Valve detection control method based on front and back ends

Abstract

The application discloses a valve detection control method based on front and rear ends, and particularly relates to the field of control valve online detection, which is used for solving the problem that the state degradation of inventory valves in long-term operation is difficult to be quantized in time and applied to real-time control; through executing small-stroke exploration in the front terminal, the valve rod angular displacement and the actuator driving amount are synchronously collected under a unified time reference, the stroke segmentation and the segment-level characteristic parameter set are obtained through incremental segmented fitting, the segment-level characteristic parameter set and process constraint data are modeled in the stroke segmentation dimension by the rear-end platform to generate a health constraint parameter set and issue the front-end terminal, the front-end terminal projects the target opening degree instruction according to the health constraint parameter set, limits the action speed, and combines the angular deviation continuous monitoring, the rollback control based on the historical safe position and the abnormal event reporting to realize the integrated management of the online diagnosis result and the control behavior of the inventory valve.

Description

Technical Field

[0001] This invention relates to the field of online control valve testing, and more specifically, to a valve testing and control method based on both the front and back ends. Background Technology

[0002] In pipeline systems used for long-term operation, such as water, gas, and oil transmission and distribution, as well as industrial installations, a large number of valves in service are gate valves, ball valves, butterfly valves, or globe valves that were configured locally years ago according to operating conditions. Their nameplate information, structural parameters, and factory specifications are often incomplete, and the original settings of some equipment are difficult to verify after multiple maintenance and modifications. These existing valves are often in high-pressure or corrosive media environments. If they experience slow operation, jamming, incomplete closure, or malfunction, it can lead to pressure imbalance, shutdown of critical units, increased leakage, and increased energy consumption, posing a threat to production safety and the stability of public services. Given rising labor costs and increasingly stringent regulations, maintenance units hope to equip existing valves with online monitoring and remote control functions, enabling visualization of operational status and proactive risk management, while minimizing or eliminating production disruptions.

[0003] In existing technologies, digital valve positioners, intelligent actuators, and partial stroke testing devices for safety shut-off valves are widely used for newly built or parameter-defined control valves. Under the premise of known valve type and nominal stroke, the stroke endpoint, friction state, and actuation capability are obtained through full stroke self-calibration or preset stroke test, and the diagnostic results are uploaded to the host computer or asset management system for analysis. At the same time, non-invasive monitoring devices that are clamped or externally attached to the valve handwheel, valve stem, or pipeline wall have also emerged to collect switch status or vibration signals and realize remote status monitoring. These solutions work well on newly built equipment or equipment with clear parameters. However, in the scenario of online upgrade of existing valves with a large scale, many models and lack of design data, if full-stroke self-calibration or large-stroke testing is directly adopted, it is easy to cause pressure fluctuations and flow changes in the pressurized pipeline network, which is difficult to perform frequently on site. If only the above-mentioned non-invasive monitoring devices are relied upon, the relationship between the driving amount and valve stem displacement of the valve in different opening ranges is not detailed enough, making it difficult to identify dead zone, high friction zone and its stroke boundary in a timely manner. Moreover, the diagnostic results are mostly in the form of alarms or maintenance suggestions, which are not tightly coupled with the process control system. When the control command is executed, it is difficult to avoid the known high-risk stroke range in a timely manner.

[0004] Under the aforementioned engineering conditions and application constraints, the core technical problem that urgently needs to be solved is how to obtain segmented parameter information that can reflect the mechanical and response characteristics of different stroke ranges using only a small number of externally measurable signals such as drive quantity and valve stem displacement, without interrupting or minimizing production and keeping the valve body structure unchanged, and how to construct safe operation constraints that can be directly adopted by process control systems on existing valves lacking original design data.

[0005] To address the aforementioned problems, a technical solution is provided. Summary of the Invention

[0006] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a valve detection and control method based on front-end and back-end. This method involves performing small-stroke trials at the front-end terminal, synchronously acquiring valve stem angular displacement and actuator drive quantity under a unified time reference. Incremental segmentation fitting yields stroke segments and a set of segment-level characteristic parameters. The back-end platform then models the set of segment-level characteristic parameters and process constraint data along with the stroke segment dimension to generate a set of health constraint parameters, which is then sent to the front-end terminal. The front-end terminal projects the target opening command into an angle and limits the action speed based on the health constraint parameter set. Combined with continuous monitoring of angle deviation, backtracking control based on historical safe positions, and reporting of abnormal events, this method achieves integrated management of online diagnostic results and control behavior for existing valves, thereby solving the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] S1: After the front-end terminal is powered on, it performs a self-test and zero-position calibration. Within the small stroke of the valve, it drives the valve stem to reciprocate micro-movements according to a preset test sequence. Under a unified time reference, it synchronously collects the valve stem angular displacement and the actuator driving amount and stores them in time order to form small stroke test data.

[0009] S2: The front-end terminal preprocesses the small stroke trial data according to a uniform time step, calculates the driving amount and displacement increment of adjacent sampling points, and performs change point detection and segment fitting to obtain multiple stroke segments on the valve stem stroke. For each stroke segment, it calculates the segment-level characteristic parameters reflecting friction and hysteresis, generates a set of segment-level characteristic parameters, and uploads it to the back-end platform.

[0010] S3: The backend platform receives the set of segment-level characteristic parameters and organizes them in the stroke segment dimension in accordance with the process constraint data of the circuit to which the valve belongs. Based on the simplified working condition model, it calculates the friction characteristics and flow capacity of each stroke segment, obtains the upper limit of safe opening and the maximum allowable action speed of each stroke segment, forms a set of health constraint parameters and sends them to the front-end terminal.

[0011] S4: When the front-end terminal receives the target opening instruction, it converts the target opening into the target angle and locates the target angle on the stroke segment. Based on the health constraint parameter set, it corrects the target angle and action speed to generate the angle trajectory. During the execution process, it monitors the angle deviation by sampling. When the angle deviation continues to exceed the set standard, it performs backtracking control based on the last safe position and reports the abnormal event to the back-end platform.

[0012] Furthermore, step S1 includes: when generating the preset test sequence, the front-end terminal drives the valve stem to reciprocate micro-motion in a local range of the valve's full stroke with the current working position as the center, and after collecting the valve stem angular displacement and the actuator driving amount, the small stroke test data with driving amount and displacement changes below the threshold are discarded, and the small stroke test data used for stroke segmentation processing are retained.

[0013] Furthermore, step S1 also includes: after the front-end terminal completes the zero-position calibration, it uses the mechanical limit or small-amplitude reciprocating rotation of the clamping mechanism to determine the reference point of the angle sensor, and assigns a unique test number to each test action under a unified time reference. At the same time, it records the test action direction, test start time, test end time, and stores them together with the small-stroke test data in the front-end local buffer.

[0014] Furthermore, step S2 includes: when the front-end terminal preprocesses the small-stroke trial data, it resamples the valve stem angular displacement and actuator drive quantity according to a uniform time step, reduces high-frequency noise through digital filtering, performs amplitude normalization processing on the resampling results, and then calculates the drive quantity increment and displacement increment of adjacent sampling points to construct an incremental sequence for change point detection.

[0015] Furthermore, step S2 also includes: after the front-end terminal completes the change point detection and segment fitting, it records the multiple stroke segments obtained by dividing the valve stem stroke in the stroke segment division table according to the stroke segment number, and writes the average slope, static friction peak value, hysteresis width and stroke segment type label corresponding to each stroke segment into the segment-level characteristic parameter set, and stores it with the stroke segment number as the index before uploading it to the back-end platform.

[0016] Furthermore, step S3 includes: the process constraint data includes at least the upstream pressure allowable range, downstream pressure allowable range, allowable differential pressure upper limit, target flow allowable range, and medium density of the circuit to which the valve belongs. The back-end platform writes the segment-level characteristic parameter set and the process constraint data into the same joint data structure in the stroke segment dimension to build a simplified operating condition model.

[0017] Furthermore, step S3 also includes: the backend platform calculates health indicators representing friction state and driving force requirements for each stroke segment based on the joint data structure, and classifies the risk level of the stroke segment according to the pre-established threshold range. The health indicators and risk levels, together with the upper limit of the safe opening of the stroke segment and the maximum allowable action speed of the stroke segment, form the stroke segment record in the health constraint parameter set.

[0018] Furthermore, step S4 includes: when the front-end terminal generates an angle trajectory by correcting the target angle and motion speed according to the health constraint parameter set, it first safely projects the target angle onto all stroke segments involved in the motion path to obtain a safe target angle, and then selects the maximum motion speed of the path from the maximum allowable motion speed of each stroke segment, and constructs a planned angle trajectory consisting of an acceleration segment, a constant speed segment, and a deceleration segment according to the safe target angle, the maximum motion speed of the path, and the sampling period.

[0019] Furthermore, step S4 also includes: when the front-end terminal monitors the angle deviation by sampling, it performs multiple reference action tests on the health valve in advance, statistically analyzes the distribution of the absolute value of the angle deviation at each sampling time, determines the angle deviation threshold based on the distribution, statistically analyzes the distribution of the number of consecutive out-of-tolerances in the health action, determines the standard for the number of consecutive out-of-tolerances based on the distribution, and uses the angle deviation threshold and the standard for the number of consecutive out-of-tolerances as the abnormal triggering conditions during online operation.

[0020] Furthermore, step S4 also includes: when the angle deviation continues to exceed the threshold and triggers backoff control, the front-end terminal records the actual angle position where the previous action ended and the angle deviation did not exceed the angle deviation threshold as the previous safe position. During the backoff control process, the previous safe position is used as the backoff endpoint, and an abnormal event record is generated after the backoff ends. The abnormal event record includes at least the original target opening instruction, the safe target angle, the action path travel segment set, the maximum angle deviation, the backoff start angle position, and the backoff end angle position, and is uploaded to the back-end platform.

[0021] The technical effects and advantages of this invention based on the valve detection and control method at both the front and back ends are as follows:

[0022] This invention, through the synergy of front-end small-stroke probing, stroke segment modeling, and back-end operating condition constraint calculation, unifies the frictional characteristics and hysteresis characteristics of the valve stem in different angle ranges, along with process parameters such as circuit pressure and flow rate, into a single stroke segment coordinate system. This forms a set of health constraint parameters with clearly defined angular boundaries, which is then distributed to the front-end terminal. When executing the target opening command, the front-end no longer acts solely based on a single set value. Instead, it first performs a safety projection on the stroke segments, then combines this with the maximum permissible action speed of each stroke segment along the path to generate an angular trajectory. The diagnostic results are directly converted into executable control constraints, realizing a coherent chain from data acquisition and status recognition to the implementation of control commands.

[0023] Meanwhile, this invention introduces continuous angle deviation monitoring and an automatic backoff mechanism based on historical safe positions at the front-end terminal. When the valve exhibits abnormal trends such as jamming or loss of synchronization within a certain stroke segment, it can immediately terminate the original planned trajectory at the local control level and backoff to a verified safe angle position. The abnormal event record, including the target command, stroke segment information, and deviation statistics, is then uploaded to the back-end. Through this combination of segmented health constraints and closed-loop abnormal reporting, this invention enables refined management of aging valves and existing valves under complex operating conditions without interrupting production or relying on frequent disassembly and inspection. This ensures that the valves operate within a safe working range that matches their inherent condition over the long term. Attached Figure Description

[0024] Figure 1 This is a schematic flowchart of the valve detection and control method based on the front and back ends of the present invention. Detailed Implementation

[0025] 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.

[0026] Example 1: Figure 1 This invention presents a valve detection and control method based on both front-end and back-end systems, comprising:

[0027] S1: After the front-end terminal is powered on, it performs a self-test and zero-position calibration. Within the small stroke of the valve, it drives the valve stem to reciprocate micro-movements according to a preset test sequence. Under a unified time reference, it synchronously collects the valve stem angular displacement and the actuator driving amount and stores them in time order to form small stroke test data.

[0028] S2: The front-end terminal preprocesses the small stroke trial data according to a uniform time step, calculates the driving amount and displacement increment of adjacent sampling points, and performs change point detection and segment fitting to obtain multiple stroke segments on the valve stem stroke. For each stroke segment, it calculates the segment-level characteristic parameters reflecting friction and hysteresis, generates a set of segment-level characteristic parameters, and uploads it to the back-end platform.

[0029] S3: The backend platform receives the set of segment-level characteristic parameters and organizes them in the stroke segment dimension in accordance with the process constraint data of the circuit to which the valve belongs. Based on the simplified working condition model, it calculates the friction characteristics and flow capacity of each stroke segment, obtains the upper limit of safe opening and the maximum allowable action speed of each stroke segment, forms a set of health constraint parameters and sends them to the front-end terminal.

[0030] S4: When the front-end terminal receives the target opening instruction, it converts the target opening into the target angle and locates the target angle on the stroke segment. Based on the health constraint parameter set, it corrects the target angle and action speed to generate the angle trajectory. During the execution process, it monitors the angle deviation by sampling. When the angle deviation continues to exceed the set standard, it performs backtracking control based on the last safe position and reports the abnormal event to the back-end platform.

[0031] In scenarios where existing valves operate under long-term pressure, the primary task is to acquire raw data reflecting the valve stem's motion state without causing significant operational disturbances. Step S1 revolves around the front-end terminal performing static detection, establishing a zero-position angle reference, generating small-stroke trial targets, and synchronous sampling and anomaly filtering on the exterior of the existing valve. The goal is to output a set of small-stroke trial data based on a unified time and angle reference. This data set will serve as the sole input source for identifying the relationship between drive increments and angle increments in subsequent steps. Therefore, in this step, each data object needs to be assigned a unique name and its usage fixed.

[0032] S101. Self-test and static data acquisition.

[0033] After the front-end terminal is powered on, the valve stem is kept in a fixed mechanical position. Within a preset static detection time interval, the output of the angle sensor is read at a fixed sampling period, forming a set of static angle sampling sequences arranged in chronological order. The difference between the maximum and minimum angle sampling values ​​in the static angle sampling sequence is defined as the static angle fluctuation value. The static angle fluctuation value is used to measure the stability of the valve stem's angle sampling under static conditions.

[0034] Within the same static detection time interval, the output of the drive acquisition unit is read at the same sampling period to form a static drive sampling sequence. The difference between the maximum and minimum drive sampling values ​​in the static drive sampling sequence is defined as the drive static fluctuation value. The angle static fluctuation value and the drive static fluctuation value are compared with preset angle stability thresholds and drive stability thresholds, respectively.

[0035] When the static fluctuation value of the angle does not exceed the angle stability threshold and the static fluctuation value of the drive does not exceed the drive stability threshold, the static detection is deemed to have passed; otherwise, the sensor or actuator is deemed to be in an abnormal state and subsequent steps are stopped.

[0036] S102. Zero-position reference angle calibration.

[0037] After the static test is passed, a reciprocating rotation command within a predefined angle range is sent to the actuator, causing the valve stem to rotate once in the forward direction and once in the reverse direction around the current operating position. During the entire reciprocating rotation process, the angle sensor output is continuously read according to the aforementioned sampling period to form a dynamic angle sampling sequence.

[0038] The minimum and maximum angle sampling values ​​appearing in the dynamic angle sampling sequence represent the two angle limits reached in this reciprocating motion. The arithmetic mean of these two angle limits is defined as the zero-point reference angle value. The zero-point reference angle value serves as the sole reference for valve stem angles throughout the invention, and all subsequent angle-related data are expressed as deviations relative to the zero-point reference angle value. To ensure consistency among different valves, the angle range of the reciprocating rotation is set during the installation and commissioning phase based on the maximum permissible trial amplitude of the valve under low-load conditions. Typically, pressure and flow fluctuations are observed by gradually increasing the reciprocating command amplitude, and the maximum amplitude that will not cause significant changes in operating conditions is selected as the configuration value without triggering an alarm.

[0039] S103. Small-scale target exploration generation.

[0040] After establishing the zero-base angle value, the current angle sensor output is read, and the zero-base angle value is subtracted from the sampled angle value to obtain the current relative angle value. An internally stored list of dimensionless small-stroke increments is used, where each increment unit represents the expected angular offset relative to the current relative angle value. The increment list monotonically increases in value. For each increment unit, two types of target relative angle values ​​are constructed: one type is the current relative angle value plus the increment unit, corresponding to the forward test target angle value; the other type is the current relative angle value minus the increment unit, corresponding to the reverse test target angle value. Each target relative angle value is added to the zero-base angle value to obtain the forward and reverse test target absolute angle values. Each pair of forward and reverse test target absolute angle values ​​is assigned a unique test number, which is arranged in ascending order according to the increment list, forming a complete queue of test target angles. To ensure that the trial actions do not exceed the valve's allowable small stroke range, the incremental list was determined through multiple tests during the configuration phase. Specifically, the relative angle offset was increased step by step. Under the premise that no alarm was triggered when monitoring changes in supply pressure and flow, a maximum offset value was selected. Then, the range from zero to the maximum offset value was divided into several levels to form the small stroke incremental list.

[0041] S104. Small-scale trial execution and synchronous sampling.

[0042] Following the order of the target angle queue, the testing actions are executed sequentially, starting with the entry with the smallest test number. For each test number, a motion trajectory consisting of angular velocities is planned. This trajectory is divided into three phases in time: an acceleration phase with gradually increasing angular velocity, a constant velocity phase with fixed angular velocity, and a deceleration phase with gradually decreasing angular velocity. The duration of each phase and the rate of change of angular velocity are configured by control engineers during installation and commissioning based on the inertia of the actuator and the mechanical characteristics of the valve. The actuator drives the valve stem along this motion trajectory from the initial angular position to the corresponding absolute angle value of the test target and then back to the initial position.

[0043] Throughout the entire process, the angle sensor output and the drive acquisition unit output are read synchronously at a fixed sampling period. The angle sampling value, drive sampling value and the absolute time of the sampling are combined into a data unit and stored in the data sequence corresponding to the current test number in an ascending order of sampling sequence number.

[0044] After all the actions of the trial numbers are executed, multiple data sequences are formed in the internal storage area of ​​the front-end terminal. Each data sequence corresponds to a trial number and contains angle sampling values, driving sampling values ​​and sampling time arranged in chronological order.

[0045] S105. Incremental calculation and anomaly screening.

[0046] To provide a discrete correspondence between driving changes and angle changes for subsequent steps, within the data sequence of each trial number, two adjacent data units are taken sequentially according to the sampling sequence number. The angle sampling value of the next data unit is subtracted from the angle sampling value of the previous data unit to obtain a series of angle increment values, forming an angle increment sequence. At the same time, the driving sampling value of the next data unit is subtracted from the driving sampling value of the previous data unit to obtain a series of driving increment values, forming a driving increment sequence.

[0047] For each angle increment sequence, the sum of the absolute values ​​of all angle increments is calculated to obtain the total angle change corresponding to that test number. Similarly, for each drive increment sequence, the sum of the absolute values ​​of all drive increments is calculated to obtain the total drive change corresponding to that test number. The total angle change is compared with a preset minimum angle change threshold. If the total angle change is less than the minimum angle change threshold, the test is considered to have not produced a valid angular displacement, and this test is deemed an idling test. The total drive change is compared with a preset maximum drive change threshold. If the total drive change is greater than the maximum drive change threshold, the actuator is considered to be approaching saturation operation in this test, and this test data is deemed unsuitable for characteristic modeling.

[0048] The above screening logic is applied to all trial numbers one by one. Data sequences that are determined to be idle trials or near-saturation trials are marked as abnormal trials and removed from the modeling dataset. Only the data sequences corresponding to the trial numbers that pass the screening are retained.

[0049] After step S1 is completed, a qualified test data set is formed in the front-end terminal. This set uses the test number as the outer index and the sampling sequence number as the inner index, and uniformly adopts the zero-position reference angle value as the angle reference. It includes static detection results, target angle information of small-stroke test, angle sample value, drive sample value, sampling time, and the total angle change and total drive change calculated from these sample values, and clearly marks abnormal tests. This data set is directly used in the subsequent step S2 to establish the segmented correspondence between drive increment and angle increment, thereby completing the segmentation of existing valve stroke and identification of high-risk areas under the condition of unknown original design parameters, providing a solid data foundation for subsequent health constraint generation.

[0050] In one embodiment, after long-term operation, the control valve of a chemical plant exhibited a slow response to the opening command. Maintenance personnel, without interrupting production, fixed a front-end terminal to a clamping mechanism near the valve stem and connected the angle sensor and actuator drive quantity acquisition line to the front-end terminal. After powering on, the front-end terminal completed a self-test and then drove the actuator to perform a small reciprocating rotation. The clamping mechanism formed a clear reference point at the mechanical limit position, and the front-end terminal used this to complete zero-position calibration and align the angle sampling and drive quantity sampling to the same time reference. After confirming on-site that the valve was near the process-allowed opening degree, the front-end terminal drove the valve stem to reciprocate slightly around the current operating position according to a small-stroke trial sequence. The front-end terminal simultaneously collected the valve stem angular displacement and actuator drive quantity, writing them in pairs to the local buffer in chronological order. It also recorded the trial number, direction of action, and start and end times for each trial action. After the trial, the front-end terminal locally marked segments with near-static angular displacement changes as invalid segments and excluded them from further analysis. On-site, the front-end terminal generated a set of trial data files arranged by trial number and entered the analysis pending state.

[0051] In step S1, a qualified set of test data has been obtained. This set records the angle sampling value, drive sampling value, and sampling time for each small-stroke test, categorized by test number and sampling sequence number, and excludes idling and saturation tests. For existing valves, to extract stroke segment characteristics usable by the backend platform from these discrete time series, it is necessary to unify the raw data from different tests and time steps onto a common reference frame and identify stroke segments with clear physical meaning in the angular coordinates. Step S2 revolves around this goal, transforming the test data into segment-level characteristic parameters indexed by angular position through time resampling, incremental construction, gradient analysis, single-test stroke segment statistics, and cross-test aggregation. This provides the foundation for operating condition modeling and health constraint generation in step S3.

[0052] S201. Experiment with data inheritance and standard time resampling.

[0053] At the start of processing, data is read from the qualified trial data set formed in step S1 in order of trial number. For any trial number, the data contains an angle sample value sequence, a drive sample value sequence, and a sampling time sequence arranged by sampling sequence number. Due to the influence of actuator response and communication delay during the actual sampling process, the sampling time intervals between adjacent samples within the same trial may differ, and the sampling time intervals between different trials are also inconsistent. Therefore, it is necessary to establish a standard time sequence within each trial number.

[0054] The specific procedure is as follows: First, extract the first and last sampling times of the current trial. Divide this time interval into several consecutive time points according to a pre-set standard time step, forming a standard time series. The standard time step can be determined during the installation and commissioning phase based on the maximum allowable angular velocity of the actuator and the desired time resolution. For example, select a step size that ensures at least ten standard time points are included in a single trial. For each time point in the standard time series, find the two nearest sampling time points in the original sampling time series. Assuming that the angle sampling value and the drive sampling value change at a uniform rate between these two sampling time points, calculate the resampled angle value and resampled drive value corresponding to that standard time point using linear interpolation. The meaning of interpolation calculation is that, for a standard time point that lies between two actual sampling moments, its angle value is equal to the angle value at the previous moment plus the product of the angle change rate and the time offset; the drive value is calculated similarly. Through this resampling process, each trial number obtains a set of resampled angle value sequences and resampled drive value sequences defined on the standard time series, thereby establishing a unified time reference between different trials.

[0055] S202. Incremental construction and gradient ratio calculation.

[0056] After standard-time resampling, angle increments and driving increments are constructed for the resampled angle value sequence and resampled driving value sequence within each trial number. For any two adjacent standard time points within a trial number, the angle increment is defined as the resampled angle value of the later standard time point minus the resampled angle value of the earlier standard time point, and the driving increment is defined as the resampled driving value of the later standard time point minus the resampled driving value of the earlier standard time point. The resulting sequence of angle increments and driving increments reflects the angle change and driving change within each time step during a trial.

[0057] To describe the response strength of driving changes to angle changes from a local perspective, the gradient ratio is calculated for each time step, provided the absolute value of the angle increment is higher than the angle increment filtering threshold. The corresponding driving increment value is then divided by the angle increment value to obtain the driving change per unit angle change. For time steps where the absolute value of the angle increment is not higher than the angle increment filtering threshold, the gradient ratio is not calculated; instead, these time steps are marked as angle stagnation points in the data record, indicating that the angle has not changed significantly within this time period.

[0058] For example, the gradient ratio can be calculated as follows:

[0059] After the unified time-based re-sample collection is completed, the control unit will handle any probe number. Two adjacent time indices inside and The angle increment is defined as the subsequent resampled angle value minus the previous resampled angle value, and the drive increment is defined as the subsequent resampled drive value minus the previous resampled drive value. The expression is as follows:

[0060] ;

[0061] ;

[0062] In the formula, the angle increment Indicates the number of the trial. In time index The angle change between the current time index and the next time index drives the increment. This represents the driving change within the same time interval. To reflect the response strength of the driving change to the angle change from a local perspective, the control unit divides the driving increment by the angle increment to obtain the gradient ratio, provided that the absolute value of the angle increment is greater than the angle increment filtering threshold. The expression is:

[0063] ;

[0064] gradient ratio in the formula Indicates the number of trials Time Index Nearby, the driving change corresponding to a unit angle change. The angle increment filtering threshold is used to distinguish between sampling moments where the angle is basically stationary and sampling moments where there is actual motion. During the on-site debugging phase, a stable trial data point can be selected, and the absolute values ​​of all angle increments can be statistically analyzed. The representative value near the lower bound of the statistical distribution is selected and multiplied by a proportionality coefficient less than one.

[0065] In one embodiment, maintenance personnel install a front-end terminal at the valve site and complete the small-stroke trial in step S1. During the trial, the front-end terminal obtains the sampling time sequence, angle sampling value sequence, and drive sampling value sequence corresponding to the same trial number. The front-end terminal first generates a standard time sequence according to the standard time step, and then generates a resampled angle value sequence and a resampled drive value sequence through linear interpolation at each standard time point. Subsequently, the front-end terminal calculates the angle increment value and drive increment value pairwise according to adjacent standard time points, and compares the angle increment value with the angle increment filtering threshold. The standard time step that meets the filtering condition is written into the gradient ratio value, and the standard time step that does not meet the filtering condition is written into the angle stagnation point mark and the stagnation drive increment table is recorded simultaneously. After the conversion is completed, the front-end terminal obtains the gradient ratio sequence and the set of angle stagnation point indices that correspond one-to-one with the original trial number, which serve as the direct input for the gradient change calculation in step S203.

[0066] S203. Gradient change calculation and single-trial candidate boundary identification.

[0067] To identify the abrupt changes in the driving angle response characteristics within each trial record, the gradient change is further calculated on the obtained gradient ratio sequence. The gradient change is defined as the absolute value of the difference between the gradient ratios of two adjacent time steps, and its physical meaning is the magnitude of the change in the driving angle response intensity per unit time step.

[0068] For each trial number, starting from the first defined gradient ratio, calculate the absolute value of the difference between adjacent gradient ratios one by one, compare these gradient changes with the gradient change judgment threshold, and collect the time indexes of gradient changes that are higher than the judgment threshold into the candidate boundary index set.

[0069] Since the candidate boundary index set often contains a series of closely spaced time indices, to reduce boundary redundancy in angular coordinates, the candidate boundary indices are clustered and compared. When the difference between two candidate indices is not higher than a preset index spacing threshold, the two indices are grouped into the same change zone. Within the change zone, the time index with the highest gradient change value is selected as the representative boundary index. In each trial record, the representative boundary index, together with the start time index and the end time index, constitute the time boundary set for a single trial, and the corresponding resampled angle value forms the angle boundary set for a single trial.

[0070] S204. Segment division of a single trial journey and calculation of segment-level characteristics.

[0071] After obtaining the set of single-trial angle boundaries, the entire time index between any two adjacent time boundaries is divided into a single-trial stroke segment according to chronological order. Each single-trial stroke segment has a corresponding starting angle position and ending angle position, as well as a sequence of angle increment values, a sequence of driving increment values, and a sequence of gradient ratios covering that segment. Within each single-trial stroke segment, the average gradient value is first calculated. All defined gradient ratios within the segment are summed, and then divided by the number of gradient ratios within the segment to obtain the average gradient value representing the overall stiffness level of that segment. Subsequently, the absolute maximum value of the driving increment value within the segment is found and used as the peak driving increment, reflecting the peak characteristics of the driving change within that segment. Simultaneously, the net angle change is calculated by summing all angle increment values ​​within the segment to obtain the total angle change from the start to the end of the segment.

[0072] The ratio of time indices marked as angle stagnation points within a segment to the total number of time indices within that segment can be calculated to obtain the angle stagnation percentage, which reflects the degree to which the angle remains stationary for an extended period within the segment. Based on the combination of average gradient value, driving peak increment, net angle change, and angle stagnation percentage, a type determination is given for each single trial stroke segment. When the net angle change is not higher than a pre-set dead zone angle threshold and the angle stagnation percentage is high, the segment is marked as a dead zone segment; when the net angle change is higher than the dead zone angle threshold, and the average gradient value exceeds the upper limit of the normal gradient reference range, and the driving peak increment is higher than the static friction determination threshold, the segment is marked as a high friction segment; other segments are marked as normal segments.

[0073] Each single trial journey segment ultimately forms a single trial segment-level characteristic record containing the starting angle position, ending angle position, average gradient value, driving peak increment, net angle change, angle stagnation percentage, and segment type label, and is stored in the intermediate characteristic data set using the trial number and segment sequence number as indexes.

[0074] S205. Aggregation across trial angle boundaries and generation of global travel segmentation characteristics.

[0075] Since the starting angle and amplitude of each probe may be different, the segments of a single probe journey formed within different probes overlap and intersect on the angular coordinates.

[0076] To establish a unified stroke segmentation across the entire valve stroke, the starting and ending angle positions of each segment are extracted from all single-trial segment-level characteristic records. These two types of angle values ​​are then merged into a raw boundary angle set. After sorting the raw boundary angle set in ascending order of angle value, the difference between two adjacent boundary angles is compared with a global angle merging threshold. For a group of consecutive boundary angles whose difference is not higher than the global angle merging threshold, they are grouped into the same angle cluster, and the arithmetic mean of all angle values ​​within this cluster is calculated. This average is used as a global stroke boundary angle. By performing this process on the entire raw boundary angle sequence, several global stroke boundary angles are formed, arranged in ascending order of angle value. The angle interval between any two adjacent global stroke boundary angles is defined as a global stroke segment.

[0077] After dividing the global travel into segments, for each global travel segment, all single-trial segment-level characteristic records are traversed, and records whose starting and ending angle positions fall entirely within the angle range of the current global travel segment are collected, forming a segmented statistical set. For each record in the statistical set, its average gradient value, driving peak increment, and net angle change are read, and the arithmetic mean is calculated for each type of characteristic to obtain the global average gradient, global average driving peak, and global average net angle change for this global travel segment. The segment type labels in the records can also be counted to determine the proportion of dead segments, high-friction segments, and normal segments, used to verify the consistency between global indicators and local types.

[0078] Then, based on the relationship between the global average gradient, global average driving peak, global average net angle change, and the aforementioned thresholds, each global travel segment is assigned a final travel segment type label. A global travel segment is identified as a dead zone segment when the global average net angle change is within the dead zone angle threshold range and the proportion of dead zone segments in the statistical set is the highest. A global travel segment is identified as a high-friction segment when both the global average gradient and global average driving peak are above the upper limit of the normal reference range and the proportion of high-friction segments is the highest. Other cases are identified as normal segments. The final segment-level characteristic parameter set is indexed by the global travel segment number, recording the start angle position, end angle position, global average gradient, global average driving peak, global average net angle change, and travel segment type label for each segment. This data is stored in front-end non-volatile storage and simultaneously sent to the back-end platform via a communication link, providing direct input for the subsequent step three: construction of the operating condition model and calculation of health constraint parameters.

[0079] After processing in step S2, the short-stroke trial timing data from step S1 is completely transformed into a stroke segment characteristic description with physical meaning on the angular coordinate system. Each global stroke segment corresponds one-to-one with a specific angular range of the valve stem, incorporating both statistically averaged gradients and drive peak information from multiple trials, as well as segment type labels obtained through threshold determination. In this way, the backend platform can directly superimpose pressure and flow constraints on specific stroke segments in step S3, generating safe opening ranges and action speed limits that conform to existing operating conditions. This ensures that health constraints are no longer limited to the overall valve level but are implemented at the fine angular range level.

[0080] In one embodiment, after the front-end terminal completes step S1, maintenance personnel can initiate stroke segment identification directly through the front-end terminal's maintenance interface without disassembling the front-end terminal on-site. The front-end terminal first realigns the angular displacement sequence and driving quantity sequence corresponding to each trial number with a unified time step, forming a directly comparable time series. Then, it performs noise reduction processing on the series to prevent abrupt changes caused by sensor jitter from participating in segmentation judgment. The front-end terminal then calculates the driving quantity increment and displacement increment at adjacent moments, forming an increment sequence arranged by time. The front-end terminal performs change point detection on the increment sequence; the change point position corresponds to the moment when the valve stem response law changes direction. Then, it performs segmented fitting on each data segment using the change point as the boundary, obtaining multiple stroke segments. The front-end terminal writes the starting angle position, ending angle position, average slope, static friction peak value, hysteresis width, and stroke segment type label of each stroke segment into a segment-level characteristic parameter set, forming a stroke segment division table indexed by the stroke segment number. Maintenance personnel can see on the interface that the list of trip segment quantity and segment boundary angle has been generated. Subsequently, the front-end terminal packages the set of segment-level characteristic parameters and puts them into the upload queue.

[0081] In the preceding steps, the front-end has already formed a set of segment-level characteristic parameters through small-stroke probing and segment identification, and uploaded it to the back-end platform. The set of segment-level characteristic parameters is indexed by the stroke segment number, recording the starting angle position, ending angle position, global average gradient, global average driving peak value, global average net angle change, and stroke segment type label for each segment. For existing valves in complex process loops, relying solely on these mechanical characteristic parameters is insufficient to directly determine which stroke segments are safe to use under the current upstream pressure, downstream pressure, allowable differential pressure, and target flow range. The back-end platform needs to introduce process constraint data, mapping it one-to-one with the segment-level characteristic parameters along the stroke segment number dimension, unifying mechanical characteristics and operating condition constraints into a single data structure. Based on this, the theoretical flow capacity, health indicators, safe opening limit, and maximum allowable operating speed of the stroke segments are calculated, forming a set of health constraint parameters that the front-end can execute.

[0082] S301. Establishment and organization of the correspondence between segment-level characteristic parameters and process constraint data.

[0083] The backend data processing unit first reads the aforementioned set of segment-level characteristic parameters. For each stroke segment number, the data processing unit extracts the starting angle position, ending angle position, global average gradient, global average drive peak value, global average net angle change, and stroke segment type label for that segment. Subsequently, the data processing unit reads the process constraint data corresponding to the loop to which the valve belongs from the process monitoring database according to the valve's unique identifier. The process constraint data includes the minimum allowable upstream pressure, the maximum allowable upstream pressure, the minimum allowable downstream pressure, the maximum allowable downstream pressure, the upper limit of allowable differential pressure, the minimum allowable target flow rate, the maximum allowable target flow rate, and the medium density. The data processing unit constructs a joint data table with the stroke segment number as the primary key, writing the segment-level characteristic fields of each stroke segment and the aforementioned process constraint fields into the same row of records, ensuring that the stroke segment number and the corresponding mechanical and operating parameters are structurally completely correlated. If the pressure unit, flow unit, or angle unit in the process constraint data is found to be inconsistent with the unit used in the segment-level characteristic parameters, the data processing unit will convert all pressure values ​​to a single pressure unit, all flow values ​​to a single flow unit, and all angle values ​​to a single angle unit according to a pre-set unit conversion table, and then write them into the joint data table.

[0084] By establishing a joint data table according to the segment numbering of the itinerary, the set of segment-level characteristic parameters and process constraint data are matched and organized segment by segment at the data level.

[0085] S302. Calculation logic of effective pressure difference and theoretical flow capacity for stroke segments.

[0086] After the joint data table is established, the modeling unit uses the table's contents to calculate the effective differential pressure and theoretical flow capacity for each stroke segment. For any stroke segment number, the modeling unit first reads the maximum allowable upstream pressure and the minimum allowable downstream pressure from the joint data table, and calculates the difference between them as the upper limit of the theoretical differential pressure. Then, it compares the upper limit of the theoretical differential pressure with the upper limit of the allowable differential pressure, and takes the smaller of the two as the effective differential pressure available for the stroke segment under the current operating conditions. Subsequently, the global average gradient needs to be converted into the equivalent flow coefficient. To this end, during the offline test phase, several reference valves with specifications similar to the target valve are selected, and standardized flow tests are conducted on these reference valves according to fixed stroke segment divisions. Stable flow and differential pressure data are collected at different segment angle positions. The flow coefficient of each reference stroke segment is calculated using the traditional flow coefficient definition, and the corresponding global average gradient is recorded. Finally, a set of discrete correspondence points between the global average gradient and the flow coefficient is formed. The modeling unit performs interpolation fitting on this set of correspondence points to obtain the functional relationship from the global average gradient to the equivalent flow coefficient, and stores the function parameters in the configuration file.

[0087] During online operation, for each stroke segment in the joint data table, the modeling unit reads the global average gradient, retrieves the aforementioned functional relationship, and calculates the equivalent flow coefficient through interpolation. Then, according to the conventional calculation method for orifice flow, the equivalent flow coefficient of the stroke segment is multiplied by the square root of the ratio of effective pressure difference to medium density to obtain the theoretical maximum flow capacity of the stroke segment per unit time. By adding a theoretical maximum flow capacity field to the joint data table, the above steps are performed sequentially for all stroke segments to establish the correspondence between stroke segments and theoretical flow capacities.

[0088] For example, the global average gradient can be converted into an equivalent flow coefficient by the following method:

[0089] Let the discrete values ​​of the global average gradient be sorted in ascending order as follows:

[0090] ;

[0091] The corresponding flow coefficient calibration values ​​are sorted as follows:

[0092] ;

[0093] Then when the global average gradient takes the value Falling into the range At that time, the equivalent flow coefficient Pick

[0094] ;

[0095] when Less than When, the equivalent flow coefficient is taken ;when Greater than When, the equivalent flow coefficient is taken .

[0096] Sort the breakpoint values ​​of the global average gradient in ascending order to get the total number of breakpoints. The sequence of breakpoints; when the global average gradient value is located at the th The breakpoint and the first When between the breakpoints, the first breakpoint is used. The equivalent flow coefficient is calculated using linear interpolation of the segment, where The value range is from 2 to -1 。

[0097] In one embodiment, the test personnel set up a closed-loop pipeline and installed a reference valve at the test station. The front-end terminal was clamped at the valve stem position of the reference valve. Following steps S1 and S2, the front-end terminal completed small-stroke probing and stroke segment identification, obtaining the global average gradient corresponding to each stroke segment number. The test personnel then drove the valve stem to the midpoint angle position of a certain stroke segment and held it stationary. By adjusting the return branch, the flowmeter reading reached a stable state. The flowmeter reading and upstream and downstream pressure measurement point readings were recorded within the stable sampling window. The pressure difference was calculated, and the flow coefficient calibration value for that stroke segment was obtained according to the definition of the flow coefficient. The test personnel repeated the above holding and recording process for each stroke segment, forming a discrete set of corresponding points between the global average gradient and the flow coefficient calibration value. The back-end platform sorted the discrete corresponding points according to the global average gradient and generated a breakpoint array, wrote it to the configuration file, and calculated the equivalent flow coefficient through piecewise linear interpolation during online runtime.

[0098] S303. Construction of health indicators for trip segments and logic of risk level classification.

[0099] To reflect the combined friction state and drive load at the stroke segment level, the modeling unit constructs stroke segment health indicators based on the global average gradient and global average drive peak value. Before formal operation, maintenance personnel collected two types of segment-level characteristic samples from several valves: healthy valve samples with good structure and lubrication, and faulty valve samples with obvious jamming, wear, or leakage. For the healthy valve samples, the global average gradient of all segments was calculated, and its concentrated distribution location was determined as the gradient reference value; the global average drive peak value of all segments was calculated, and its concentrated distribution location was determined as the drive reference value.

[0100] The centralized distribution location is defined as follows: A sample set is constructed using the global average gradient of all stroke segments in the healthy valve sample. The sample set is sorted from smallest to largest value, and samples at both ends of the sorted sequence with a predetermined proportion are removed, retaining the middle sample subset. The arithmetic mean of the middle sample subset is taken to obtain the gradient reference value. Another sample set is constructed using the global average driving peak value of all stroke segments in the healthy valve sample. The same sorting, removal, and arithmetic averaging process is used to obtain the driving reference value. The predetermined proportions are given by the maintenance procedure, and the setting principle is to remove a small number of abnormal segments with significant deviations, while retaining a sufficient number of normal segments for calculating the reference value.

[0101] In one embodiment, maintenance personnel select several valves with good structure and lubrication conditions within the same factory area as healthy valve samples. They install front-end terminals on each valve and execute steps S1 and S2, uploading the segment-level characteristic parameter set for each valve to the back-end platform. The back-end platform extracts the global average gradient from the segment-level characteristic parameter set of each healthy valve sample and merges them into a gradient sample set. The gradient sample set is sorted, and samples at both ends with a preset ratio are removed to obtain an intermediate sample subset. The arithmetic mean of this subset is calculated and written into the gradient reference value field. The back-end platform processes the global average driving peak value in the same way to obtain the driving reference value field. During online operation, the back-end platform reads the global average gradient and global average driving peak value for any stroke segment and calculates the ratios with the gradient reference value and driving reference value, respectively, as inputs for subsequent health indicator construction.

[0102] During online operation, for any stroke segment number, the modeling unit reads the global average gradient and global average drive peak value from the joint data table, divides them by the gradient reference value and drive reference value respectively, to obtain the average gradient ratio and drive peak value ratio. Then, the average gradient ratio and drive peak value ratio are multiplied to obtain a comprehensive resistance factor. Finally, the health index value is obtained by dividing by one and adding the comprehensive resistance factor. The health index naturally falls between zero and one. The closer the value is to one, the closer the friction state and drive load of that segment are to the healthy valve reference level; the closer the value is to zero, the closer the segment is to a severely hindered state.

[0103] To transform health indicators into discrete risk levels, the modeling unit determines two health indicator thresholds using the health indicator distributions of the aforementioned healthy valve samples and faulty valve samples. Specifically, the segmented health indicators of all healthy samples are sorted, and the values ​​at the lower and upper predetermined proportions of the sorted sequence are selected as the boundaries of the healthy sample intervals. Similarly, the segmented health indicators of all faulty samples are sorted, and the values ​​at the upper predetermined proportions of the sorted sequence are selected as the boundaries of the faulty sample intervals. Then, by comparing the interval boundaries of the two types of samples, low-risk, medium-risk, and high-risk zones on the health indicator axis are determined.

[0104] During online operation, for each travel segment number, the modeling unit divides the risk level based on the interval where the health indicator is located and combines it with the travel segment type label. When the health indicator falls within the low-risk area and the travel segment type label is a normal segment, the risk level is recorded as low risk; when the health indicator falls within the medium-risk area or the travel segment type label is a normal segment but the health indicator is close to the high-risk boundary, the risk level is recorded as medium risk; when the health indicator falls within the high-risk area or the travel segment type label is a dead segment or a high-friction segment, the risk level is recorded as high risk.

[0105] The risk level field is written back to the combined data table so that each travel segment has both a health indicator and a risk level.

[0106] S304. Solving the upper limit of safety opening and the maximum allowable operating speed and generating the health constraint parameter set.

[0107] After the theoretical maximum flow capacity, health indicator, and risk level of each travel segment have been written into the combined data table, the modeling unit uses this information to generate the upper limit of safety opening and the maximum allowable operating speed for the front end. First, calculate the geometric opening ratio of the travel segment by dividing the angle range of the travel segment by the total angle range of the effective stroke of the valve, obtaining the geometric opening ratio of this segment. Subsequently, calculate the segment flow utilization ratio by dividing the maximum allowable value of the target flow by the theoretical maximum flow capacity of the travel segment, obtaining the relative capacity required when only considering the flow demand. When the segment flow utilization ratio is lower than one, theoretically, the target flow can be satisfied without using the full segment; when the segment flow utilization ratio is higher than one, it indicates that using this segment alone cannot bear the target flow and multiple segments need to work together.

[0108] The modeling unit constructs an opening utilization coefficient by comparing the segment flow utilization ratio with the health indicator and taking the smaller value of the two as the opening utilization coefficient, so that segments with lower health indicators are only allowed to participate in work with a limited opening even if their flow capacity is sufficient.

[0109] The angle position of the upper limit of safety opening is determined as follows. Based on the starting angle position of the travel segment, calculate the product of the angle range of the travel segment and the opening utilization coefficient, and add the product value to the starting angle position to obtain the angle position of the upper limit of safety opening. For travel segments with a high-risk level, the modeling unit can add a reduction coefficient to the above opening utilization coefficient and preset the reduction ratio in the configuration file to impose stricter opening restrictions on high-risk segments in health constraints.

[0110] The logic for determining the maximum permissible motion speed is as follows: first, during the actuator selection phase, the reference motion speed is calculated from the standard full-stroke motion time provided by the manufacturer, and then the entire effective angle range is divided by the standard full-stroke time to obtain the reference motion speed.

[0111] During online operation, the modeling unit multiplies the reference action speed with the health index of each stroke segment to obtain the maximum allowable action speed. This allows segments with a health index close to one to operate at the reference action speed, while segments with a lower health index have their action speed limited to a lower level. This enables the valve to move slowly within stroke segments with poor mechanical conditions, reducing the risk of impact and jamming.

[0112] The modeling unit organizes the starting and ending angle positions, upper limit angle positions of safe opening, maximum permissible action speed, risk level, and health indicators of each stroke segment into health constraint records. These records are then written sequentially into the health constraint parameter set, indexed by the stroke segment number. After the health constraint parameter set is generated, the backend platform performs a consistency check. This check includes verifying whether the upper limit angle positions of the safe opening of adjacent stroke segments form a continuous, monotonic safe interval; whether the upper limit of the safe opening of high-risk segments has shrunk to the preset safe range; and whether all stroke segments have generated maximum permissible action speeds and risk levels. Once the checks pass, the health constraint parameter set is packaged together with the valve's unique identifier and sent to the corresponding frontend terminal via the communication link. A copy is also stored in the backend storage medium for future reference in operational condition traceability and parameter adjustment.

[0113] Through the processing in step S3 above, the set of segment-level characteristic parameters and process constraint data are completely transformed into a set of health constraint parameters for control execution at the stroke segment level. Each stroke segment has a safe opening limit and a maximum permissible action speed that match the current operating conditions, along with health indicators and risk level descriptions. In step S4, the front-end terminal can directly project and correct the target opening and action time in the process control command based on the set of health constraint parameters, ensuring that the actual action trajectory of the existing valves is consistently limited to the stroke range that is permissible under operating conditions and where the mechanical state is relatively healthy, thereby reducing the risk of jamming and malfunction shutdown.

[0114] In one embodiment, after the backend platform receives the set of segment-level characteristic parameters uploaded from the field front-end terminal, the on-duty engineer can view the newly added record of the corresponding valve's unique identifier on the maintenance terminal in the control room. The backend platform locates the loop to which the valve belongs in the asset ledger and reads the upstream pressure allowable range, downstream pressure allowable range, allowable differential pressure limit, target flow allowable range, and medium characteristics of the loop from the historical database of the process control device. The backend platform writes these process constraint data and the set of segment-level characteristic parameters into the same joint data structure according to the stroke segment number. The backend platform then calls the global average gradient and equivalent flow coefficient relationship stored in the offline calibration file, converts the global average gradient of each stroke segment into the equivalent flow coefficient within the joint data structure, and calculates the flow capacity of each stroke segment in combination with the process constraint data. The backend platform then uses the gradient reference value and drive reference value formed by the healthy valve sample as a benchmark to generate health indicators for each stroke segment and give a risk level. Subsequently, it calculates the safe opening upper limit angle position and maximum allowable action speed of each stroke segment to form a health constraint parameter set. After the on-duty engineer confirms that the health constraint parameter set matches the valve's unique identifier, the backend platform sends the health constraint parameter set to the corresponding frontend terminal. The frontend terminal maintenance interface displays that the health constraint parameter set has been updated and can be used for control command processing.

[0115] In steps S1 to S3, the front-end terminal has completed the small-stroke trial data collection and stroke segment characteristic identification. The back-end platform has also generated a health constraint parameter set based on the segment-level characteristic parameter set and process constraint data, and distributed it to the front-end terminal according to the valve's unique identifier. In the health constraint parameter set, the stroke segment number is used as an index to provide each stroke segment with the stroke segment's starting angle position, stroke segment's ending angle position, stroke segment's upper limit angle position for safe opening, stroke segment's maximum allowable action speed, and stroke segment's risk level.

[0116] In actual operation, existing valves still rely on the process control device to issue target opening percentages and expected action times. If the front-end terminal directly converts the target opening percentage into angle commands and drives the actuator, the valve stem may move at excessively high speeds in the high-friction stroke segment or exceed the upper limit angle of the safe opening for that stroke segment. Step S4 involves mapping the health constraint parameter set to the control commands from the process control device segment by stroke segment within the front-end terminal, projecting and correcting them in both the angle target and action speed dimensions, and monitoring deviations during operation. If a deviation from the health constraints occurs, a safety backoff action is triggered, and an abnormal event record is generated and uploaded to the back-end.

[0117] S401. Control command parsing and motion path travel segment set determination.

[0118] After the health constraint parameter set has been stored in the front-end terminal, whenever the process control device issues a new control command, the front-end first receives the target opening percentage and the expected action time. The current valve stem angle position is given by the angle sensor and is uniformly referred to as the current valve stem angle position. The fully closed valve angle position and the fully open valve angle position have been written into the front-end configuration during the commissioning phase.

[0119] The control unit first calculates the target angle position based on the target opening percentage and the full stroke angle range. The calculation process is as follows: first, calculate the full stroke angle range, that is, subtract the valve's fully closed angle position from the valve's fully open angle position; then, divide the target opening percentage by one hundred to obtain the target opening coefficient; then, multiply the full stroke angle range by the target opening coefficient to obtain the angle increment; finally, add the angle increment to the valve's fully closed angle position to obtain the target angle position.

[0120] The direction of the action is determined by the relationship between the current valve stem angle and the target angle. If the target angle is greater than the current valve stem angle, the action is open; if the target angle is less than the current valve stem angle, the action is closed. In the health constraint parameter set, each stroke segment number corresponds one-to-one with the starting and ending angle positions of the stroke segment. All stroke segment numbers are traversed sequentially, forming an angle interval between the current valve stem angle and the target angle. Then, it is determined whether the angle interval of each stroke segment intersects with this angle interval. Specifically, if the ending angle of a stroke segment is not less than the smaller of the current valve stem angle and the target angle, and the starting angle of that stroke segment is not greater than the larger of the current valve stem angle and the target angle, then the current action path is considered to pass through this stroke segment, and the corresponding stroke segment number is added to the action path stroke segment set.

[0121] Through this process, the front-end terminal obtains a set of motion path travel segments, which are used to constrain the angle target and motion speed throughout the entire motion path.

[0122] S402. Determination of the angular position and planned angular velocity of the safe target under health constraint projection.

[0123] After the set of motion path travel segments has been determined, the control unit needs to use the upper limit angle position of the safe opening of the travel segment and the maximum allowable motion speed of the travel segment recorded in the health constraint parameter set to constrain the target angle position and motion speed. In the health constraint parameter set, each travel segment number records the upper limit angle position of the safe opening of the travel segment.

[0124] First, the system checks if any stroke segment in the motion path travel segment set has a safe upper limit angle position that is between the current valve stem angle position and the target angle position and is close to the target angle position in the direction of motion. If the safe upper limit angle position of each stroke segment in the motion path travel segment set is not lower than the target angle position, then the safe target angle position is directly taken as the target angle position. If one or more stroke segments in the motion path travel segment set have a safe upper limit angle position that is lower than the target angle position, then the safe upper limit angle position that is closest to the current valve stem angle position and is located in the direction of motion is selected from these stroke segments as the safe target angle position for this action. This ensures that subsequent actions will not exceed the safe upper limit angle position of any stroke segment in the motion path.

[0125] After determining the target safety angle position, the control unit calculates the average commanded angular velocity based on the difference between the target safety angle position and the current valve stem angle position, as well as the expected action time. The calculation process is as follows: take the absolute value of the difference between the target safety angle position and the current valve stem angle position, divide it by the expected action time, and obtain the average commanded angular velocity. Simultaneously, the control unit reads the maximum permissible action speed of each stroke segment from the set of stroke segments along the action path, and selects the smallest value as the maximum path angular velocity. The rule for determining the planned angular velocity is: if the average commanded angular velocity is not higher than the maximum path angular velocity, then the planned angular velocity is equal to the average commanded angular velocity; if the average commanded angular velocity is higher than the maximum path angular velocity, then the planned angular velocity is equal to the maximum path angular velocity.

[0126] Through the above two steps, the target angular position is projected as a safe target angular position, and the action speed is limited to the planned angular velocity that does not exceed the maximum permissible action speed of any stroke segment on the action path.

[0127] S403. Logic for constructing the planning angle position sequence and monitoring the angle deviation.

[0128] Once the target angle position and planned angular velocity are determined, the control unit needs to generate a sequence of planned angle positions and monitor the angle deviation between the actual and planned angle positions in real time during execution. The front-end terminal is pre-configured with a sampling period, which can be set according to the angle sensor refresh rate and the dynamic response capability of the actuator.

[0129] First, the total execution time is calculated based on the angle difference between the target safety angle position and the current valve stem angle position, as well as the planned angular velocity. The calculation process is as follows: take the absolute value of the difference between the target safety angle position and the current valve stem angle position, divide it by the planned angular velocity, and obtain the total execution time. If the total execution time is less than the expected execution time, the total execution time is used as the execution time for this action; if the total execution time is greater than the expected execution time, the total execution time is still used as the execution time, but it can be marked in the backend statistics as not meeting the original time requirement. Then, the total execution time is divided by the sampling period and rounded up to obtain the number of samples.

[0130] When constructing the planned angle position sequence, the current valve stem angle position is used as the first planned angle position. If the action is opening, the planned angle position at each sampling moment is the product of the planned angular velocity and the sampling period added to the previous planned angle position, until the safe target angle position is reached or the last sampling moment is reached. If the action is closing, the planned angle position at each sampling moment is the product of the planned angular velocity and the sampling period subtracted from the previous planned angle position, similarly until the safe target angle position is reached or the last sampling moment is reached. This forms a planned angle position sequence distributed over the total action execution time.

[0131] In actual execution, the front-end terminal collects the angle sensor output at each sampling moment to form an actual angle position sequence. The control unit calculates the angle deviation at each sampling moment, which is defined as the difference between the actual angle position and the planned angle position at the corresponding moment.

[0132] During online operation, the control unit checks whether the absolute value of the angle deviation exceeds the angle deviation threshold at each sampling moment, and simultaneously counts the number of times the absolute value of the angle deviation exceeds the threshold within consecutive sampling moments. The consecutive deviation standard is also obtained during the commissioning phase by counting the maximum consecutive deviation number in the reference action of the healthy valve and multiplying it by a safety factor. When the number of deviations at consecutive sampling moments exceeds the consecutive deviation standard, the current action is considered to have deviated from the healthy constraint, and a safety rollback action needs to be initiated.

[0133] S404. Safety rollback actions and abnormal event recording and reporting under abnormal trigger conditions.

[0134] When the absolute value of the angle deviation at consecutive sampling times exceeds the angle deviation threshold more than the consecutive exceedance count standard, the control unit immediately stops executing according to the original planned angle position sequence and initiates a safety backoff action. To ensure a clear target angle position for backoff in abnormal situations, the front-end terminal records a historical safe angle position after each normal operation. The historical safe angle position is defined as the actual angle position at the end of an operation, provided that the absolute value of the angle deviation never exceeds the angle deviation threshold and the drive quantity does not saturate throughout the entire operation.

[0135] When an anomaly occurs, the control unit plans a one-way retraction trajectory, using the current actual angle position as the retraction starting point and the historical safe angle position as the retraction ending point. The retraction angular velocity of the trajectory is selected based on the principle of not exceeding the maximum angular velocity of the path in the set of stroke segments corresponding to the current action, and not exceeding the planned angular velocity of the current action. Specifically, the planned angular velocity is compared with the maximum angular velocity of the path, and the smaller value is taken as the retraction angular velocity. The retraction trajectory uses the same sampling period as the normal action, advancing the planned angle position in the opposite direction at the retraction angular velocity at each sampling moment until it approaches the vicinity of the historical safe angle position. During the retraction process, the front-end terminal continues to collect the actual angle position and drive quantity. Once it detects that the absolute value of the difference between the actual angle position and the historical safe angle position is lower than the preset tolerance, and the absolute value of the angle deviation drops back to within the angle deviation threshold, the retraction action is considered complete, and the current actual angle position is updated to the new historical safe angle position to provide a reference for the next abnormal retraction.

[0136] While completing the safe rollback, the front-end terminal constructs an exception event record.

[0137] An exception log should include at least the following:

[0138] First, the original target opening percentage and the target angle position calculated from it, as well as the safe target angle position used in this case;

[0139] Second, the segment number of each segment in the motion path segment set and the corresponding risk level of the segment;

[0140] Third, an explanation of the statistical basis used during the initialization of the motion planning angular velocity, maximum path angular velocity, angular deviation threshold, and consecutive out-of-tolerance number standard;

[0141] Fourth, the absolute value of the maximum angular deviation that occurs during the operation and the number of consecutive samplings that exceed the tolerance;

[0142] Fifth, the starting angle position, ending angle position, and duration of the safe retraction action;

[0143] Sixth, the time of the exception trigger and the time of the exception event record generation.

[0144] After abnormal events are recorded and stored locally on the front-end terminal, they are uploaded to the back-end platform via a communication link. The back-end classifies and aggregates the abnormal events in the asset management module using the valve's unique identifier, and then compares them with the health constraint parameter set and segment-level characteristic parameter set generated in step S3 to determine whether the health constraint parameters need to be recalculated or whether the valve needs to be scheduled for offline maintenance.

[0145] Through the specific implementation of step S4 above, the front-end terminal strictly incorporates the upper limit angle position of the safe opening angle of the stroke segment and the maximum allowable action speed of the stroke segment from the health constraint parameter set into the calculation of the target angle position and action speed during the execution of control commands. Angle and speed constraints are simultaneously applied to the action path spanning multiple stroke segments. At the execution level, the angle deviation between the actual angle position and the planned angle position is monitored through an angle deviation threshold and a consecutive deviation count standard constructed based on the statistical results of the health valve. Once the action deviates from the health constraints, a safety backoff is performed according to the historical safe angle position, and a closed-loop data link between the front-end and back-end is formed through abnormal event recording. Thus, the segment-level characteristic parameter set and the health constraint parameter set obtained in steps S1 to S3 are implemented into specific angle trajectory and speed control rules in step S4, ensuring that the existing valves truly remain within the safe stroke segments defined by health constraints during long-term operation.

[0146] In one embodiment, production personnel issue a new target opening command to the valve and set the desired action time on the process control device. After receiving the command, the front-end terminal first converts the target opening into a target angular position, and then uses the current valve stem angular position and the target angular position to determine the set of stroke segments involved in the action path. The front-end terminal reads the upper limit angular position of the safe opening and the maximum allowable action speed corresponding to the set of stroke segments in the health constraint parameter set, projects the target angular position as the safe target angular position, and selects the most stringent maximum allowable action speed on the action path as the planned angular velocity. Subsequently, it generates an angular trajectory and drives the actuator to move along the trajectory. During the action, the front-end terminal continuously collects the valve stem angular displacement and the actuator drive amount, calculates the angular deviation between the actual angular trajectory and the angular trajectory, and judges it according to the pre-established angular deviation threshold and the consecutive out-of-tolerance number threshold. If the angular deviation continuously meets the abnormal triggering condition, the front-end terminal stops the original angular trajectory and performs backtracking control according to the historical safe position. At the same time, it generates an abnormal event record, which includes the original target opening command, the safe target angular position, the set of stroke segments of the action path, the maximum angular deviation, the backtracking start angular position, and the backtracking end angular position. In the control room, maintenance personnel can see that the backend platform receives abnormal event records and archives the records to the maintenance page of the valve. The field front-end terminal displays that the rollback is complete and updates the new historical safe position.

[0147] Specifically, the above are merely preferred embodiments of this application and are not intended to limit this application.

[0148] The angle stability threshold, drive stability threshold, minimum angle change threshold, maximum drive change threshold, angle increment filtering threshold, gradient change judgment threshold, index spacing threshold, dead zone angle threshold, static friction judgment threshold, global angle merging threshold, angle deviation threshold, and safety factor can be pre-calibrated through offline simulation testing or set to fixed values ​​according to on-site operating procedures.

[0149] In the description of this specification, references to terms such as "an embodiment," "example," and "specific example" indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0150] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A valve detection and control method based on front-end and back-end, characterized in that, Including the following steps: S1: After the front-end terminal is powered on, it performs a self-test and zero-position calibration. Within the small stroke of the valve, it drives the valve stem to reciprocate micro-movements according to a preset test sequence. Under a unified time reference, it synchronously collects the valve stem angular displacement and the actuator driving amount and stores them in time order to form small stroke test data. S2: The front-end terminal preprocesses the small stroke trial data according to a uniform time step, calculates the driving amount and displacement increment of adjacent sampling points, and performs change point detection and segment fitting to obtain multiple stroke segments on the valve stem stroke. For each stroke segment, it calculates the segment-level characteristic parameters reflecting friction and hysteresis, generates a set of segment-level characteristic parameters, and uploads it to the back-end platform. S3: The backend platform receives the set of segment-level characteristic parameters and organizes them in the stroke segment dimension in accordance with the process constraint data of the circuit to which the valve belongs. Based on the simplified working condition model, it calculates the friction characteristics and flow capacity of each stroke segment, obtains the upper limit of safe opening and the maximum allowable action speed of each stroke segment, forms a set of health constraint parameters and sends them to the front-end terminal. S4: When the front-end terminal receives the target opening instruction, it converts the target opening into the target angle and locates the target angle on the stroke segment. Based on the health constraint parameter set, it corrects the target angle and action speed to generate the angle trajectory. During the execution process, it monitors the angle deviation by sampling. When the angle deviation continues to exceed the set standard, it performs backtracking control based on the last safe position and reports the abnormal event to the back-end platform.

2. The valve detection and control method based on front-end and back-end as described in claim 1, characterized in that: Step S1 includes: when generating a preset test sequence, the front-end terminal drives the valve stem to reciprocate micro-motion in a local range of the valve's full stroke, with the current working position as the center, in a manner that gradually increases the stroke amplitude. After collecting the valve stem angular displacement and the actuator drive amount, the small stroke test data with drive amount and displacement changes below the threshold are discarded, and the small stroke test data used for stroke segmentation processing are retained.

3. The valve detection and control method based on front-end and back-end as described in claim 2, characterized in that: Step S1 also includes: after the front-end terminal completes the zero-position calibration, it uses the mechanical limit or small-amplitude reciprocating rotation of the clamping mechanism to determine the reference point of the angle sensor, and assigns a unique test number to each test action under a unified time reference. At the same time, it records the test action direction, test start time, test end time, and stores them together with the small-stroke test data in the front-end local buffer.

4. The valve detection and control method based on front-end and back-end as described in claim 3, characterized in that: Step S2 includes: when the front-end terminal preprocesses the small-stroke trial data, it resamples the valve stem angular displacement and actuator drive quantity according to a uniform time step, reduces high-frequency noise through digital filtering, normalizes the amplitude of the resampling results, and then calculates the drive quantity increment and displacement increment of adjacent sampling points to construct an incremental sequence for change point detection.

5. The valve detection and control method based on front-end and back-end as described in claim 4, characterized in that: Step S2 also includes: after the front-end terminal completes the change point detection and segment fitting, it records the multiple stroke segments obtained by dividing the valve stem stroke in the stroke segment division table according to the stroke segment number, and writes the average slope, static friction peak value, hysteresis width and stroke segment type label corresponding to each stroke segment into the segment-level characteristic parameter set, and stores it with the stroke segment number as the index before uploading it to the back-end platform.

6. The valve detection and control method based on front-end and back-end as described in claim 5, characterized in that: Step S3 includes: The process constraint data includes at least the upstream pressure allowable range, downstream pressure allowable range, allowable differential pressure limit, target flow allowable range, and medium density of the circuit to which the valve belongs. The back-end platform writes the segment-level characteristic parameter set and the process constraint data into the same joint data structure in the stroke segment dimension to build a simplified operating condition model.

7. The valve detection and control method based on front-end and back-end as described in claim 6, characterized in that: Step S3 also includes: the backend platform calculates health indicators representing friction state and driving force requirements for each stroke segment based on the joint data structure, and divides the stroke segment into risk levels according to the pre-established threshold range. The health indicators and risk levels, together with the upper limit of the safe opening of the stroke segment and the maximum allowable speed of the stroke segment, form the stroke segment record in the health constraint parameter set.

8. The valve detection and control method based on front-end and back-end as described in claim 7, characterized in that: Step S4 includes: When the front-end terminal generates an angle trajectory by correcting the target angle and motion speed according to the health constraint parameter set, it first safely projects the target angle onto all stroke segments involved in the motion path to obtain a safe target angle, and then selects the maximum motion speed of the path from the maximum allowable motion speed of each stroke segment, and constructs a planned angle trajectory consisting of an acceleration segment, a constant speed segment and a deceleration segment according to the safe target angle, the maximum motion speed of the path and the sampling period.

9. The valve detection and control method based on front-end and back-end as described in claim 8, characterized in that: Step S4 also includes: when the front-end terminal monitors the angle deviation by sampling, it performs multiple reference action tests on the health valve in advance, statistically analyzes the distribution of the absolute value of the angle deviation at each sampling time, determines the angle deviation threshold based on the distribution, statistically analyzes the distribution of the number of consecutive out-of-tolerances in the health action, determines the standard for the number of consecutive out-of-tolerances based on the distribution, and uses the angle deviation threshold and the standard for the number of consecutive out-of-tolerances as the abnormal triggering conditions during online operation.

10. The valve detection and control method based on front-end and back-end as described in claim 9, characterized in that: Step S4 further includes: when the angle deviation continues to exceed the threshold and triggers backoff control, the front-end terminal records the actual angle position where the previous action ended and the angle deviation did not exceed the angle deviation threshold as the previous safe position. During the backoff control process, the previous safe position is used as the backoff endpoint, and an abnormal event record is generated after the backoff ends. The abnormal event record includes at least the original target opening instruction, the safe target angle, the action path travel segment set, the maximum angle deviation, the backoff start angle position, and the backoff end angle position, and is uploaded to the back-end platform.

Images