A pressure intelligent regulating system for butterfly valve and method thereof

Abstract

This invention relates to the field of industrial process control technology, and discloses a pressure intelligent regulation system and method for butterfly valves. The method includes acquiring spatiotemporally synchronized operating status data of the butterfly valve and establishing a dynamic pressure behavior model. By decomposing the pressure field of the model, the pressure components caused by changes in gas state and gas flow characteristics are separated. Based on the mechanical pressure components, a valve action influence chain is constructed, and reverse pressure disturbance tracing is performed to identify key valve action nodes causing pressure deviations. Based on this, an adaptive regulation strategy designed to counteract the disturbance is generated and compiled into a control command sequence and issued to the execution unit. This method can accurately identify the physical causes of pressure disturbances and trace them back to the specific mechanical action root cause, thereby achieving source compensation for pressure deviations and improving the accuracy of regulation and system stability.

Description

Technical Field

[0001] This invention relates to the field of industrial process control technology, specifically to a pressure intelligent regulation system and method for butterfly valves. Background Technology

[0002] In industrial process control systems, pressure regulation of butterfly valves is a crucial aspect of maintaining process stability. Currently, the commonly used technology is a closed-loop control strategy based on pressure sensor feedback. These methods treat the detected system pressure as a whole signal, and the controller directly calculates and outputs adjustment commands for the valve opening based on the deviation between this signal and the setpoint. Its control logic relies on a simplified model that establishes a direct correspondence between pressure and valve opening.

[0003] The core flaw of existing technical solutions lies in their inability to distinguish the different physical sources of pressure disturbance signals. Pressure changes in a real system are the combined result of gas dynamics and valve mechanical actions. Gas factors include flow fluctuations and pump disturbances; mechanical factors include the inertia of valve plate movement, friction, and actuator hysteresis. Traditional methods treat these two completely different types of pressure changes together, leading to the control system's inability to accurately compensate for complex disturbances. Its adjustment actions can only compensate for pressure deviations, rather than targeting the source of the deviation, easily causing problems such as regulation lag, overshoot oscillations, and even misoperation of valves to correct gas disturbances, resulting in unnecessary mechanical wear and energy consumption. Its fundamental limitation lies in the lack of decoupling capability for the components of the pressure signal and the lack of a causal tracing mechanism for how historical valve operations specifically affect the current pressure state.

[0004] There is a need for an intelligent regulation method that can isolate the causes of pressure disturbances and trace them back to the root cause of specific valve actions, so as to achieve adaptive control that can accurately offset them from the source. Summary of the Invention

[0005] The purpose of this invention is to provide a pressure intelligent regulation system and method for butterfly valves to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides a method for intelligent pressure regulation of a butterfly valve, the method comprising:

[0007] For the target butterfly valve, acquire a set of spatiotemporally synchronized butterfly valve operating status data;

[0008] Based on the spatiotemporally synchronized butterfly valve operating status data set, a dynamic pressure behavior model describing the response relationship between pressure changes and valve actions is established.

[0009] The dynamic pressure behavior model is subjected to pressure field decomposition processing to separate the pressure component caused by changes in gas state and the pressure component caused by changes in gas flow characteristics, and pressure component separation results are generated.

[0010] Based on the pressure component separation results, valve action influence chain analysis is performed on the pressure component caused by the change in gas flow characteristics to generate a valve action influence chain containing multiple influence nodes.

[0011] Using the valve action influence chain, reverse pressure disturbance tracing calculation is performed on the real-time pressure data and real-time opening data to identify the key valve action nodes that cause the pressure to deviate from the predetermined state and generate a pressure disturbance tracing report.

[0012] Based on the pressure disturbance source report, perform strategy generation calculation based on pressure field reconstruction to synthesize a butterfly valve adaptive adjustment strategy aimed at counteracting the pressure disturbance;

[0013] The adaptive adjustment strategy of the butterfly valve is compiled into a sequence of control instructions that can be executed by the butterfly valve controller, and the sequence of control instructions is sent to the pressure regulation execution unit of the target butterfly valve.

[0014] Preferably, the step of acquiring a spatiotemporally synchronized butterfly valve operating status data set for the target butterfly valve includes:

[0015] A multi-dimensional dynamic pressure field is constructed for the target butterfly valve. The multi-dimensional dynamic pressure field integrates real-time pressure data collected by the pressure sensor, real-time gas flow data collected by the gas flow meter, and real-time opening data fed back by the valve positioner.

[0016] Spatiotemporal mapping processing is performed within the multi-dimensional dynamic pressure field to map the real-time pressure data, real-time gas flow data, and real-time opening data to a unified spatiotemporal coordinate system, generating a spatiotemporally synchronized butterfly valve operating status data set.

[0017] The construction of the multi-dimensional dynamic pressure field for the target butterfly valve includes:

[0018] Raw pressure readings are periodically collected from multiple pressure sensing nodes distributed upstream and downstream of the target butterfly valve.

[0019] The original gas flow reading at the corresponding timestamp is synchronously collected from the gas flow metering unit connected in series with the target butterfly valve;

[0020] The valve positioning unit that drives the target butterfly valve synchronously acquires the original opening degree feedback reading at the corresponding timestamp;

[0021] The original pressure readings, original gas flow readings, and original opening feedback readings are subjected to time synchronization calibration to eliminate time offsets caused by differences in signal transmission paths and generate time-aligned original datasets.

[0022] Within a pre-defined virtual pressure field space, a five-dimensional coordinate system is assigned to each time-aligned raw data point, comprising time dimension, spatial location dimension, pressure value dimension, gas flow rate value dimension, and opening value dimension, thereby completing the construction of the multi-dimensional dynamic pressure field.

[0023] Preferably, the spatiotemporal mapping process performed within the multidimensional dynamic pressure field includes:

[0024] A series of time anchors are created based on the timestamp of each data point in the original time-aligned dataset;

[0025] The spatial location, pressure value, gas flow rate value, and opening value corresponding to each time anchor point are mapped onto a two-dimensional grid plane with time and pipeline location as axes to generate a preliminary spatiotemporal distribution map.

[0026] For data gaps caused by missing data or noise in the preliminary spatiotemporal distribution map, a spatiotemporal kriging interpolation algorithm is used to repair the data and generate a continuous spatiotemporal state surface.

[0027] The spatiotemporal consistency of the continuous spatiotemporal state surface is verified to ensure that the changes in pressure, gas flow rate and opening degree on any continuous spatiotemporal path conform to the preset physical continuity constraints. After verification, the spatiotemporally synchronized butterfly valve operation status data set is output.

[0028] Preferably, the establishment of a dynamic pressure behavior model describing the response relationship between pressure changes and valve action includes:

[0029] From the spatiotemporally synchronized butterfly valve operating status data set, extract pressure change event segments and corresponding valve opening change event segments;

[0030] Calculate the pressure change rate curve for each pressure change event segment, and the opening change rate curve for the corresponding valve opening change event segment;

[0031] Perform cross-correlation analysis on the pressure change rate curve and the opening change rate curve, calculate the cross-correlation coefficients under multiple time lags, and generate cross-correlation analysis results.

[0032] Based on the cross-correlation analysis results, the characteristic time lag that causes the cross-correlation coefficient to reach its peak value is identified. The characteristic time lag characterizes the delay characteristic from valve action to pressure response.

[0033] Based on the characteristic time lag and the shape of the pressure change rate curve and the opening change rate curve, a nonlinear transfer function is fitted, which is the core of the dynamic pressure behavior model.

[0034] Preferably, the step of performing pressure field decomposition processing on the dynamic pressure behavior model includes:

[0035] From the spatiotemporally synchronized butterfly valve operating status data set, extract pressure and gas flow data that do not include obvious valve action periods, and train a gas pressure sub-model to describe pressure fluctuations in a pure gas state.

[0036] The gas pressure sub-model is embedded into the dynamic pressure behavior model. The predicted pressure value generated by the gas pressure sub-model under the same input conditions is subtracted from the predicted output of the dynamic pressure behavior model to obtain the residual pressure component.

[0037] The residual pressure component is attributed to the change in gas flow characteristics, thereby completing the separation of the pressure component caused by the change in gas state and the pressure component caused by the change in gas flow characteristics, and obtaining the pressure component separation result.

[0038] Preferably, the step of performing valve action influence chain analysis on the pressure component caused by the change in gas flow characteristics includes:

[0039] From the spatiotemporally synchronized butterfly valve operating status data set, extract the valve opening change sequence that is temporally correlated with the pressure component caused by the change in gas flow characteristics;

[0040] The valve opening change sequence is discretized into a series of micro valve action nodes arranged in time sequence;

[0041] Analyze the immediate and delayed effects of each micro-valve actuation node on the pressure reading at the downstream pressure sensing node, and label the range and intensity of the effect for each micro-valve actuation node;

[0042] Based on the temporal sequence of the micro-valve action nodes and the mutual influence transmission relationship, they are connected into a directed graph structure, which is the valve action influence chain containing multiple influence nodes.

[0043] Preferably, the step of using the valve action influence chain to perform reverse pressure disturbance tracing calculation on the real-time pressure data and real-time opening data includes:

[0044] When a pressure anomaly event is detected that deviates from the predetermined state, the pressure anomaly characteristics of the pressure anomaly event are recorded, including the abnormal amplitude, the time of occurrence of the anomaly, and the spatial location of the anomaly.

[0045] In the valve action influence chain, a reverse search is performed on valve action nodes that occurred before the time of the anomaly and have a connected influence path to the spatial location of the anomaly, and these nodes are marked as candidate source nodes.

[0046] Calculate the pressure influence weight of each candidate tracing node on the anomaly occurrence time and anomaly spatial location. The pressure influence weight is obtained by propagation calculation based on the preset influence intensity in the valve action influence chain.

[0047] Candidate traceability nodes whose pressure influence weight exceeds a preset traceability threshold are selected and identified as the key valve action nodes.

[0048] The pressure disturbance source tracing report is generated by summarizing the action parameters, influence paths, and calculated pressure influence weights of all key valve action nodes.

[0049] Preferably, the execution of strategy generation calculation based on pressure field reconstruction includes:

[0050] The pressure disturbance source tracing report is analyzed to obtain the expected pressure correction amount for each key valve action node. The expected pressure correction amount is the correction amount that is opposite to the direction of the current pressure deviation.

[0051] Using the key valve action node as the strategy application point, a virtual compensating valve action is simulated and applied in the multi-dimensional dynamic pressure field, and the pressure change generated by the virtual action under the action of the dynamic pressure behavior model is calculated.

[0052] The amplitude and timing of the virtual compensating valve action are iteratively adjusted until the calculated pressure change matches the expected pressure correction within a preset tolerance range.

[0053] The amplitude, timing, and action node information of the virtual compensating valve action obtained in the final iteration are encoded into a preliminary regulation strategy.

[0054] The preliminary adjustment strategies are comprehensively optimized to ensure that the simultaneous execution of these strategies does not trigger new pressure conflicts in the multi-dimensional dynamic pressure field, and finally the adaptive adjustment strategy of the butterfly valve is synthesized.

[0055] Preferably, the comprehensive optimization of the multiple preliminary adjustment strategies includes:

[0056] Each initial adjustment strategy is defined as a valve action vector at a specific node and at a specific time on the valve action influence chain;

[0057] A strategy conflict detection model is established to simulate the pressure field superposition effect under the dynamic pressure behavior model when multiple valve action vectors are executed in parallel.

[0058] Calculate the pressure field superposition effect between any two valve action vectors, determine whether pressure oscillation or cancellation phenomenon exceeding the preset conflict threshold will occur, and mark the strategy pair with such phenomenon as conflict strategy pair.

[0059] To minimize the number of inter-strategy conflicts and maximize the overall pressure correction effect, a strategy combinatorial optimization problem is constructed.

[0060] Solving the strategy combination optimization problem yields a preliminary set of adjustment strategies that do not conflict with each other and have the best overall pressure correction effect. This preliminary set of adjustment strategies is the final adaptive adjustment strategy for the butterfly valve used in the synthesis.

[0061] Preferably, the present invention also includes a pressure intelligent regulation system for a butterfly valve, the system including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor, when executing the computer program, implements the steps of the pressure intelligent regulation method for a butterfly valve as described above.

[0062] Compared with the prior art, the beneficial effects of the present invention are:

[0063] By establishing a dynamic pressure behavior model and performing pressure field decomposition, the monitored comprehensive pressure field is separated into independent components caused by changes in gas state and gas flow characteristics. This technology achieves precise decoupling of the physical causes of pressure disturbances. The effect is that the control system can clearly identify whether the current pressure deviation originates from changes in external gas conditions or from the historical influence of valve actions. This provides a precise basis for formulating targeted regulation strategies, avoiding misjudging gas disturbances as valve malfunctions or using inappropriate mechanical actions for compensation, thus improving the targeting and accuracy of regulation.

[0064] Based on the separated valve mechanical pressure components, a valve action influence chain is constructed, and reverse pressure disturbance tracing calculations are performed. This technique establishes a causal mapping model from historical valve actions to the current pressure state and achieves reverse tracing. Its effect is the ability to accurately identify key historical action nodes that cause the current pressure to deviate from the predetermined state. This allows the control strategy to directly correct or compensate for the root cause of the problem, rather than simply responding to the final pressure deviation value. This achieves a shift in the regulation mode from "suppressing the phenomenon" to "eliminating the cause," improving system stability and regulation efficiency, and reducing ineffective or excessive regulation actions. Attached Figure Description

[0065] Figure 1 This is a schematic diagram illustrating the working principle of the intelligent pressure regulation method for butterfly valves described in this invention.

[0066] Figure 2 A flowchart for spatiotemporal mapping processing within a multidimensional dynamic pressure field;

[0067] Figure 3 A flowchart for performing pressure field decomposition on a dynamic pressure behavior model;

[0068] Figure 4 The convergence graph of the objective function for the butterfly valve compensation strategy iteration;

[0069] Figure 5 This is a heatmap showing the conflict coefficients among multiple strategies. Detailed Implementation

[0070] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0071] Please see Figure 1 This invention provides a method for intelligent pressure regulation of butterfly valves. The method includes: acquiring a spatiotemporally synchronized butterfly valve operating state data set for a target butterfly valve; establishing a dynamic pressure behavior model describing the response relationship between pressure changes and valve actions based on this data set; performing pressure field decomposition processing on the dynamic pressure behavior model to separate the pressure component caused by changes in gas state and the pressure component caused by changes in gas flow characteristics, generating a pressure component separation result; performing valve action influence chain analysis on the pressure component caused by changes in gas flow characteristics based on the pressure component separation result, generating a valve action influence chain containing multiple influence nodes; using the valve action influence chain, performing reverse pressure disturbance tracing calculation on real-time pressure data and real-time opening data to identify key valve action nodes that cause pressure deviation from the predetermined state, generating a pressure disturbance tracing report; performing strategy generation calculation based on pressure field reconstruction based on the pressure disturbance tracing report to synthesize a butterfly valve adaptive regulation strategy aimed at offsetting pressure disturbances; compiling the butterfly valve adaptive regulation strategy into a control command sequence executable by the butterfly valve controller, and sending the control command sequence to the pressure regulation execution unit of the target butterfly valve.

[0072] Example 1: For a target butterfly valve, a spatiotemporally synchronized butterfly valve operating status data set is obtained. This process includes constructing a multi-dimensional dynamic pressure field for the target butterfly valve. This multi-dimensional dynamic pressure field integrates real-time pressure data collected by pressure sensors, real-time gas flow data collected by gas flow meters, and real-time opening data fed back by the valve positioner. Spatiotemporal mapping processing is performed within the multi-dimensional dynamic pressure field to map the real-time pressure data, real-time gas flow data, and real-time opening data to a unified spatiotemporal coordinate system, generating a spatiotemporally synchronized butterfly valve operating status data set. Constructing the multi-dimensional dynamic pressure field for the target butterfly valve includes periodically collecting raw pressure readings from multiple pressure sensing nodes distributed upstream and downstream of the target butterfly valve. Simultaneously collecting raw gas flow readings at corresponding timestamps from a gas flow metering unit connected in series with the target butterfly valve. Simultaneously acquiring raw opening feedback readings at corresponding timestamps from the valve positioning unit driving the target butterfly valve. Time synchronization calibration processing is performed on the raw pressure readings, raw gas flow readings, and raw opening feedback readings to eliminate time offsets caused by differences in signal transmission paths, generating a time-aligned raw dataset. Within a pre-defined virtual pressure field space, each time-aligned raw data point is assigned a five-dimensional coordinate system containing time dimension, spatial location dimension, pressure value dimension, gas flow rate value dimension, and opening value dimension, thereby completing the construction of a multi-dimensional dynamic pressure field.

[0073] In practical implementation, the process of acquiring a spatiotemporally synchronized butterfly valve operating status data set for the target butterfly valve involves constructing a multi-dimensional dynamic pressure field that integrates multi-source data and performing data alignment and fusion within it. Constructing the multi-dimensional dynamic pressure field for the target butterfly valve is a fundamental step. This multi-dimensional dynamic pressure field integrates real-time pressure data collected by pressure sensors, real-time gas flow data collected by gas flow meters, and real-time opening data fed back by valve positioners. In some embodiments, the multi-dimensional dynamic pressure field is constructed as follows: raw pressure readings are collected from multiple pressure sensing nodes distributed upstream and downstream of the target butterfly valve at a fixed sampling period. Raw flow readings with corresponding timestamps are synchronously collected from a gas flow metering unit installed in series with the target butterfly valve. Raw opening feedback readings at the same timestamp are synchronously acquired from the valve positioning unit driving the target butterfly valve. Optionally, time synchronization calibration processing is performed on the raw pressure readings, raw gas flow readings, and raw opening feedback readings. This processing aims to eliminate time offsets caused by differences in different signal transmission paths, generating a time-aligned raw dataset. Within the preset virtual pressure field space, each data point in the time-aligned original dataset is assigned a unique coordinate. This coordinate is a five-dimensional coordinate that includes time dimension, spatial location dimension, pressure value dimension, gas flow value dimension, and opening value dimension, thereby completing the construction of a multi-dimensional dynamic pressure field.

[0074] In practical implementation, performing spatiotemporal mapping within a multi-dimensional dynamic pressure field is a key operation to unify the aforementioned multi-source heterogeneous data into a single analytical framework. In some embodiments, the spatiotemporal mapping process uses the timestamp of each data point in the time-aligned original dataset as a reference to create a series of equally or unequally spaced time anchor points. The spatial location information, pressure value, gas flow rate value, and opening value corresponding to each time anchor point are mapped onto a two-dimensional grid plane with absolute time as the horizontal axis and the physical location of the pipeline as the vertical axis, generating a preliminary spatiotemporal distribution map. Due to sensor failure or signal interference, the preliminary spatiotemporal distribution map may contain missing data points or noisy data points, forming data gaps. Optionally, a spatiotemporal kriging interpolation algorithm is used to repair the data gaps. This algorithm uses spatiotemporally adjacent valid data points to estimate the value of the gap location, generating a continuous spatiotemporal state surface. It is understandable that performing spatiotemporal consistency verification on the continuous spatiotemporal state surface is a necessary step. The verification process ensures that the changes in the three physical quantities of pressure, gas flow rate and opening degree meet the preset gas mechanics continuity constraints on any continuous spatiotemporal path. After verification, the final spatiotemporally synchronized butterfly valve operating status data set is output.

[0075] In practical implementation, time synchronization calibration requires precise calculation and compensation for clock deviations and transmission delays from each data source. One calibration method involves sending synchronization time pulse signals to all data acquisition units during system initialization and recording the time offset from when each unit receives the pulse to when the data is reported to the central processing unit. The timestamp of each subsequently acquired raw data reading is then subtracted from the time offset of its corresponding data acquisition unit to align all data with the central processing unit's time reference.

[0076] Example 2: See Figure 2 Within the multi-dimensional dynamic pressure field, spatiotemporal mapping processing is performed, including creating a series of time anchor points based on the timestamps of each data point in the time-aligned original dataset. The spatial location, pressure value, gas flow rate value, and opening degree value corresponding to each time anchor point are mapped onto a two-dimensional grid plane with time and pipeline location as axes, generating a preliminary spatiotemporal distribution map. Data gaps in the preliminary spatiotemporal distribution map caused by missing data or noise are repaired using a spatiotemporal kriging interpolation algorithm, generating a continuous spatiotemporal state surface. The continuous spatiotemporal state surface is then subjected to spatiotemporal consistency verification to ensure that changes in pressure, gas flow rate, and opening degree conform to preset physical continuity constraints along any continuous spatiotemporal path. Upon successful verification, a spatiotemporally synchronized butterfly valve operating status data set is output.

[0077] A dynamic pressure behavior model describing the response relationship between pressure changes and valve actions is established. This process includes extracting pressure change event segments and corresponding valve opening change event segments from a spatiotemporally synchronized butterfly valve operating state data set. The pressure change rate curve for each pressure change event segment and the corresponding valve opening change rate curve are calculated. Cross-correlation analysis is performed on the pressure and opening rate curves, calculating the cross-correlation coefficients under multiple time lags, generating cross-correlation analysis results. Based on the cross-correlation analysis results, the characteristic time lag that causes the cross-correlation coefficient to reach its peak is identified. This characteristic time lag characterizes the delay between valve action and pressure response. Based on the characteristic time lag and the shapes of the pressure and opening rate curves, a nonlinear transfer function is fitted; this nonlinear transfer function is the core of the dynamic pressure behavior model.

[0078] In practical implementation, spatiotemporal mapping processing and the establishment of a dynamic pressure behavior model are performed within a multi-dimensional dynamic pressure field. These two operations together lay the foundation for understanding the relationship between pressure and valve action. In practice, spatiotemporal mapping processing creates a series of time anchors based on the timestamps of each data point in the time-aligned raw dataset. For example, in an industrial vacuum pipeline gas pressure regulation scenario, time anchors can be set to one per second, corresponding to a set of pressure, flow rate, and opening snapshot data collected every second. The spatial location, pressure value, gas flow rate value, and opening value corresponding to each time anchor are mapped onto a two-dimensional grid plane with time and pipeline location as axes to generate a preliminary spatiotemporal distribution map. In practice, the horizontal axis of this grid is the timestamp sequence, and the vertical axis is the physical location coordinates of the sensors arranged along the pipeline. Each grid point stores the triplet data of pressure, gas flow rate, and opening at that specific time and location. In some embodiments, for pressure data gaps caused by sensor malfunctions in the initial spatiotemporal distribution map, a spatiotemporal kriging interpolation algorithm is used for data repair. The spatiotemporal kriging interpolation algorithm comprehensively considers valid data points that are temporally and spatially adjacent to the gap, and its interpolation weights... Determined by the following formula:

[0079]

[0080] in: Indicates the spatial location to be interpolated and time The estimated value on, Indicates the first The spatial location of the nearest valid data points and time The observed values ​​on These are weighting coefficients calculated from the spatiotemporal variogram model, which sum to 1. The summation sign is... The process involves analyzing neighboring valid data points. All voids are filled using a spatiotemporal kriging interpolation algorithm to generate a continuous spatiotemporal state surface. Optionally, spatiotemporal consistency verification is performed on the continuous spatiotemporal state surface. For example, it checks whether the rate of change of pressure between any two consecutive time anchor points and two adjacent spatial measurement points exceeds the theoretical limit of instantaneous gas change, and whether the product of gas flow rate and valve opening conforms to the continuity equation. After verification, a spatiotemporally synchronized butterfly valve operating status data set is output.

[0081] In practical implementation, establishing a dynamic pressure behavior model describing the response relationship between pressure changes and valve actions begins with extracting characteristic event segments from a spatiotemporally synchronized butterfly valve operating status data set. For example, from a ten-minute data segment, three event segments can be extracted where the valve opening jumps from 30% to 40%, along with three corresponding pressure rise event segments monitored at specific downstream measuring points. The pressure change rate curve for each pressure change event segment is calculated, i.e., the first derivative sequence of pressure with respect to time, along with the opening change rate curve for the corresponding valve opening change event segment. In some embodiments, cross-correlation analysis is performed on the pressure change rate curve and the opening change rate curve to calculate the cross-correlation coefficient under multiple time lags, generating cross-correlation analysis results. Based on the cross-correlation analysis results, the characteristic time lag that causes the cross-correlation coefficient to reach its peak is identified. It can be understood that in an industrial pipeline pressure control example, this characteristic time lag might be 2 seconds, indicating that approximately 2 seconds after the valve action command is issued, its main pressure effect appears at the downstream measuring point.

[0082] The dynamic pressure behavior model possesses adaptive adjustment capabilities, enabling it to perceive dynamic operating parameters such as gas flow fluctuations and pipeline pressure changes in real time, and automatically optimize the nonlinear transfer function parameters within the model. When abrupt changes occur in operating conditions, the model can quickly respond and update key parameters such as characteristic time lag, ensuring an accurate description of the relationship between pressure changes and valve action response under different dynamic operating conditions. This avoids the adjustment lag or overshoot problems that occur in fixed models under complex operating conditions, and always maintains the matching between the adjustment strategy and real-time operating conditions.

[0083] Example 3: See Figure 3During the operation of the dynamic pressure behavior model, early warning judgments of pressure changes are performed simultaneously. The model predicts pressure change trends over a future period in real time and compares these trends with preset pressure safety thresholds. When the predicted pressure may exceed the threshold range, an early warning is triggered. This warning information is linked to the pressure field decomposition stage, providing priority guidance for subsequent pressure component separation and disturbance source tracing. The dynamic pressure behavior model undergoes pressure field decomposition processing. This process includes extracting pressure and gas flow data from a spatiotemporally synchronized butterfly valve operating status dataset that does not contain significant valve operation periods, and training a gas pressure sub-model describing pressure fluctuations in a pure gas state. This gas pressure sub-model is embedded into the dynamic pressure behavior model. The predicted pressure value generated by the gas pressure sub-model under the same input conditions is subtracted from the predicted output of the dynamic pressure behavior model to obtain the residual pressure component. This residual pressure component is attributed to changes in gas flow characteristics, thus completing the separation of the pressure component caused by changes in gas state and the pressure component caused by changes in gas flow characteristics, yielding the pressure component separation result.

[0084] A valve action influence chain analysis is performed on the pressure component caused by changes in gas flow characteristics. This process involves extracting a sequence of valve opening changes that are temporally correlated with the pressure component caused by these changes from a spatiotemporally synchronized butterfly valve operating status data set. This sequence is then discretized into a series of micro-valve action nodes arranged in chronological order. The immediate and delayed effects of each micro-valve action node on the pressure readings at downstream pressure sensing nodes are analyzed, and the influence range and intensity of each node are labeled. Based on the temporal order and inter-node influence transmission relationships, these micro-valve action nodes are connected into a directed graph structure, which constitutes the valve action influence chain containing multiple influencing nodes.

[0085] In practical implementation, performing pressure field decomposition on the dynamic pressure behavior model requires first identifying and utilizing data from periods when the system is in a relatively stable operating condition. For example, in an industrial vacuum pipeline application scenario, when the valve opening changes by less than 1% over 300 consecutive seconds, it can be determined that this period does not contain significant valve action. Pressure and gas flow data sequences within this period without significant valve action are extracted from a spatiotemporally synchronized butterfly valve operating status dataset. These data are then used to train a gas pressure sub-model describing pressure fluctuations in a pure gas state. The gas pressure sub-model can be a pressure calculation model based on flow rate, gas density, and pipeline characteristics. In practical implementation, the gas pressure sub-model is embedded into the dynamic pressure behavior model. Under the same time series input conditions, the predicted pressure value generated by the gas pressure sub-model is subtracted from the predicted total pressure output of the dynamic pressure behavior model. Its mathematical expression is:

[0086]

[0087] in: This represents the residual pressure component at time t. This represents the total stress value predicted by the dynamic stress behavior model at time t. This represents the pressure value predicted by the gas pressure sub-model at time t based on current gas flow rate and other conditions. It can be understood that the residual pressure component... This is attributed to changes in gas flow characteristics, thereby completing the separation of the pressure component caused by changes in gas state and the pressure component caused by changes in gas flow characteristics, and obtaining the pressure component separation result.

[0088] In practical implementation, performing valve action influence chain analysis on the pressure component caused by changes in gas flow characteristics first requires correlation matching in the time dimension. In some embodiments, a valve opening change sequence that is temporally correlated with the pressure component caused by changes in gas flow characteristics is extracted from a spatiotemporally synchronized butterfly valve operating status data set. For example, if the pressure component separation result shows a significant residual pressure rise between timestamps T1 and T2, then the valve opening sample value sequence within the time window [T1-Δt, T2] is extracted, where Δt is a preset traceback time window. The valve opening change sequence is discretized into a series of micro valve action nodes arranged in chronological order. Each micro valve action node represents the amount and direction of opening change within a small time period. For example, a node can represent "at 10:01:03.200, the opening increased by 0.05% in 0.1 seconds". In some embodiments, the immediate and delayed effects of each micro-valve actuation node on the pressure reading at a specific downstream pressure sensing node are analyzed. For example, by calculating the transfer function or correlation between the actuation node and subsequent pressure changes, the influence range and intensity of each micro-valve actuation node are labeled. The influence range can be quantified as the spatial attenuation distance of the pressure response, and the influence intensity can be quantified as the magnitude of the resulting pressure change. Optionally, based on the temporal sequence and mutual influence transmission relationships of the micro-valve actuation nodes, they are connected into a directed graph structure. The nodes in the graph are micro-valve actuation nodes, and the directed edges represent the pressure influence of the previous node being transmitted and superimposed on the pressure field affected by the next node. This directed graph structure is a valve actuation influence chain containing multiple influencing nodes.

[0089] Example 4: Utilizing the valve action influence chain, reverse pressure disturbance tracing calculations are performed on real-time pressure data and real-time opening data. The process includes recording the pressure anomaly characteristics of the detected pressure anomaly event, including the anomaly amplitude, anomaly occurrence time, and anomaly spatial location. In the valve action influence chain, a reverse search is performed on valve action nodes that occurred before the anomaly occurrence time and have a connected influence path to the anomaly spatial location; these are marked as candidate tracing nodes. The pressure influence weight of each candidate tracing node on the anomaly occurrence time and anomaly spatial location is calculated. The pressure influence weight is obtained through propagation calculation based on the preset influence intensity in the valve action influence chain. Candidate tracing nodes with pressure influence weights exceeding a preset tracing threshold are selected and identified as key valve action nodes. The action parameters, influence paths, and calculated pressure influence weights of all key valve action nodes are summarized to generate a pressure disturbance tracing report.

[0090] The strategy generation calculation based on pressure field reconstruction is performed. This process includes parsing pressure disturbance source reports to obtain the expected pressure correction for each key valve action node. This expected pressure correction is the correction amount opposite to the current pressure deviation direction. Using the key valve action nodes as the policy application points, a virtual compensating valve action is simulated and applied in a multi-dimensional dynamic pressure field. The pressure change generated by the virtual action under the dynamic pressure behavior model is calculated. The amplitude and timing of the virtual compensating valve action are iteratively adjusted until the calculated pressure change matches the expected pressure correction within a preset tolerance range. The amplitude, timing, and application node information of the final iteratively obtained virtual compensating valve action are encoded into a preliminary adjustment strategy. Multiple preliminary adjustment strategies are comprehensively optimized to ensure that simultaneous execution of these strategies does not trigger new pressure conflicts within the multi-dimensional dynamic pressure field, ultimately synthesizing an adaptive adjustment strategy for the butterfly valve.

[0091] In practical implementation, the valve action influence chain is used to perform reverse pressure disturbance tracing calculations on real-time pressure data and real-time opening data. For example, in a scenario of controlling the outlet pressure of an industrial vacuum pipeline, when the system detects that the pressure at a downstream point deviates from the set value by more than a threshold at time T, a pressure anomaly event is triggered. The pressure anomaly characteristics of the pressure anomaly event are recorded, including the anomaly amplitude, the time of anomaly occurrence, and the spatial location of anomaly. In practical implementation, the valve action influence chain is searched for valve action nodes that occur before the anomaly occurrence time and have a connected influence path to the anomaly spatial location. For example, assuming the pressure anomaly occurs at 10:05:00 and the location is "downstream sensor S4", the valve action influence chain is traversed in reverse, marking all valve action nodes with timestamps before 10:05:00 and whose influence paths can reach "downstream sensor S4" according to the directed graph structure. These nodes are marked as candidate tracing nodes. Optionally, the pressure influence weight of each candidate source node on the anomaly occurrence time and spatial location is calculated. The pressure influence weight is obtained by propagation calculation based on the preset influence intensity in the valve action influence chain. The propagation calculation multiplies the influence intensity of upstream nodes along the influence path by an attenuation coefficient. In essence, candidate source nodes whose pressure influence weight exceeds a preset source threshold are selected and identified as key valve action nodes causing pressure anomalies due to changes in gas flow characteristics. Refer to Table 1, which shows a simplified excerpt of a pressure disturbance source tracing report.

[0092] Table 1: Excerpt from the Pressure Disturbance Source Tracing Report

[0093]

[0094] Summarize the action parameters, influence paths, and calculated pressure influence weights of all key valve action nodes to generate a complete pressure disturbance tracing report.

[0095] In practical implementation, the strategy generation calculation based on pressure field reconstruction begins with analyzing the pressure disturbance source report to obtain the expected pressure correction for each key valve action node. The expected pressure correction is numerically equal to and opposite in direction to the current pressure deviation. In some embodiments, the key valve action node is used as the strategy application point, and a virtual compensating valve action is simulated in the multi-dimensional dynamic pressure field. For example, for the key valve action node "Valve_Act_142" in Table 1, its original action is an increase in opening by 0.6%, so the virtual compensating valve action could be a reverse reduction in opening. The pressure change generated by the virtual action under the dynamic pressure behavior model is calculated. The amplitude and timing of the virtual compensating valve action are iteratively adjusted until the calculated pressure change matches the expected pressure correction within a preset tolerance range. The matching process minimizes the objective function. :

[0096]

[0097] in: It is the simulated pressure change. This represents the expected pressure correction. It can be understood that the amplitude, timing, and actuation node information of the virtual compensating valve action obtained from the final iteration are encoded into a preliminary control strategy. Optionally, multiple preliminary control strategies are comprehensively optimized to ensure that executing these strategies simultaneously does not trigger new pressure conflicts within the multi-dimensional dynamic pressure field, ultimately synthesizing an adaptive control strategy for the butterfly valve.

[0098] See Figure 4 This is a convergence graph of the objective function in the butterfly valve compensation strategy iteration. The blue curve represents the change in the objective function value of Valve_Act_138, and the black dashed line is the "tolerance threshold (0.1 kPa)," used to determine whether the strategy iteration has met the target. The objective function value decreases linearly with the number of iterations, reaching the tolerance threshold (0.1 kPa) in the second iteration and approaching 0 in the third iteration. The smaller the objective function value, the smaller the deviation between the "simulated pressure correction" and the "expected pressure correction," and the better the strategy performance. This graph is used for iterative optimization verification of the butterfly valve compensation strategy: by tracking the convergence process of the objective function, it confirms whether the compensation strategy matches the expected correction within the preset tolerance, which is a key evaluation criterion in the "pressure field reconstruction strategy generation" stage.

[0099] Example 5: Comprehensive optimization of multiple preliminary adjustment strategies. The process includes defining each preliminary adjustment strategy as a valve action vector at a specific node and time in the valve action influence chain. A strategy conflict detection model is established to simulate the pressure field superposition effect under a dynamic pressure behavior model when multiple valve action vectors are executed in parallel. The pressure field superposition effect between any two valve action vectors is calculated to determine whether pressure oscillations or cancellation phenomena exceeding a preset conflict threshold will occur. Strategy pairs exhibiting such phenomena are marked as conflicting strategy pairs. A strategy combination optimization problem is constructed with the objective of minimizing the number of conflicts between strategies and maximizing the overall pressure correction effect. Solving this strategy combination optimization problem yields a set of preliminary adjustment strategies that do not conflict with each other and have the optimal overall pressure correction effect. This set of preliminary adjustment strategies is the final adaptive adjustment strategy for the synthesized butterfly valve.

[0100] In practical implementation, the method described in this embodiment involves a process of comprehensively optimizing multiple preliminary adjustment strategies. This process aims to coordinate multiple compensatory actions to prevent mutual interference between strategies and achieve the overall optimal adjustment effect. In practice, the comprehensive optimization of multiple preliminary adjustment strategies first defines each preliminary adjustment strategy as a valve action vector at a specific node and at a specific time on the valve action influence chain. For example, in a complex vacuum pipeline network with multiple adjustment points, a preliminary adjustment strategy can be formally represented as a vector. ,in: This indicates the identifier of the action node in the chain affected by the valve action. Indicates the preset time for strategy execution. Indicates at node The amount of valve opening change that needs to be executed.

[0101] A strategy conflict detection model is established to simulate the pressure field superposition effect under a dynamic pressure behavior model when multiple valve action vectors are executed in parallel. The core of the strategy conflict detection model is to call the dynamic pressure behavior model, input a sequence of valve action vectors executed in parallel, and calculate the spatiotemporal distribution of pressure changes at all locations throughout the pipeline network and over a future period. In some embodiments, calculating the pressure field superposition effect between any two valve action vectors involves vector superimposing the pressure change fields generated by the individual execution of the two vectors and comparing it with the actual pressure change fields calculated by the dynamic pressure behavior model when they are executed sequentially or simultaneously. It is determined whether pressure oscillations or cancellation phenomena exceeding a preset conflict threshold will occur. For example, the preset conflict threshold can be a percentage of the pressure change amplitude. If the pressure fluctuation peak generated by the combined action of two valve action vectors at a certain point exceeds a certain proportion of the sum of their peak values ​​when they act alone, or causes the pressure change direction to be opposite, thus significantly weakening the regulation effect, then the strategy pair exhibiting such phenomena is marked as a conflicting strategy pair.

[0102] In some embodiments, a policy combination optimization problem is constructed with the objective of minimizing the number of inter-policy conflicts and maximizing the overall pressure adjustment effect. It can be understood that the decision variable in this optimization problem is a binary choice vector, representing whether each initial adjustment policy is ultimately adopted, and its objective function is... This can be expressed as:

[0103]

[0104] in: It is a binary decision variable; a value of 1 indicates the adoption of the first decision variable. The initial adjustment strategy is set to 0, indicating that it will not be adopted. It is an instruction strategy With strategy The coefficient indicates whether there is a conflict; if there is a conflict, it is 1, otherwise it is 0. Indicates the first The quantified value of the expected stress correction effect when the initial adjustment strategy is implemented alone; and These are the weighting coefficients that balance minimizing conflict with maximizing effect. The summation symbol iterates through all candidate initial adjustment strategies. Optionally, solving the strategy combinatorial optimization problem can employ an integer programming solver or a genetic algorithm to search for the objective function while satisfying all physical and logical constraints. The optimal combination of decision variables is determined by solving the strategy combination optimization problem. This yields a preliminary set of control strategies that do not conflict with each other and have the best overall pressure correction effect. This preliminary set of control strategies is the final adaptive control strategy for the butterfly valve used in the synthesis. Optionally, the comprehensive optimization process may also consider the actuator's operating frequency limitations, incorporating them as constraints into the optimization problem, such as limiting the maximum number of actions of a single valve per unit time. It can be understood that by solving this optimization problem, the final output adaptive control strategy for the butterfly valve is a set of coordinated valve control commands that can be executed in parallel or sequentially, aiming to systematically counteract identified pressure disturbances. Energy consumption control logic is integrated into the data fusion algorithm and control strategy generation process. The system statistically analyzes the energy consumption patterns of valve actions, combining the valve action frequency and opening change amplitude corresponding to different control strategies. While ensuring pressure regulation accuracy, it prioritizes the strategy combination with lower energy consumption. Simultaneously, by optimizing the timing and amplitude of valve actions, unnecessary frequent actions are reduced, lowering the energy consumption of the actuator. The system regularly summarizes energy consumption data and feeds it back to the regulation strategy generation stage, continuously optimizing strategy parameters to achieve the dual goals of precise pressure regulation and low-energy operation. The system supports human-machine interaction, allowing operators to view key information such as the dynamic pressure behavior model's operating status, pressure field decomposition results, and regulation strategy execution through a visual interface. Operators can also manually input parameters such as pressure thresholds and regulation priorities to intervene in the regulation process. Simultaneously, the system possesses remote data transmission and monitoring capabilities, enabling the remote transmission of real-time pressure data, early warning information, and energy consumption statistics to the monitoring center via the network, achieving remote real-time monitoring of the butterfly valve's pressure regulation process. When an abnormal warning occurs, the monitoring center can remotely issue emergency regulation commands to ensure stable system operation even in unattended scenarios.

[0105] See Figure 5This is a heatmap of conflict coefficients among multiple strategies. Both rows and columns contain six strategies (Strategy 1-Strategy 6), and the cell values ​​represent the "conflict coefficient." Darker colors indicate higher conflict coefficients (0.8 being the darkest), corresponding to the color scale on the right. This graph is used for conflict detection in the adaptive control strategies of butterfly valves: by visualizing the conflict intensity between strategies, it filters out strategy combinations with no or low conflict, serving as a key tool in the "strategy comprehensive optimization" stage and preventing new pressure disturbances caused by the parallel execution of multiple strategies.

[0106] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0107] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent pressure regulation of a butterfly valve, characterized in that, The method includes: For the target butterfly valve, acquire a set of spatiotemporally synchronized butterfly valve operating status data; Based on the spatiotemporally synchronized butterfly valve operating status data set, a dynamic pressure behavior model describing the response relationship between pressure changes and valve actions is established. The dynamic pressure behavior model is subjected to pressure field decomposition processing to separate the pressure component caused by changes in gas state and the pressure component caused by changes in gas flow characteristics, and pressure component separation results are generated. Based on the pressure component separation results, valve action influence chain analysis is performed on the pressure component caused by the change in gas flow characteristics to generate a valve action influence chain containing multiple influence nodes. Using the valve action influence chain, reverse pressure disturbance tracing calculations are performed on real-time pressure data and real-time opening data to identify key valve action nodes that cause pressure to deviate from the predetermined state and generate a pressure disturbance tracing report. Based on the pressure disturbance source report, perform strategy generation calculation based on pressure field reconstruction to synthesize a butterfly valve adaptive adjustment strategy aimed at counteracting the pressure disturbance; The adaptive adjustment strategy of the butterfly valve is compiled into a sequence of control instructions that can be executed by the butterfly valve controller, and the sequence of control instructions is sent to the pressure regulation execution unit of the target butterfly valve.

2. The intelligent pressure regulation method for a butterfly valve according to claim 1, characterized in that, The acquisition of a spatiotemporally synchronized butterfly valve operating status data set for the target butterfly valve includes: A multi-dimensional dynamic pressure field is constructed for the target butterfly valve. The multi-dimensional dynamic pressure field integrates real-time pressure data collected by the pressure sensor, real-time gas flow data collected by the gas flow meter, and real-time opening data fed back by the valve positioner. Spatiotemporal mapping processing is performed within the multi-dimensional dynamic pressure field to map the real-time pressure data, real-time gas flow data, and real-time opening data to a unified spatiotemporal coordinate system, generating a spatiotemporally synchronized butterfly valve operating status data set. The construction of the multi-dimensional dynamic pressure field for the target butterfly valve includes: Raw pressure readings are periodically collected from multiple pressure sensing nodes distributed upstream and downstream of the target butterfly valve. The original gas flow reading at the corresponding timestamp is synchronously collected from the gas flow metering unit connected in series with the target butterfly valve; The valve positioning unit that drives the target butterfly valve synchronously acquires the original opening degree feedback reading at the corresponding timestamp; The original pressure readings, original gas flow readings, and original opening feedback readings are subjected to time synchronization calibration to eliminate time offsets caused by differences in signal transmission paths and generate time-aligned original datasets. Within a pre-defined virtual pressure field space, a five-dimensional coordinate system is assigned to each time-aligned raw data point, comprising time dimension, spatial location dimension, pressure value dimension, gas flow rate value dimension, and opening value dimension, thereby completing the construction of the multi-dimensional dynamic pressure field.

3. The intelligent pressure regulation method for a butterfly valve according to claim 2, characterized in that, The spatiotemporal mapping process performed within the multidimensional dynamic pressure field includes: A series of time anchors are created based on the timestamp of each data point in the original time-aligned dataset; The spatial location, pressure value, gas flow rate value, and opening value corresponding to each time anchor point are mapped onto a two-dimensional grid plane with time and pipeline location as axes to generate a preliminary spatiotemporal distribution map. For data gaps caused by missing data or noise in the preliminary spatiotemporal distribution map, a spatiotemporal kriging interpolation algorithm is used to repair the data and generate a continuous spatiotemporal state surface. The spatiotemporal consistency of the continuous spatiotemporal state surface is verified to ensure that the changes in pressure, gas flow rate and opening degree on any continuous spatiotemporal path conform to the preset physical continuity constraints. After verification, the spatiotemporally synchronized butterfly valve operation status data set is output.

4. The intelligent pressure regulation method for a butterfly valve according to claim 1, characterized in that, The establishment of a dynamic pressure behavior model describing the response relationship between pressure changes and valve action includes: From the spatiotemporally synchronized butterfly valve operating status data set, extract pressure change event segments and corresponding valve opening change event segments; Calculate the pressure change rate curve for each pressure change event segment, and the opening change rate curve for the corresponding valve opening change event segment; Perform cross-correlation analysis on the pressure change rate curve and the opening change rate curve, calculate the cross-correlation coefficients under multiple time lags, and generate cross-correlation analysis results. Based on the cross-correlation analysis results, the characteristic time lag that causes the cross-correlation coefficient to reach its peak value is identified. The characteristic time lag characterizes the delay characteristic from valve action to pressure response. Based on the characteristic time lag and the shape of the pressure change rate curve and the opening change rate curve, a nonlinear transfer function is fitted, which is the core of the dynamic pressure behavior model.

5. The intelligent pressure regulation method for a butterfly valve according to claim 1, characterized in that, The process of performing pressure field decomposition on the dynamic pressure behavior model includes: From the spatiotemporally synchronized butterfly valve operating status data set, extract pressure and gas flow data that do not include obvious valve action periods, and train a gas pressure sub-model to describe pressure fluctuations in a pure gas state. The gas pressure sub-model is embedded into the dynamic pressure behavior model. The predicted pressure value generated by the gas pressure sub-model under the same input conditions is subtracted from the predicted output of the dynamic pressure behavior model to obtain the residual pressure component. The residual pressure component is attributed to the change in gas flow characteristics, thereby completing the separation of the pressure component caused by the change in gas state and the pressure component caused by the change in gas flow characteristics, and obtaining the pressure component separation result.

6. The intelligent pressure regulation method for a butterfly valve according to claim 1, characterized in that, The valve action influence chain analysis performed on the pressure component caused by the change in gas flow characteristics includes: From the spatiotemporally synchronized butterfly valve operating status data set, extract the valve opening change sequence that is temporally correlated with the pressure component caused by the change in gas flow characteristics; The valve opening change sequence is discretized into a series of micro valve action nodes arranged in time sequence; Analyze the immediate and delayed effects of each micro-valve actuation node on the pressure reading at the downstream pressure sensing node, and label the range and intensity of the effect for each micro-valve actuation node; Based on the temporal sequence of the micro-valve action nodes and the mutual influence transmission relationship, they are connected into a directed graph structure, which is the valve action influence chain containing multiple influence nodes.

7. The intelligent pressure regulation method for a butterfly valve according to claim 1, characterized in that, The step of using the valve action influence chain to perform reverse pressure disturbance tracing calculation on the real-time pressure data and real-time opening data includes: When a pressure anomaly event is detected that deviates from the predetermined state, the pressure anomaly characteristics of the pressure anomaly event are recorded, including the abnormal amplitude, the time of occurrence of the anomaly, and the spatial location of the anomaly. In the valve action influence chain, a reverse search is performed on valve action nodes that occurred before the time of the anomaly and have a connected influence path to the spatial location of the anomaly, and these nodes are marked as candidate source nodes. Calculate the pressure influence weight of each candidate tracing node on the anomaly occurrence time and anomaly spatial location. The pressure influence weight is obtained by propagation calculation based on the preset influence intensity in the valve action influence chain. Candidate traceability nodes whose pressure influence weight exceeds a preset traceability threshold are selected and identified as the key valve action nodes. The pressure disturbance source tracing report is generated by summarizing the action parameters, influence paths, and calculated pressure influence weights of all key valve action nodes.

8. The intelligent pressure regulation method for a butterfly valve according to claim 7, characterized in that, The execution of strategy generation calculation based on pressure field reconstruction includes: The pressure disturbance source tracing report is analyzed to obtain the expected pressure correction amount for each key valve action node. The expected pressure correction amount is the correction amount that is opposite to the direction of the current pressure deviation. Using the key valve action node as the strategy application point, a virtual compensating valve action is simulated and applied in a multi-dimensional dynamic pressure field, and the pressure change generated by the virtual action under the action of the dynamic pressure behavior model is calculated. The amplitude and timing of the virtual compensating valve action are iteratively adjusted until the calculated pressure change matches the expected pressure correction within a preset tolerance range. The amplitude, timing, and action node information of the virtual compensating valve action obtained in the final iteration are encoded into a preliminary regulation strategy. The preliminary adjustment strategies are comprehensively optimized to ensure that the simultaneous execution of these strategies does not trigger new pressure conflicts in the multi-dimensional dynamic pressure field, and finally the adaptive adjustment strategy of the butterfly valve is synthesized.

9. The intelligent pressure regulation method for a butterfly valve according to claim 8, characterized in that, The comprehensive optimization of the multiple preliminary adjustment strategies includes: Each initial adjustment strategy is defined as a valve action vector at a specific node and at a specific time on the valve action influence chain; A strategy conflict detection model is established to simulate the pressure field superposition effect under the dynamic pressure behavior model when multiple valve action vectors are executed in parallel. Calculate the pressure field superposition effect between any two valve action vectors, determine whether pressure oscillation or cancellation phenomenon exceeding the preset conflict threshold will occur, and mark the strategy pair with such phenomenon as conflict strategy pair. To minimize the number of inter-strategy conflicts and maximize the overall pressure correction effect, a strategy combinatorial optimization problem is constructed. Solving the strategy combination optimization problem yields a preliminary set of adjustment strategies that do not conflict with each other and have the best overall pressure correction effect. This preliminary set of adjustment strategies is the final adaptive adjustment strategy for the butterfly valve used in the synthesis.

10. A pressure intelligent regulation system for a butterfly valve, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the intelligent pressure regulation method for butterfly valves as described in any one of claims 1 to 9.

Images