the opening degree deviation value at the previous sampling time through time-series backtracking and then calculating the deviation change, the deviation slope can be stably captured, obtaining an accurate rate of change of opening degree deviation, fully reflecting the dynamic change trend of opening degree deviation. Comparing the current standard value of differential pressure with the standard value at the initial sampling time to complete offset detection can accurately extract the characteristic value of differential pressure fluctuation, intuitively reflecting the actual fluctuation state of differential pressure. The entire feature decomposition process is logically rigorous and numerically deterministic, and can efficiently output three core control features: opening deviation value, opening deviation change rate, and differential pressure fluctuation characteristic value. This provides reliable input for feedforward control and compensation control, and improves the accuracy and stability of control calculations.

[0085] III. Based on the valve opening setpoint and the differential pressure fluctuation characteristic value of the multi-source data, perform multi-dimensional addressing matching on the feedforward table of the valve actuator to obtain the feedforward adjustment amount of the feedforward table, and perform nonlinear weight coordination between the feedforward adjustment amount and the opening deviation value to obtain the feedforward component of the valve actuator;

[0086] In this embodiment of the invention, the process of performing multi-dimensional addressing matching on the feedforward table of the valve actuator based on the valve opening setpoint and the differential pressure fluctuation characteristic value from the multi-source data to obtain the feedforward adjustment amount of the feedforward table includes:

[0087] The valve opening setting value of the multi-source data is taken as the first component, and the differential pressure fluctuation characteristic value is taken as the second component. The first component and the second component are bound in component order to obtain a two-dimensional coordinate pair of the first component and the second component.

[0088] The two-dimensional coordinate pairs are vectorized and assembled to obtain the two-dimensional addressing vector of the feedforward table in the valve actuator;

[0089] Based on the two-dimensional addressing vector, a joint interval assignment determination is performed on the row interval set and column interval set of the feedforward table, and the row interval into which the first component falls is taken as the target row interval, and the column interval into which the second component falls is taken as the target column interval.

[0090] Based on the spatial intersection of the target row interval and the target column interval, address decoding is triggered on the storage unit of the feedforward table to obtain the feedforward adjustment amount of the feedforward table.

[0091] The step of performing nonlinear weighted coordination between the feedforward adjustment amount and the opening deviation value to obtain the feedforward component of the valve actuator includes:

[0092] Filter out the weighted segments corresponding to the magnitude of the opening deviation value;

[0093] The weight association table bound to the weight segment is decoded and output to obtain the collaborative weight value of the weight segment;

[0094] Based on the cooperative weight value, the feedforward adjustment amount is gain-scaled to obtain the scaled feedforward adjustment amount.

[0095] The scaled feedforward adjustment and the opening deviation value are superimposed by voltage to obtain the feedforward component of the valve actuator.

[0096] The valve opening setting value from the multi-source data is taken as the first component, and the differential pressure fluctuation characteristic value is taken as the second component. The first component and the second component are sequentially bound to obtain a two-dimensional coordinate pair between the first component and the second component. First, the preset valve opening setting value is extracted from the multi-source data. This setting value is a fixed value that is set and stored in advance by the staff according to the actual working needs of the valve actuator. It is consistent with the dimension of the standard value in the opening standard subsequence. This valve opening setting value is determined as the first component. At the same time, the differential pressure fluctuation characteristic value of the previously obtained standard sequence is extracted. This characteristic value is the difference between the standard value at the current sampling time and the standard value at the initial sampling time in the differential pressure standard subsequence. This differential pressure fluctuation characteristic value is determined as the second component. The component sequential binding process is to bind the first component and the second component in a fixed order of "first component first, second component last". After binding, a set of combined data containing two values ​​is formed. This combined data is the two-dimensional coordinate pair between the first component and the second component. During the binding process, it is ensured that the correspondence between the first component and the second component is not reversed. Each set of two-dimensional coordinate pairs uniquely corresponds to a set of first component and second component, with no misbinding or omission.

[0097] The two-dimensional coordinate pairs are vectorized and assembled to obtain the two-dimensional addressing vector of the feedforward table in the valve actuator. First, the assembly rules of the two-dimensional addressing vector are defined. The rule is that the first component of the two-dimensional coordinate pair is used as the first element of the vector, and the second component of the two-dimensional coordinate pair is used as the second element of the vector. The element order of the vector is completely consistent with the binding order of the components in the two-dimensional coordinate pair, and the vector contains only these two elements, with no extra or missing elements. The vectorization assembly process is to fill the two components of the two-dimensional coordinate pair into a preset vector structure in sequence according to the above rules. This vector structure is a fixed format set in advance to carry the two component information required for two-dimensional addressing. After assembly, the resulting vector is the two-dimensional addressing vector of the feedforward table in the valve actuator. This vector can accurately reflect the numerical information of the first and second components, providing a unique basis for the subsequent addressing of the feedforward table.

[0098] Based on the two-dimensional addressing vector, a joint interval assignment determination is performed on the row interval set and column interval set of the feedforward table. The row interval to which the first component falls is taken as the target row interval, and the column interval to which the second component falls is taken as the target column interval. First, the preset rules for the row interval set and column interval set of the feedforward table are defined. The row interval set of the feedforward table consists of multiple continuous and non-overlapping intervals. Each row interval corresponds to a fixed opening set value range. The interval division is based on the normal opening working range of the valve actuator, 0% to 100%. This range is divided into 10 row intervals, with each row interval having a span of 10%. For example, the first row interval is 0% to 10%, the second row interval is 10% to 20%, and so on. All row intervals together form the row interval set. Similarly, the column interval set of the feedforward table consists of multiple continuous and non-overlapping intervals. Each column interval corresponds to a fixed differential pressure fluctuation characteristic value range. The interval division is based on the normal range of differential pressure fluctuation, -0.5 to 0.5. The range is divided into 10 column intervals, each with a span of 0.1. For example, the first column interval is -0.5 to -0.4, the second column interval is -0.4 to -0.3, and so on. All column intervals together form a column interval set. The joint interval assignment determination process is as follows: extract the first element, i.e., the first component, from the two-dimensional addressing vector. Compare the first component with each row interval in the row interval set to determine whether the value of the first component falls between the upper and lower limits of a certain row interval. If it falls within a certain row interval, then the row interval is determined as the target row interval. At the same time, extract the second element, i.e., the second component, from the two-dimensional addressing vector. Compare the second component with each column interval in the column interval set to determine whether the value of the second component falls between the upper and lower limits of a certain column interval. If it falls within a certain column interval, then the column interval is determined as the target column interval. During the determination process, it is ensured that each component can uniquely fall into one interval, without falling into multiple intervals simultaneously or not falling into any interval.

[0099] Based on the spatial intersection of the target row interval and the target column interval, the address decoding of the storage unit of the feedforward table is triggered to obtain the feedforward adjustment amount of the feedforward table. First, the storage structure of the feedforward table is defined. The feedforward table adopts a row-column matrix storage, where each row interval corresponds to a row of the matrix, and each column interval corresponds to a column of the matrix. The intersection of the row and column corresponds to a unique storage unit. Each storage unit stores the corresponding feedforward adjustment amount in advance. This feedforward adjustment amount is a fixed value that is pre-calibrated and stored according to the working characteristics of the valve actuator, and is used to adjust the operating state of the valve actuator. The address decoding triggering process is to find the spatial intersection of the determined target row interval and target column interval in the matrix storage structure. This intersection position is the address of the corresponding storage unit. Then, the address decoding operation of the storage unit is triggered. The address decoding operation is to accurately locate the storage unit according to the storage unit address and read the pre-stored value in the storage unit. The read value is the feedforward adjustment amount of the feedforward table. During the triggering process, it is ensured that the address positioning is accurate and there are no positioning errors or reading errors. Finally, a unique feedforward adjustment amount is obtained.

[0100] First, the weighted segments corresponding to the amplitude of the opening deviation value are selected. Then, the opening deviation value at the current sampling time of the previously obtained standard sequence is extracted, and the amplitude of this opening deviation value is calculated. The amplitude calculation method is as follows: if the opening deviation value is positive, the amplitude is the opening deviation value itself; if the opening deviation value is negative, the amplitude is the absolute value of the opening deviation value, ensuring that the amplitude is always non-negative. Next, the preset weighted segment division rules are defined. The weighted segments are divided according to the amplitude range of the opening deviation value, resulting in a total of 5 weighted segments. Each segment corresponds to a fixed amplitude range, specifically as follows: amplitudes from 0 to 0.1 correspond to the... The system has five weighted segments: a first weighted segment with an amplitude of 0.1 to 0.2, a second weighted segment with an amplitude of 0.2 to 0.3, a third weighted segment with an amplitude of 0.3 to 0.4, a fourth weighted segment with an amplitude of 0.4 to 0.5, and a fifth weighted segment with an amplitude of 0.4 to 0.5. All weighted segments are continuous and non-overlapping, covering all possible amplitude ranges of the opening deviation value. The screening process involves comparing the calculated opening deviation value amplitude with the amplitude range of each weighted segment to determine which weighted segment the amplitude falls into and then selecting that weighted segment. This ensures that each amplitude uniquely corresponds to a weighted segment, with no screening errors or omissions.

[0101] The weight association table bound to the weight segment is decoded and output to obtain the collaborative weight value of the weight segment. First, it is clarified that each weight segment is pre-bound with a unique weight association table. This table is pre-built and stored, containing only the unique collaborative weight value corresponding to that weight segment. The collaborative weight value is set based on the influence of the opening deviation amplitude on the feedforward adjustment amount; the smaller the amplitude, the larger the collaborative weight value, and vice versa. Specifically, the collaborative weight value bound to the first weight segment is 0.9, and the collaborative weight value bound to the second weight segment is... The first weight is 0.7, the second weight is 0.5, the third weight is 0.3, and the fifth weight is 0.1. Each weight association table corresponds one-to-one with a weight segment, with no duplication or confusion. The decoding output process involves finding the weight association table bound to the selected weight segment, decoding the weight association table, which means reading the collaborative weight value stored in the weight association table and outputting the value directly. The output value is the collaborative weight value of the weight segment. During the decoding process, it is ensured that the weight association table matches accurately, with no decoding or output errors.

[0102] Based on the cooperative weight value, the feedforward adjustment amount is gain-scaled to obtain the scaled feedforward adjustment amount. First, the previously obtained cooperative weight value and feedforward adjustment amount are extracted to ensure that both are valid values. The cooperative weight value is a fixed value between 0.1 and 0.9, and the feedforward adjustment amount is a fixed adjustment value read from the feedforward table. The gain scaling process involves multiplying the feedforward adjustment amount by the cooperative weight value. During multiplication, the value of the feedforward adjustment amount is directly multiplied by the value of the cooperative weight value. The resulting product is the scaled feedforward adjustment amount. During the scaling process, the multiplication calculation is ensured to be accurate and without calculation errors. The scaled value can reflect the gain control effect of the cooperative weight value on the feedforward adjustment amount, so that the amplitude of the feedforward adjustment amount matches the opening deviation value.

[0103] The scaled feedforward adjustment and the opening deviation value are voltage superimposed to obtain the feedforward component of the valve actuator. First, the previously obtained scaled feedforward adjustment and the opening deviation value at the current sampling time are extracted to ensure that the dimensions of the two are consistent, both being dimensions related to the valve actuator control voltage, allowing for direct superposition. The voltage superposition process involves adding the scaled feedforward adjustment and the opening deviation value, considering both the positive and negative signs. If both values ​​are positive, the sum is the sum of the two values; if one is positive and the other is negative, the sum is the subtraction of the two values, with the sign consistent with the larger value; if both are negative, the sum is the addition of the two values ​​with a negative sign. The sum obtained after superposition is the feedforward component of the valve actuator. This feedforward component is used for subsequent precise adjustment of the valve actuator's operating state. The superposition process ensures accurate calculation without omitting any positive or negative signs, ultimately yielding a unique feedforward component.

[0104] The beneficial effects include: by combining the valve opening setpoint and differential pressure fluctuation characteristic value into a two-dimensional coordinate pair and generating a two-dimensional addressing vector, precise multi-dimensional addressing of the feedforward table can be achieved, quickly locating the target row and column intervals, accurately obtaining the matching feedforward adjustment amount, and improving the response speed and matching accuracy of feedforward regulation. Based on the opening deviation value amplitude, weight segments are selected and decoded to obtain cooperative weight values, enabling adaptive matching of weights, ensuring that the weight configuration highly matches the current deviation state. Scaling the feedforward adjustment amount with the cooperative weight value allows the feedforward regulation intensity to adapt to actual operating conditions. Then, voltage superposition merges the scaled feedforward adjustment amount with the opening deviation value to form a stable and single control voltage signal, obtaining the feedforward component. The entire process, through coherent processing of two-dimensional addressing, adaptive weighting, gain scaling, and signal superposition, achieves precise generation of the feedforward control amount, effectively suppressing control deviations caused by external disturbances, and improving the stability and response efficiency of the valve actuator's feedforward control.

[0105] IV. Construct a two-dimensional state point from the opening deviation value and the opening deviation change rate, and calculate the vertical projection distance of the two-dimensional state point on the preset compensation plane to obtain the projection distance of the two-dimensional state point.

[0106] In this embodiment of the invention, the step of constructing a two-dimensional state point from the opening deviation value and the opening deviation change rate, and calculating the vertical projection distance of the two-dimensional state point on a preset compensation plane to obtain the projection distance of the two-dimensional state point includes:

[0107] The opening deviation value is assigned to the first coordinate slot, and the opening deviation change rate is assigned to the second coordinate slot. The first coordinate slot and the second coordinate slot are then paired and anchored to obtain the two-dimensional state point, which includes the first coordinate value. Second coordinate value ;

[0108] By extracting coefficients from the planar expression of the preset compensation plane, the first normal coefficient of the preset compensation plane is obtained. Second normal coefficient and distance constant ;

[0109] Calculate the vertical projection distance from the two-dimensional state point to the preset compensation plane. The calculation formula is as follows:

[0110] ;

[0111] The vertical projection distance is... The first coordinate value of the two-dimensional state point is the opening deviation value. The second coordinate value of the two-dimensional state point is the opening deviation change rate. Let be the coefficient of the normal vector of the preset compensation plane in the first dimension. The coefficient of the normal vector of the preset compensation plane in the second dimension. The position constant of the preset compensation plane. The magnitude of the normal vector is used for distance normalization.

[0112] Fill the opening deviation value into the preset first coordinate slot, fill the opening deviation change rate into the preset second coordinate slot, and pair and lock the values ​​in the two coordinate slots in a fixed position to form fixed point data containing the first coordinate value and the second coordinate value, thus obtaining a two-dimensional state point.

[0113] From the plane expression of the preset compensation plane, the corresponding normal correlation coefficient and position correlation constant are extracted from the plane expression through hardware analysis circuit. The separated values ​​are determined as the first normal coefficient, the second normal coefficient and the distance constant, respectively, to obtain the first normal coefficient, the second normal coefficient and the distance constant of the preset compensation plane.

[0114] The hardware distance calculation circuit calculates the projected distance of the two-dimensional state point by using the coordinate values ​​of the two-dimensional state point, the first normal coefficient and the second normal coefficient of the preset compensation plane and the distance constant, according to the physical calculation logic of the perpendicular distance from the point to the plane.

[0115] The beneficial effects include filling the opening deviation value and the rate of change of opening deviation into fixed coordinate slots and locking them together, forming two-dimensional state points with clear positions and stable values. This provides standardized state input for compensation calculations, avoiding calculation errors caused by ambiguous state descriptions. By accurately extracting the first normal coefficient, second normal coefficient, and distance constant from the preset compensation plane through hardware analysis circuitry, the accuracy and reliability of compensation plane parameters are ensured, providing stable benchmark conditions for distance calculations. Based on the coordinates of the two-dimensional state point and the relevant constants of the compensation plane, the calculation is completed using the physical logic of the perpendicular distance from the point to the plane, resulting in a uniquely determined projected distance, providing an objective and unified basis for compensation positioning. This process, through standardized state construction, hardware parameter analysis, and physical distance calculation, ensures accurate and stable projected distance output, providing reliable support for subsequent compensation coefficient positioning and compensation component generation, and improving the accuracy and consistency of valve actuator compensation control.

[0116] V. Based on the projection distance, perform one-dimensional coordinate projection positioning on the compensation curve of the valve actuator to obtain the compensation coefficient of the projection distance, and modulate and couple the compensation coefficient with the opening deviation value to obtain the compensation component of the valve actuator;

[0117] In this embodiment of the invention, the step of performing one-dimensional coordinate projection positioning on the compensation curve of the valve actuator based on the projection distance to obtain the compensation coefficient of the projection distance includes:

[0118] The projection distance is projected onto the origin of a one-dimensional coordinate axis to obtain the normalized projection coordinate value of the projection distance;

[0119] Based on the normalized projection coordinate values, serial interval landing point matching is performed on the segmented boundary threshold sequence of the compensation curve, and the normalized projection coordinate values ​​and the threshold windows in the boundary threshold sequence are subjected to segment-by-segment membership adjudication to obtain the target segmented interval of the compensation curve.

[0120] Based on the ordinal number of the target segment interval, the coefficient storage slots bound to the ordinal number are selected to obtain the selected coefficient storage slots of the compensation curve.

[0121] The storage depth of the coefficient amplitude in the coefficient storage slot is captured, and the coefficient amplitude is extracted through the read channel activated by the gating switch to obtain the compensation coefficient of the projection distance.

[0122] The step of modulating and coupling the compensation coefficient with the opening deviation value to obtain the compensation component of the valve actuator includes:

[0123] Using the compensation coefficient as the amplitude control quantity and the opening deviation value as the carrier signal, the carrier signal is weighted and mapped based on the amplitude control quantity to obtain the weighted mapped amplitude-modulated signal of the carrier signal.

[0124] The amplitude-modulated signal is subjected to absolute value rectification transformation to obtain the rectified envelope signal of the weighted mapped amplitude-modulated signal;

[0125] The rectified envelope signal is subjected to low-pass smoothing filtering to remove the carrier frequency component in the amplitude modulation signal, thereby obtaining the compensation component of the valve actuator.

[0126] The projected distance is mapped to the position corresponding to the origin of the coordinate axis according to the scale rules of the one-dimensional coordinate axis. The projected distance is then converted into coordinate values ​​that conform to the scale range of the one-dimensional coordinate axis through a hardware normalization processing circuit, thus obtaining the normalized projected coordinate value of the projected distance.

[0127] The normalized projected coordinate values ​​are compared with the pre-set segment boundary threshold sequence of the compensation curve. The range of the normalized projected coordinate values ​​is determined one by one according to the order of the threshold sequence. The segment interval to which the normalized projected coordinate values ​​uniquely belong is determined, and the target segment interval of the compensation curve is obtained.

[0128] Read the fixed sequence number corresponding to the target segment interval, and use the hardware gating circuit to turn on the coefficient storage slot that has a fixed binding relationship with the sequence number, so that the slot is in a readable state, and obtain the selected coefficient storage slot of the compensation curve.

[0129] The hardware data reading circuit reads the values ​​stored in the selected coefficient storage slots, and outputs the complete coefficient values ​​inside the slots through the conductive reading channel to obtain the compensation coefficient for the projection distance.

[0130] The compensation coefficient is set as the amplitude control value, and the opening deviation value is set as the carrier signal. The amplitude of the carrier signal is adjusted according to the amplitude control value through the hardware weighted mapping circuit, so that the amplitude of the carrier signal and the compensation coefficient form a fixed proportional relationship, thus obtaining the weighted mapped amplitude modulation signal of the carrier signal.

[0131] The weighted mapped amplitude modulation signal is unidirectionally converted by a hardware rectifier circuit, which converts all inverse amplitudes in the signal into unidirectional amplitudes, preserving the overall amplitude change profile of the signal, and thus obtaining the rectified envelope signal of the weighted mapped amplitude modulation signal.

[0132] The rectified envelope signal is frequency-filtered by a hardware low-pass filter circuit, which completely blocks the carrier frequency components in the signal that are higher than the preset cutoff frequency, and retains only the low-frequency smooth signal that meets the control requirements, thus obtaining the compensation component of the valve actuator.

[0133] The beneficial effects include normalizing the projected distance and mapping it to one-dimensional coordinates, which unifies the compensation positioning under a standard scale and improves the stability of interval matching. By comparing the normalized coordinates with the segmented thresholds of the compensation curve segment by segment, the unique target segment interval can be accurately located. Then, by selecting the corresponding coefficient storage slot through sequence numbering, the matching compensation coefficient can be quickly read, achieving adaptive and precise matching of the compensation amount. Adjusting the opening deviation signal with the compensation coefficient allows the compensation intensity to adapt to the current deviation state. Rectification eliminates reverse amplitude interference and retains the effective change profile. Low-pass filtering removes high-frequency components to obtain a smooth control signal, ultimately generating a stable and reliable compensation component. The entire process relies on continuous execution of hardware circuitry, ensuring accurate positioning, rapid response, effective avoidance of signal distortion, and improved compensation control accuracy and smoothness, providing stable adaptive compensation support for valve actuators.

[0134] VI. Based on the historical performance evaluation value of the valve actuator in the previous cycle, the initial weight of the valve actuator is corrected according to performance orientation to obtain the current cycle weight of the valve actuator;

[0135] In this embodiment of the invention, the step of adjusting the initial weight of the valve actuator based on the historical performance evaluation value of the valve actuator in the previous cycle to obtain the current cycle weight of the valve actuator includes:

[0136] The previous cycle historical performance evaluation value of the valve actuator is used as the performance quantification index to classify the interval and obtain the correction level interval of the previous cycle.

[0137] Based on the performance correction parameter library of the valve actuator, the offset storage unit bound to the correction level interval is selected and collected by unit selection to obtain the weight offset of the correction level interval.

[0138] Symbolic weight injection is performed on the initial weight of the valve actuator and the weight offset to obtain the corrected intermediate weight of the initial weight;

[0139] The modified intermediate weights are subjected to boundary clamping constraints to obtain the current cycle weights of the valve actuator.

[0140] The historical performance evaluation value of the valve actuator in the previous cycle is used as the sole quantitative indicator of performance. It is compared with multiple fixed grade intervals that have been pre-divided according to the control performance standard to determine the unique grade interval into which the evaluation value falls, thus obtaining the correction grade interval for the previous cycle.

[0141] From the performance correction parameter library pre-established by the valve actuator, locate the weight offset storage unit that has a unique binding relationship with the correction level range, turn on the storage unit through the hardware gating circuit and read the internal data to obtain the weight offset of the correction level range.

[0142] The weight offset is directly superimposed on the initial weight value of the valve actuator according to its own positive or negative sign, so that the initial weight changes according to the performance orientation, forming a weight value after one adjustment, and thus obtaining the intermediate weight after the correction of the initial weight.

[0143] The corrected intermediate weight is compared with the preset minimum and maximum boundary values. When the corrected intermediate weight is less than the minimum boundary value, it is limited to the minimum boundary value. When the corrected intermediate weight is greater than the maximum boundary value, it is limited to the maximum boundary value. When the corrected intermediate weight is between the two boundary values, the original value is maintained. This completes the boundary clamping constraint and obtains the current cycle weight of the valve actuator.

[0144] The beneficial effects include matching historical performance evaluation values ​​from the previous cycle with fixed grading intervals, objectively determining correction level ranges, providing clear performance basis for weight correction, and avoiding ambiguity in correction direction. Selecting and binding weight offsets from the performance correction parameter library ensures precise correspondence between correction values ​​and control effects, improving the targeting of weight adjustments. Directly symbolically adding weight offsets to the initial weights enables rapid performance-oriented numerical correction, allowing control parameters to adapt to changes in operating conditions in real time. Clamping intermediate weights with minimum and maximum boundary values ​​prevents weight values ​​from exceeding reasonable ranges, ensuring control parameters remain within a safe and effective range. The overall process achieves closed-loop adaptive correction of weights, improving the rationality and stability of control parameters, ensuring the multivariable control of the valve actuator remains in optimal condition, and improving long-term adjustment accuracy and reliability.

[0145] VII. Based on the current cycle weight and the preset communication protocol, the feedforward component and the compensation component are encapsulated by protocol mapping to obtain the adaptive control command of the valve actuator;

[0146] In this embodiment of the invention, the step of performing protocol mapping and encapsulation on the feedforward component and the compensation component based on the current cycle weight and a preset communication protocol to obtain the adaptive control command of the valve actuator includes:

[0147] Based on the current cycle weight, the feedforward component and the compensation component are weighted and fused to obtain the composite control quantity of the feedforward component and the compensation component.

[0148] Based on the data field format of the preset communication protocol, the value range of the synthetic control quantity is mapped to the allowed value range of the protocol field to obtain the protocol payload of the synthetic control quantity.

[0149] Based on the frame structure definition of the preset communication protocol, the protocol payload is assembled into protocol frames to obtain a complete protocol frame of the protocol payload;

[0150] According to the bit order rules of the preset communication protocol, the complete protocol frame is serialized into a bit stream to obtain the adaptive control command of the valve actuator.

[0151] According to the current cycle weight setting ratio, the amplitude of the feedforward component and the compensation component are combined by the hardware fusion circuit, so that the two components are combined into a single control value according to a fixed weight ratio, and the composite control quantity of the feedforward component and the compensation component is obtained.

[0152] Based on the data field length and value range standards specified in the preset communication protocol, the value of the synthesized control quantity is converted into a value form that conforms to the protocol's allowed range through a hardware mapping circuit, so that the synthesized control quantity is adapted to the transmission requirements of the protocol field, and the protocol payload of the synthesized control quantity is obtained.

[0153] According to the frame structure definition of the preset communication protocol, the protocol payload is combined with the frame header, frame trailer, check bits and other parts specified by the protocol in sequence through the hardware frame assembly circuit to form a complete data frame that conforms to the protocol specification, and thus obtain the complete protocol frame of the protocol payload.

[0154] Following the bit order rules of the preset communication protocol, the parallel data of the complete protocol frame is converted into serial data transmitted in a fixed bit order through a hardware serial orchestration circuit, thus obtaining the adaptive control command of the valve actuator.

[0155] The beneficial effects include merging the amplitudes of the feedforward and compensation components according to the current cycle weights, enabling the two types of control components to be fused in the optimal ratio to obtain a stable and unified composite control quantity, ensuring the coordination of control output. Mapping the composite control quantity to the protocol payload according to the communication protocol standard allows the control values ​​to adapt to the transmission specifications, avoiding data overflow or parsing anomalies. Assembling the frame header, frame tail, and check bit according to the protocol frame structure forms a complete and standardized data frame, improving the reliability and recognizability of data transmission. Converting parallel data into serial data conforming to bit order rules adapts to industrial fieldbus transmission methods, ensuring stable command issuance and accurate execution. The entire process, through weight allocation, protocol mapping, frame assembly, and serial arrangement, achieves a complete conversion from control quantity to standard command, improving the output efficiency, transmission stability, and equipment compatibility of adaptive control commands, and ensuring the precise and reliable execution of valve actuator control actions.

[0156] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0157] This application embodiment can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0158] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A multivariable adaptive control method for a valve actuator, characterized in that, The method includes: I. Synchronously normalize the multi-source data of the valve actuator to obtain a standard sequence of the multi-source data; II. Perform deviation fluctuation characteristic decomposition on the standard sequence to obtain the opening deviation value, opening deviation change rate and pressure difference fluctuation characteristic value of the standard sequence; III. Based on the valve opening setpoint and the differential pressure fluctuation characteristic value of the multi-source data, perform multi-dimensional addressing matching on the feedforward table of the valve actuator to obtain the feedforward adjustment amount of the feedforward table, and perform nonlinear weight coordination between the feedforward adjustment amount and the opening deviation value to obtain the feedforward component of the valve actuator; IV. Construct a two-dimensional state point from the opening deviation value and the opening deviation change rate, and calculate the vertical projection distance of the two-dimensional state point on the preset compensation plane to obtain the projection distance of the two-dimensional state point. V. Based on the projection distance, perform one-dimensional coordinate projection positioning on the compensation curve of the valve actuator to obtain the compensation coefficient of the projection distance, and modulate and couple the compensation coefficient with the opening deviation value to obtain the compensation component of the valve actuator; VI. Based on the historical performance evaluation value of the valve actuator in the previous cycle, the initial weight of the valve actuator is corrected according to performance orientation to obtain the current cycle weight of the valve actuator; VII. Based on the current cycle weight and the preset communication protocol, the feedforward component and the compensation component are encapsulated by protocol mapping to obtain the adaptive control command of the valve actuator.

2. The multivariable adaptive control method for a valve actuator as described in claim 1, characterized in that, The process of synchronously normalizing the multi-source data from the valve actuator to obtain a standard sequence of the multi-source data includes: Time-scale consistency correction is performed on the signal timestamps of the valve actuator in the multi-source data to obtain the time-aligned multi-source data of the valve actuator. The values ​​of the data sources in the time-aligned multi-source data are mapped to a pre-stored normalized mapping table to obtain the standard values ​​of the data sources; Based on the chronological order of the timestamps, the standard values ​​are sorted in ascending order of time to obtain the standard sequence of the multi-source data.

3. The multivariable adaptive control method for a valve actuator as described in claim 2, characterized in that, The step of performing deviation fluctuation characteristic decomposition on the standard sequence to obtain the aperture deviation value, aperture deviation change rate, and differential pressure fluctuation characteristic value of the standard sequence includes: Based on the identification field of the data source, the standard sequence is decoupled from the signal channel to obtain the standard subsequence of the opening degree and the standard subsequence of the pressure difference. The current sampling time standard value in the opening standard subsequence is compared with the valve opening setting value of the multi-source data to determine the amplitude difference, thereby locking the opening deviation value of the current sampling time of the standard sequence. The standard value of the previous sampling time in the standard subsequence of valve opening is compared with the valve opening set value by time-series backtracking and difference identification to determine the opening deviation value of the standard sequence at the previous sampling time. The deviation slope of the current sampling time opening deviation value and the previous sampling time opening deviation value is captured to obtain the opening deviation change rate of the standard sequence. The pressure difference fluctuation characteristic value of the standard sequence is obtained by offset detection between the current sampling time standard value and the initial sampling time standard value in the pressure difference standard subsequence.

4. The multivariable adaptive control method for a valve actuator as described in claim 1, characterized in that, Based on the valve opening setpoint and the differential pressure fluctuation characteristic value from the multi-source data, multi-dimensional addressing matching is performed on the feedforward table of the valve actuator to obtain the feedforward adjustment amount of the feedforward table, including: The valve opening setting value of the multi-source data is taken as the first component, and the differential pressure fluctuation characteristic value is taken as the second component. The first component and the second component are bound in component order to obtain a two-dimensional coordinate pair of the first component and the second component. The two-dimensional coordinate pairs are vectorized and assembled to obtain the two-dimensional addressing vector of the feedforward table in the valve actuator; Based on the two-dimensional addressing vector, a joint interval assignment determination is performed on the row interval set and column interval set of the feedforward table, and the row interval into which the first component falls is taken as the target row interval, and the column interval into which the second component falls is taken as the target column interval. Based on the spatial intersection of the target row interval and the target column interval, address decoding is triggered on the storage unit of the feedforward table to obtain the feedforward adjustment amount of the feedforward table.

5. The multivariable adaptive control method for a valve actuator as described in claim 1, characterized in that, The step of performing nonlinear weighted coordination between the feedforward adjustment amount and the opening deviation value to obtain the feedforward component of the valve actuator includes: Filter out the weighted segments corresponding to the magnitude of the opening deviation value; The weight association table bound to the weight segment is decoded and output to obtain the collaborative weight value of the weight segment; Based on the cooperative weight value, the feedforward adjustment amount is gain-scaled to obtain the scaled feedforward adjustment amount. The scaled feedforward adjustment and the opening deviation value are superimposed by voltage to obtain the feedforward component of the valve actuator.

6. The multivariable adaptive control method for a valve actuator as described in claim 1, characterized in that, The step of constructing a two-dimensional state point from the opening deviation value and the opening deviation change rate, and calculating the vertical projection distance of the two-dimensional state point on the preset compensation plane to obtain the projection distance of the two-dimensional state point includes: The opening deviation value is assigned to the first coordinate slot, and the opening deviation change rate is assigned to the second coordinate slot. The first coordinate slot and the second coordinate slot are then paired and anchored to obtain the two-dimensional state point, which includes the first coordinate value. Second coordinate value ; By extracting coefficients from the planar expression of the preset compensation plane, the first normal coefficient of the preset compensation plane is obtained. Second normal coefficient and distance constant ; Calculate the vertical projection distance from the two-dimensional state point to the preset compensation plane. The calculation formula is as follows: ； The vertical projection distance is... The first coordinate value of the two-dimensional state point is the opening deviation value. The second coordinate value of the two-dimensional state point is the rate of change of the opening deviation. Let be the coefficient of the normal vector of the preset compensation plane in the first dimension. The coefficient of the normal vector of the preset compensation plane in the second dimension. The position constant of the preset compensation plane. The magnitude of the normal vector is used for distance normalization.

7. The multivariable adaptive control method for a valve actuator as described in claim 1, characterized in that, The step of performing one-dimensional coordinate projection positioning on the compensation curve of the valve actuator based on the projection distance to obtain the compensation coefficient of the projection distance includes: The projection distance is projected onto the origin of a one-dimensional coordinate axis to obtain the normalized projection coordinate value of the projection distance; Based on the normalized projection coordinate values, serial interval landing point matching is performed on the segmented boundary threshold sequence of the compensation curve, and the normalized projection coordinate values ​​and the threshold windows in the boundary threshold sequence are subjected to segment-by-segment membership adjudication to obtain the target segmented interval of the compensation curve. Based on the ordinal number of the target segment interval, the coefficient storage slots bound to the ordinal number are selected to obtain the selected coefficient storage slots of the compensation curve. The storage depth of the coefficient amplitude in the coefficient storage slot is captured, and the coefficient amplitude is extracted through the read channel activated by the gating switch to obtain the compensation coefficient of the projection distance.

8. The multivariable adaptive control method for a valve actuator as described in claim 1, characterized in that, The step of modulating and coupling the compensation coefficient with the opening deviation value to obtain the compensation component of the valve actuator includes: Using the compensation coefficient as the amplitude control quantity and the opening deviation value as the carrier signal, the carrier signal is weighted and mapped based on the amplitude control quantity to obtain the weighted mapped amplitude-modulated signal of the carrier signal. The amplitude-modulated signal is subjected to absolute value rectification transformation to obtain the rectified envelope signal of the weighted mapped amplitude-modulated signal; The rectified envelope signal is subjected to low-pass smoothing filtering to remove the carrier frequency component in the amplitude modulation signal, thereby obtaining the compensation component of the valve actuator.

9. The multivariable adaptive control method for a valve actuator as described in claim 1, characterized in that, The initial weights of the valve actuator are adjusted based on the historical performance evaluation value of the valve actuator in the previous cycle to obtain the current cycle weights of the valve actuator, including: The previous cycle historical performance evaluation value of the valve actuator is used as the performance quantification index to classify the interval and obtain the correction level interval of the previous cycle. Based on the performance correction parameter library of the valve actuator, the offset storage unit bound to the correction level interval is selected and collected by unit selection to obtain the weight offset of the correction level interval. Symbolic weight injection is performed on the initial weight of the valve actuator and the weight offset to obtain the corrected intermediate weight of the initial weight; The modified intermediate weights are subjected to boundary clamping constraints to obtain the current cycle weights of the valve actuator.

10. The multivariable adaptive control method for a valve actuator as described in claim 1, characterized in that, The adaptive control command for the valve actuator is obtained by protocol mapping and encapsulating the feedforward component and the compensation component based on the current cycle weight and a preset communication protocol, including: Based on the current cycle weight, the feedforward component and the compensation component are weighted and fused to obtain the composite control quantity of the feedforward component and the compensation component. Based on the data field format of the preset communication protocol, the value range of the synthetic control quantity is mapped to the allowed value range of the protocol field to obtain the protocol payload of the synthetic control quantity; Based on the frame structure definition of the preset communication protocol, the protocol payload is assembled into protocol frames to obtain a complete protocol frame of the protocol payload; According to the bit order rules of the preset communication protocol, the complete protocol frame is serialized into a bit stream to obtain the adaptive control command of the valve actuator.

Images