Working condition adaptive pressure calibration method and system for high-pressure pressure regulating valve

Abstract

The application relates to the technical field of pressure measurement calibration, and particularly discloses a working condition adaptive pressure calibration method and system for a high-pressure pressure regulating valve, which comprises the following steps: acquiring multi-dimensional working condition parameters and pressure measurement values of a valve body under a running working condition, and acquiring pressure pulsation characteristics according to the multi-dimensional working condition parameters, wherein the multi-dimensional working condition parameters comprise a valve front pressure sequence, a valve rear pressure sequence, a valve body temperature sequence, a valve body vibration signal sequence and a valve core position signal sequence; and acquiring a working condition classification result according to the pressure pulsation characteristics. According to the application, the valve front pressure, the valve rear pressure, the valve body temperature, the valve body vibration and the valve core position and other multi-dimensional working condition parameters are acquired, the pressure difference gradient, the temperature change rate, the vibration energy and the pressure pulsation characteristics based on wavelet multi-scale decomposition and multiple fractal analysis are extracted from the multi-dimensional working condition parameters, and different flow state working conditions such as laminar flow, transition flow and turbulent flow are identified.

Description

Technical Field

[0001] This invention relates to the field of pressure measurement and calibration technology, and in particular to a working condition adaptive pressure calibration method and system for a high-pressure regulating valve. Background Technology

[0002] High-pressure regulating valves are core regulating devices used in industrial pipeline systems for precise control of fluid pressure. They are widely used in high-pressure pipeline systems in fields such as petrochemicals, natural gas transmission, thermal power generation, and aerospace fuel supply.

[0003] When existing high-pressure regulating valves close or open rapidly, the fluid inside the valve cavity undergoes adiabatic compression or expansion. Simultaneously, the pressure changes dramatically within a very short time, and the fluid temperature also changes transiently: rapid closure causes the fluid to compress, leading to a sharp temperature increase; rapid opening causes the fluid to expand, leading to a sharp temperature decrease. This transient temperature change caused by fluid thermodynamics affects measurement accuracy through heat conduction on the pressure sensor diaphragm, resulting in a temperature-coupled error in the pressure measurement that is unrelated to the actual pressure. Traditional externally mounted temperature sensors can only sense temperature changes on the valve body's outer wall and cannot capture the instantaneous temperature field of the fluid inside the valve cavity. Furthermore, this error dynamically changes with the valve's operating speed and pressure change rate, and static calibration methods cannot compensate for this error source. This leads to a significant decrease in pressure measurement accuracy under rapid valve operation, making it difficult to meet the high-precision measurement requirements of high-pressure regulating valves under complex dynamic conditions. Summary of the Invention

[0004] The purpose of this invention is to provide a working condition adaptive pressure calibration method and system for a high-pressure regulating valve, so as to solve the technical problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: An adaptive pressure calibration method for a high-pressure regulating valve includes: The system acquires multi-dimensional operating parameters and pressure measurement values ​​of the valve body under operating conditions, and obtains pressure pulsation characteristics based on the multi-dimensional operating parameters. The multi-dimensional operating parameters include the valve inlet pressure sequence, valve outlet pressure sequence, valve body temperature sequence, valve body vibration signal sequence, and valve core position signal sequence. Based on the pressure pulsation characteristics, the operating condition classification result is obtained, and based on the operating condition classification result and the pressure measurement value, the basic calibration parameters are obtained. The dynamic response compensation amount is obtained based on the pressure measurement value and multi-dimensional operating condition parameters, and the benchmark calibration value is obtained based on the dynamic response compensation amount and the basic calibration parameters. The flow-related error compensation amount is obtained based on the multi-dimensional operating condition parameters and operating condition classification results, and the comprehensive calibration coefficient is obtained based on the benchmark calibration value and the flow-related error compensation amount. The pressure measurement value is calibrated and corrected according to the comprehensive calibration coefficient to obtain the calibrated pressure value; The basic calibration parameters are adaptively updated based on the calibration deviation between the calibrated pressure value and the pressure reference standard.

[0006] Preferably, the step of obtaining pressure pulsation characteristics based on the multidimensional operating condition parameters includes: The differential pressure sequence is obtained based on the inlet and outlet pressure sequences, and the differential pressure gradient characteristics are obtained based on the differential pressure sequence. The temperature change rate characteristics are obtained based on the valve body temperature sequence; Vibration energy characteristics are obtained based on the valve body vibration signal sequence; Pressure pulsation characteristics are obtained based on the pre-valve pressure sequence and post-valve pressure sequence.

[0007] Preferably, the step of obtaining pressure pulsation characteristics based on the pre-valve pressure sequence and the post-valve pressure sequence includes: The pressure pulsation components before and after the valve are obtained based on the pressure sequence before and after the valve, respectively. The pressure pulsation sequence is obtained based on the pre-valve pressure pulsation component and the post-valve pressure pulsation component. The pressure pulsation sequence is decomposed into wavelet multi-scale values ​​to obtain wavelet coefficient energies at multiple scales. The scaling exponent and multifractal spectrum width are obtained based on the wavelet coefficient energy at the multiple scales. The pulsation frequency characteristics are obtained based on the scaling index, and the pressure pulsation characteristics are obtained based on the multifractal spectrum width.

[0008] Preferably, the step of obtaining the operating condition classification result based on the pressure pulsation characteristics includes: Obtain the pulsation feature vector based on the pressure pulsation characteristics; Obtain template feature vectors for multiple preset flow state categories, wherein the flow state categories include laminar steady-state category, transitional flow state category, and turbulent flow state category; A distance set is obtained based on the distance between the pulsating feature vector and each of the template feature vectors; Obtain the working condition classification result by identifying the working condition flow state category corresponding to the minimum distance in the distance set.

[0009] Preferably, the step of obtaining the dynamic response compensation amount based on the pressure measurement value and multi-dimensional operating condition parameters includes: The sensor output sequence is obtained based on the pressure measurement value; Obtain the sensor's natural frequency and damping ratio, and construct the sensor transfer function based on the natural frequency and damping ratio; The inverse filter is obtained based on the sensor transfer function; A preliminary deconvolution sequence is obtained based on the inverse filter and the sensor output sequence; The regularization parameter is obtained from the initial deconvolution sequence, and the initial deconvolution sequence is regularized according to the regularization parameter to obtain the pressure estimation sequence after deconvolution. The dynamic response compensation amount is obtained based on the deconvolutioned pressure estimation sequence and the sensor output sequence.

[0010] Preferably, the step of obtaining the flow-related error compensation amount based on the multi-dimensional operating condition parameters and operating condition classification results includes: Obtain the specific heat ratio and current pressure measurement of the fluid; The pressure change rate is obtained based on the pre-valve pressure sequence and the post-valve pressure sequence. The estimated adiabatic temperature rise is obtained based on the pressure change rate and specific heat ratio. The temperature sensitivity coefficient of the sensor is obtained, and the fluid adiabatic effect compensation component is obtained based on the estimated adiabatic temperature rise and the temperature sensitivity coefficient. The sensor deviation compensation component is obtained based on the multi-dimensional operating parameters. The flow-related error compensation amount is obtained based on the fluid adiabatic effect compensation component and the sensor deviation compensation component.

[0011] Preferably, the step of obtaining the inter-sensor deviation compensation component based on the multi-dimensional operating condition parameters includes: Acquire multiple measurement values ​​from multiple sensors under the same operating conditions; Obtain the deviation matrix between the measurement values ​​based on the multiple measurement values; The mean deviation vector and the covariance deviation matrix are obtained from the deviation matrix. The deviation covariance matrix is ​​decomposed into eigenvalues ​​to obtain eigenvalues ​​and eigenvectors. The consistent deviation component is obtained by projecting the deviation along the direction of the eigenvector corresponding to the largest eigenvalue. The difference deviation component is obtained by removing the uniform deviation component from the remaining deviation. The difference deviation component is used to obtain the inter-sensor deviation compensation component.

[0012] Preferably, the step of obtaining the basic calibration parameters based on the operating condition classification results and pressure measurement values ​​includes: Obtain the preset flow regime-coefficient mapping relationship, and query the flow regime-coefficient mapping relationship according to the working condition classification result to obtain the reference coefficient under the corresponding flow regime; The sensor signal quality assessment value is obtained based on the pressure measurement value and the reference coefficient; The sensor confidence weight is obtained based on the sensor signal quality assessment value; The basic calibration parameters are obtained based on the reference coefficients and sensor confidence weights.

[0013] Preferably, the step of adaptively updating the basic calibration parameters based on the calibration deviation between the calibrated pressure value and the pressure reference standard includes: The residual is obtained based on the calibrated pressure value and the pressure reference standard; Determine whether the absolute value of the residual is greater than a preset update threshold; If the absolute value of the residual is greater than the preset update threshold, then the corresponding element in the flow-coefficient mapping relationship is updated according to the residual and the preset learning rate; If the absolute value of the residual is not greater than the preset update threshold, then obtain the consecutive non-updated count and determine whether the consecutive non-updated count is greater than the preset forgetting threshold. If the number of consecutive unupdated values ​​is greater than the preset forgetting threshold, then the corresponding elements in the flow-coefficient mapping relationship are decayed and updated according to the preset forgetting factor.

[0014] This invention also discloses an adaptive pressure calibration system for a high-pressure regulating valve, comprising: The data acquisition module is used to acquire multi-dimensional operating condition parameters and pressure measurement values ​​of the valve body under operating conditions, and to acquire pressure pulsation characteristics based on the multi-dimensional operating condition parameters. The multi-dimensional operating condition parameters include the valve inlet pressure sequence, valve outlet pressure sequence, valve body temperature sequence, valve body vibration signal sequence, and valve core position signal sequence. The parameter acquisition module is used to obtain the working condition classification result based on the pressure pulsation characteristics, and to obtain the basic calibration parameters based on the working condition classification result and the pressure measurement value. The basic calibration module is used to obtain the dynamic response compensation amount based on the pressure measurement value and multi-dimensional operating condition parameters, and to obtain the benchmark calibration value based on the dynamic response compensation amount and basic calibration parameters. The benchmark acquisition module is used to obtain the flow-related error compensation amount based on the multi-dimensional working condition parameters and working condition classification results, and to obtain the comprehensive calibration coefficient based on the benchmark calibration value and the flow-related error compensation amount. The calibration output module is used to calibrate and correct the pressure measurement value according to the comprehensive calibration coefficient, and obtain the calibrated pressure value. The update module is used to adaptively update the basic calibration parameters based on the calibration deviation between the calibrated pressure value and the pressure reference standard.

[0015] The beneficial effects of this application are as follows: This invention acquires multi-dimensional operating condition parameters such as inlet pressure, outlet pressure, valve body temperature, valve body vibration, and valve core position, and extracts pressure gradient, temperature change rate, vibration energy, and pressure pulsation characteristics (scaling exponent and multifractal spectrum width) based on wavelet multi-scale decomposition and multifractal analysis to identify different flow regimes such as laminar flow, transitional flow, and turbulent flow. For sensor dynamic response hysteresis, a transfer function is constructed based on the sensor's natural frequency and damping ratio, and dynamic response compensation is obtained through regularized deconvolution. For fluid adiabatic effects, an estimated adiabatic temperature rise is calculated based on the specific heat ratio and pressure change rate, and the influence of temperature sensitivity is compensated. For multi-sensor common-mode interference, consistent deviation and differential deviation are separated through deviation matrix eigenvalue decomposition, compensating only for individual differences. Finally, the above compensations are combined to obtain calibration coefficients to correct the pressure measurement value, and the basic calibration parameters are adaptively evolved through an "update-forget" dual mechanism based on the deviation between the calibrated pressure and the reference standard. This invention achieves high-precision pressure calibration under all operating conditions without relying on external reference signals by transforming pressure pulsation from "noise" into flow regime discrimination features, deconvolution to compensate for sensor dynamic hysteresis, thermodynamic model to compensate for adiabatic effects, eigenvalue decomposition to eliminate common mode deviation, and parameter adaptive evolution. This significantly improves the measurement accuracy and long-term reliability of high-pressure regulating valves under complex dynamic conditions. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of a method flow according to an embodiment of this application.

[0017] Figure 2 This is a schematic diagram of the system structure according to an embodiment of this application.

[0018] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0019] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0020] like Figure 1 As shown, this application provides a condition-adaptive pressure calibration method for a high-pressure regulating valve, comprising: S1. Obtain multi-dimensional operating condition parameters and pressure measurement values ​​of the valve body under operating conditions, and obtain pressure pulsation characteristics based on the multi-dimensional operating condition parameters, wherein the multi-dimensional operating condition parameters include the valve inlet pressure sequence, valve outlet pressure sequence, valve body temperature sequence, valve body vibration signal sequence and valve core position signal sequence. For example, a preliminary dynamic response compensation amount is obtained based on the pressure measurement value, and the pressure measurement value is subjected to preliminary dynamic compensation to obtain a compensated pressure sequence. Pressure pulsation characteristics are obtained based on the compensated pressure sequence and the multi-dimensional operating condition parameters.

[0021] S2. Obtain the operating condition classification result based on the pressure pulsation characteristics, and obtain the basic calibration parameters based on the operating condition classification result and the pressure measurement value; S3. Obtain the dynamic response compensation amount based on the pressure measurement value and multi-dimensional operating condition parameters, and obtain the benchmark calibration value based on the dynamic response compensation amount and basic calibration parameters. S4. Obtain the flow-related error compensation amount based on the multi-dimensional working condition parameters and working condition classification results, and obtain the comprehensive calibration coefficient based on the benchmark calibration value and the flow-related error compensation amount. S5. The pressure measurement value is calibrated and corrected according to the comprehensive calibration coefficient to obtain the calibrated pressure value; S6. Based on the calibration deviation between the calibrated pressure value and the pressure reference standard, the basic calibration parameters are adaptively updated.

[0022] As described in steps S1-S6 above, existing high-pressure regulating valve pressure calibration technology mainly adopts a static calibration method before delivery, that is, obtaining fixed calibration coefficients in a standard laboratory environment for long-term use after installation. However, in actual industrial operation, high-pressure regulating valves face a variety of complex operating conditions: the dynamic response lag of the sensor during rapid opening and closing leads to amplitude attenuation and phase delay; the adiabatic compression or expansion of the fluid in the valve cavity during rapid valve action causes transient temperature changes, which in turn affect the temperature characteristics of the pressure sensor diaphragm; when multiple pressure sensors are simultaneously subjected to interference such as common-mode vibration and common-mode temperature drift, a systematic consistency deviation that cannot be eliminated by traditional weighted fusion is generated; the error characteristics of pressure measurement systems under different flow states—laminar, transitional, and turbulent—are fundamentally different, and traditional methods treat pressure pulsations as noise filtering, discarding crucial flow state information.

[0023] This invention obtains pressure pulsation characteristics through pre-valve and post-valve pressure sequences, including the scaling exponent and multifractal spectrum width after wavelet multi-scale decomposition. This transforms turbulent pulsation signals, traditionally considered noise, into key features characterizing the flow regime, providing a physically meaningful basis for subsequent operational condition classification. This achieves a shift in flow regime information from being "filtered out" to being "utilized." Based on the pressure pulsation characteristics, operational condition classification results are obtained. Then, based on the operational condition classification results and pressure measurements, basic calibration parameters are obtained, enabling differentiated calibration parameter selection based on flow regime identification. This solves the problem of different measurement deviation characteristics under different flow regimes. Simultaneously, by dynamically adjusting the sensor confidence weights through signal-to-noise ratio, the interference of low-quality data on calibration parameters is reduced.

[0024] The preliminary dynamic response compensation adopts a simplified form of the dynamic response compensation amount described in step S3 (a first-order deconvolution based on the sensor's nominal natural frequency and damping ratio). The purpose is to eliminate the main distortion of the pressure pulsation signal by the sensor's frequency response characteristics before flow pattern identification, and to avoid misjudging turbulence as laminar flow due to high-frequency attenuation of the sensor.

[0025] In the subsequent step S3, a fine calculation of dynamic response compensation will be performed based on more accurate real-time parameters, and the compensation result will be fused with the preliminary compensation result to improve the calibration accuracy under dynamic operating conditions.

[0026] The dynamic response compensation is obtained based on pressure measurements and multi-dimensional operating parameters. A transfer function is constructed using the sensor's natural frequency and damping ratio. Frequency domain deconvolution processing is then performed using Tikhonov regularization to recover the true pressure signal from the sensor output sequence, eliminating amplitude attenuation and phase delay errors introduced by the sensor's second-order inertial system under rapid transition conditions. A benchmark calibration value is obtained based on the dynamic response compensation and fundamental calibration parameters, enabling the calibration benchmark to track real pressure changes under dynamic operating conditions.

[0027] The flow-related error compensation amount is obtained based on multi-dimensional operating parameters and operating condition classification results. On the one hand, the adiabatic temperature rise estimate is calculated based on the ideal gas adiabatic process equation, and the fluid adiabatic effect compensation component is obtained by combining the sensor temperature sensitivity coefficient. For the first time, the fluid adiabatic effect is introduced into the pressure calibration compensation system, filling the gap in the existing technology that ignores transient thermodynamic processes. On the other hand, multiple sensor measurements are collected to construct a deviation matrix and perform eigenvalue decomposition. The deviation projection in the direction of the eigenvector corresponding to the largest eigenvalue is removed as a consistent deviation component, and the remaining differential deviation component is used as the inter-sensor deviation compensation component, eliminating the systematic consistent deviation caused by multi-sensor common-mode interference.

[0028] The pressure measurement value is calibrated and corrected based on the comprehensive calibration coefficient to obtain the calibrated pressure value, ensuring high accuracy of the calibration results under steady-state, transient, and complex flow conditions. Based on the calibration deviation between the calibrated pressure value and the pressure reference standard, the basic calibration parameters are adaptively updated: when the residual exceeds a threshold, updates are made incrementally according to the learning rate; when there has been no update for a long period, updates are made attenuated according to the forgetting factor. This dual "update-forgetting" mechanism enables online evolution of the calibration parameters, allowing for rapid correction of deviations and automatic decay of outdated parameters, enhancing the system's long-term adaptability to sensor aging and operational drift.

[0029] This invention does not rely on external standard signal sources, but performs continuous calibration entirely through multi-dimensional operating parameters within the valve body. This allows the high-pressure regulating valve to maintain accurate pressure measurement even in industrial environments without external reference signals. By fusing multi-source sensor data, including upstream pressure sequence, downstream pressure sequence, valve body temperature sequence, valve body vibration signal sequence, and valve core position signal sequence, the system can adapt to changes in operating conditions in real time. When the valve is in different flow states, it automatically matches differentiated calibration strategies to ensure measurement accuracy across the entire operating range. Through multifractal analysis of pressure pulsations, dynamic deconvolution compensation of sensors, fluid adiabatic effect compensation, and inter-sensor deviation decomposition, this invention not only uses static benchmark calibration parameters but also incorporates multiple real-time signals such as flow state information, dynamic response, and thermo-coupling for calibration. This enables the system to compensate for measurement errors caused by flow state changes, rapid actions, temperature transients, and sensor drift, enhancing the adaptability of the high-pressure regulating valve in complex industrial environments.

[0030] In one embodiment, the step of obtaining pressure pulsation characteristics based on the multidimensional operating condition parameters includes: S101. Obtain the differential pressure sequence based on the pre-valve pressure sequence and the post-valve pressure sequence, and obtain the differential pressure gradient characteristics based on the differential pressure sequence; For example, the pre-valve pressure sequence and the post-valve pressure sequence are obtained, at each sampling point number. The formula for calculating the pressure difference sequence is: ,in, Indicates the sampling point number. Represents the pressure difference sequence. This represents the pre-valve pressure sequence. Indicates the pressure sequence after the valve; Then, by pre-selecting the time window width (Taking 100 sampling periods), perform central difference operation on the pressure difference sequence within the sliding window to obtain the pressure difference gradient feature. The formula for calculating the pressure difference gradient feature is as follows: ; in, Indicates the characteristics of pressure gradient. Indicates the width of the pre-selected time window.

[0031] S102. Obtain the temperature change rate characteristics based on the valve body temperature sequence; S103. Obtain vibration energy characteristics based on the valve body vibration signal sequence; S104. Obtain pressure pulsation characteristics based on the pre-valve pressure sequence and post-valve pressure sequence.

[0032] As described in steps S101-S104 above, this invention extracts five characteristic state quantities from multi-dimensional operating parameters, namely, differential pressure gradient, temperature change rate, vibration energy, pulsation frequency, and pulsation intensity. These characteristics each reflect different aspects of the valve's operating state.

[0033] The differential pressure gradient is calculated by subtracting the upstream and downstream pressures to obtain a differential pressure sequence, and then calculating its rate of change. The differential pressure itself reflects the energy loss of fluid passing through the valve's throttling section, while the differential pressure gradient further characterizes how quickly this energy loss changes. When the valve is rapidly adjusted or the upstream gas supply fluctuates, the differential pressure changes significantly in a short period, resulting in a high differential pressure gradient value. When the valve is operating in a steady state, the differential pressure gradient approaches zero. This characteristic provides a basis for subsequently determining whether the operating condition is in a transient process.

[0034] The rate of temperature change is a differential and smoothed representation of the valve body temperature sequence. Valve body temperature changes are typically slow, and directly using temperature values ​​is insufficient to capture the dynamic changes in the temperature field. The rate of temperature change, however, reflects the activity level of the thermal process. When a valve actuates rapidly, causing adiabatic compression or expansion of the fluid, the internal temperature of the valve cavity undergoes a transient change. Although this change is not immediately and completely transmitted to the valve body surface, abnormal fluctuations in the rate of temperature change can serve as a reference signal for activating thermal compensation strategies.

[0035] Vibration energy characteristics involve time-frequency analysis of vibration signals. Valves are subjected to various mechanical excitations during operation: fluid impact on the valve core generates mid-to-high frequency vibrations, external pipeline vibrations are transmitted to the valve body through connecting flanges, and cavitation produces vibrations at characteristic frequencies. By decomposing the vibration signal into different frequency bands and calculating the energy of each band, a vibration energy characteristic vector containing multiple components can be obtained. When the energy of a certain frequency band abnormally increases, a corresponding interference source can be identified, providing a basis for subsequent identification and compensation of common-mode interference between sensors.

[0036] Pressure pulsation characteristics involve directly filtering out the fluctuating components of the pressure signal as noise. First, the pulsating component is separated from the original pressure signal. Then, continuous wavelet transform is used to analyze the distribution characteristics of the pulsating energy across different frequency bands. The scaling exponent reflects whether the energy is concentrated at low frequencies or dispersed to high frequencies, while the multifractal spectral width reflects the strength of the intermittency of the pulsating signal. By jointly analyzing these two indices, different flow regimes such as laminar, transitional, and turbulent flows can be distinguished. For example, in laminar flow, the pulsation is weak and concentrated at low frequencies, with a large scaling exponent and a small spectral width; in turbulent flow, the pulsation is strong and the frequency band is broadened, with a small scaling exponent and a large spectral width.

[0037] In one embodiment, the step of obtaining pressure pulsation characteristics based on the pre-valve pressure sequence and the post-valve pressure sequence includes: S1041. Obtain the pressure pulsation component before the valve and the pressure pulsation component after the valve according to the pressure sequence before the valve and the pressure sequence after the valve, respectively. S1042. Obtain the pressure pulsation sequence based on the pressure pulsation component before the valve and the pressure pulsation component after the valve. S1043. Perform wavelet multi-scale decomposition on the pressure pulsation sequence to obtain wavelet coefficient energy at multiple scales; S1044. Obtain the scaling exponent and multifractal spectrum width based on the wavelet coefficient energy at the multiple scales; S1045. Obtain the pulsation frequency characteristics based on the scaling index, and obtain the pressure pulsation characteristics based on the multifractal spectrum width.

[0038] As described in steps S1041-S1045 above, this invention separates the steady-state component and the pulsating component from the original pressure signal by performing low-pass filtering on the pre-valve pressure sequence and the post-valve pressure sequence, respectively. The steady-state component corresponds to the slow change trend of pressure, while the pulsating component retains the high-frequency fluctuation components caused by fluid turbulence, valve opening and closing actions, etc. After combining the pre-valve and post-valve pulsating components into a pressure pulsation sequence, the sequence is subjected to continuous wavelet transform using the Morlet wavelet basis function to obtain the wavelet coefficient energy distribution at different scales. The scale of the wavelet transform is inversely proportional to the frequency of the signal. Small scales correspond to high-frequency components, reflecting small eddies and rapid pulsations in the fluid, while large scales correspond to low-frequency components, reflecting large-scale eddies and slow fluctuations in the fluid.

[0039] The scaling exponent is obtained by power-law fitting of wavelet coefficient energy and scale. This exponent reflects the concentration of pressure pulsation energy in different frequency bands. When the scaling exponent is large, the energy is mainly concentrated in the low-frequency band, indicating that the fluid flow is relatively stable and the pulsations mainly originate from large-scale vortices. When the scaling exponent is small, the energy diffuses to the high-frequency band, indicating that the fluid flow contains more small-scale vortices and rapid pulsations. Based on this, multifractal analysis is further performed on the pressure pulsation sequence. The intermittent characteristics of the pressure pulsation are described by calculating the multifractal spectral width. The larger the spectral width, the more complex the signal structure, indicating that events of different intensities alternate in the pressure pulsation, reflecting the non-uniform distribution of the scale and intensity of vortices in the flow field. When the spectral width approaches zero, it indicates that the pressure pulsation is close to uniform distribution and the flow field structure is relatively simple.

[0040] Using the scaling exponent as the pulsation frequency characteristic and the multifractal spectrum width as the pulsation intensity characteristic, these two elements complement each other to describe the essential characteristics of pressure pulsation. Existing technologies typically treat pressure pulsation as noise to be filtered out, discarding the flow regime information it contains. This embodiment, however, extracts the scaling exponent through wavelet multi-scale decomposition and the spectrum width through multifractal analysis, transforming the pulsation signal, traditionally considered interference, into quantifiable feature parameters. These features are independent of external reference signals and are entirely obtained from the pressure signal collected by the valve itself, enabling the system to autonomously perceive the flow state of the fluid within the valve. In practical applications, a large scaling exponent and a small multifractal spectrum width indicate a relatively regular fluid flow state, with less pulsation interference in pressure measurement. Conversely, as the scaling exponent gradually decreases and the spectrum width gradually increases, it indicates a change in the fluid flow state and increased turbulence intensity, requiring more flow regime-related compensation considerations in pressure measurement.

[0041] In one embodiment, the step of obtaining the operating condition classification result based on the pressure pulsation characteristics includes: S201. Obtain the pulsation feature vector based on the pressure pulsation characteristics; S202. Obtain template feature vectors of multiple preset working condition flow state categories, wherein the working condition flow state categories include laminar steady state category, transitional flow state category and turbulent flow state category; S203. Obtain a distance set based on the distance between the pulsating feature vector and each of the template feature vectors; S204. Obtain the working condition classification result by identifying the working condition flow state category corresponding to the minimum distance in the distance set.

[0042] As described in steps S201-S204 above, this invention obtains a pulsation feature vector through pressure pulsation characteristics, combining the pulsation frequency feature and the pulsation intensity feature into a two-dimensional feature vector. These two features describe the nature of pressure pulsation from different perspectives. The pulsation frequency feature reflects the concentration of pulsation energy in the frequency domain, while the pulsation intensity feature reflects the non-uniformity of the pulsation signal. Template feature vectors for multiple preset flow state categories are obtained, where laminar steady-state, transitional flow, and turbulent flow categories represent different flow states of the fluid within the valve cavity. In laminar flow, the fluid flows regularly with a uniform velocity distribution. The transitional flow is an intermediate state between laminar and turbulent flow, while in turbulent flow, the fluid flows chaotically with numerous vortices. These template feature vectors are obtained by offline collection of historical operating data and cluster analysis, representing typical values ​​of pressure pulsation characteristics under various flow states.

[0043] Calculate the Euclidean distance between the current pulsating feature vector and each template feature vector. The magnitude of the Euclidean distance reflects the similarity between the current flow characteristics and the characteristics of various typical flow states; the smaller the distance, the closer the current flow state is to that category. Select the flow state category corresponding to the smallest distance from the distance set as the operating condition classification result, that is, consider the current flow state of the fluid in the valve to be the closest to that typical flow state.

[0044] In existing technologies, pressure pulsations are typically treated as noise to be filtered out, discarding the flow regime information they contain. This embodiment, however, achieves automatic flow regime identification by matching pressure pulsation characteristics with a preset flow regime template. When the pulsation frequency is high and the pulsation intensity is low, the current feature vector is close to the template feature vector of the laminar steady-state category, and the system determines it as a laminar steady-state condition. As the pulsation frequency gradually decreases and the pulsation intensity gradually increases, the feature vector moves closer to the transitional or turbulent flow regime template, and the system automatically switches the condition classification result. This flow regime identification method based on feature distance does not rely on external reference signals; it judges entirely based on the characteristics of the pressure pulsations themselves, enabling the system to autonomously perceive changes in the flow state of the fluid within the valve. This provides a basis for subsequently selecting differentiated calibration parameters based on the flow regime. In practical applications, when the valve opening is small and the fluid velocity is low, the system automatically identifies it as a laminar steady-state condition and adopts a calibration strategy suitable for smooth flow. When the valve opens rapidly and the fluid velocity increases, the system detects the change in flow state and automatically switches the calibration parameters to match the calibration strategy with the current flow state.

[0045] In one embodiment, the step of obtaining the basic calibration parameters based on the operating condition classification result and the pressure measurement value includes: S205. Obtain a preset flow state-coefficient mapping relationship, and query the flow state-coefficient mapping relationship according to the working condition classification result to obtain the reference coefficient under the corresponding flow state; S206. Obtain the sensor signal quality evaluation value based on the pressure measurement value and the reference coefficient; S207. Obtain the sensor confidence weight based on the sensor signal quality evaluation value; S208. Obtain the basic calibration parameters based on the reference coefficients and sensor confidence weights.

[0046] As described in steps S205-S208 above, this invention uses a preset flow regime-coefficient mapping relationship, taking the operating condition classification results obtained in step S2 as a query index, and extracting the benchmark coefficient corresponding to the current flow regime from the mapping table. Since the systematic error characteristics of pressure measurement differ fundamentally under different flow regimes—laminar flow has uniform pressure distribution and measurement deviation mainly originates from temperature drift, while turbulent flow has severe pressure pulsation and measurement deviation mainly originates from vibration coupling and hydrodynamic effects—pre-setting differentiated benchmark coefficients for different flow regimes allows the selection of subsequent calibration parameters to match the current flow regime, avoiding the use of the same set of coefficients to adapt to all operating conditions.

[0047] Based on the obtained baseline coefficients, the short-time signal-to-noise ratio (SNR) is calculated using the pressure measurements as an evaluation value for sensor signal quality. The mean of the pressure measurements within the time window reflects the effective component of the signal, while the variance of the difference between the pressure measurements and their moving mean reflects the noise level. The ratio of these two values ​​is logarithmically transformed to obtain the SNR. When the sensor is subjected to common-mode vibration, electromagnetic interference, or local flow field disturbances, the noise component in the measurements increases significantly, and the SNR decreases accordingly. In this case, the sensor signal quality is low, and the weight of this sensor needs to be reduced in subsequent fusion. When the sensor operates in a good environment, the SNR remains at a high level, indicating that the measurement values ​​are highly reliable.

[0048] The short-time signal-to-noise ratio (SNR) is converted into sensor confidence weights using a mapping function. This mapping function has the following characteristics: when the SNR is lower than a preset SNR benchmark, the output approaches 0, indicating that the sensor signal quality is poor and unsuitable for calibration; when the SNR is higher than the benchmark, the output approaches 1, indicating that the sensor signal quality is good and can be used as the primary calibration basis; near the benchmark, the output transitions continuously with changes in SNR, avoiding abrupt changes in calibration parameters caused by sudden weight shifts. Through this mapping method, the system can adaptively adjust the confidence level of each sensor according to signal quality, reducing its weight when the signal quality is poor and increasing its weight when the signal quality is good.

[0049] Finally, the basic calibration parameters are obtained based on the reference coefficients and sensor confidence weights. When the sensor confidence weight is high, the basic calibration parameters are mainly determined by the reference coefficients corresponding to the current flow state. In this case, the system trusts the current sensor measurement value and considers the reference coefficients applicable under the current flow state. When the sensor confidence weight is low, the basic calibration parameters regress to the default calibration coefficients. In this case, the system considers the sensor signal to be significantly affected by interference and should not rely excessively on the current measurement value, so it adopts the more conservative default coefficients. This approach preserves the calibration characteristics of flow state differences while avoiding interference from low-quality data on the calibration parameters. In practical applications, when the sensor signal is affected by external interference, causing a decrease in the signal-to-noise ratio, the system can automatically reduce the impact of that sensor on the calibration parameters, preventing erroneous data from contaminating the calibration results. When the signal returns to normal, the system can quickly restore its trust in the flow state reference coefficients. This adaptive weighting mechanism based on signal quality enables the calibration parameters to remain stable and reliable under different operating conditions and signal quality conditions.

[0050] In one embodiment, the step of obtaining the dynamic response compensation amount based on the pressure measurement value and multi-dimensional operating condition parameters includes: S301. Obtain the sensor output sequence based on the pressure measurement value; S302. Obtain the sensor's natural frequency and damping ratio, and construct the sensor transfer function based on the natural frequency and damping ratio; For example, the sensor transfer function process is as follows: Obtain the natural angular frequency, damping ratio, and Laplace operator of the pressure sensor, and construct the sensor transfer function: ; in, Represents the sensor transfer function. Represents the Laplace operator. This represents the sensor's natural angular frequency. This indicates the damping ratio.

[0051] S303. Obtain the inverse filter based on the sensor transfer function; S304. Obtain a preliminary deconvolution sequence based on the inverse filter and the sensor output sequence; S305. Obtain regularization parameters based on the initial deconvolution sequence, and perform regularization processing on the initial deconvolution sequence based on the regularization parameters to obtain the pressure estimation sequence after deconvolution. S306. Obtain the dynamic response compensation amount based on the deconvolutioned pressure estimation sequence and the sensor output sequence.

[0052] As described in steps S301-S306 above, the present invention constructs a sensor transfer function by using the pressure sensor output sequence, the sensor's natural frequency, and the damping ratio, and obtains an inverse filter based on the transfer function. A preliminary deconvolution sequence is obtained through the inverse filter and the sensor output sequence. Then, a regularization parameter is obtained based on the preliminary deconvolution sequence to perform regularization processing on the preliminary deconvolution sequence, resulting in a deconvolutioned pressure estimation sequence. Finally, the dynamic response compensation amount is obtained based on the deconvolutioned pressure estimation sequence and the sensor output sequence.

[0053] A pressure sensor is a second-order inertial system with an inherent natural frequency and damping ratio. When the pressure changes rapidly, the sensor's output cannot immediately keep up with the actual pressure change, but rather exhibits a transition process: during a step increase in pressure, the sensor output shows a rise delay and overshoot; during rapid pressure fluctuations, the sensor output shows amplitude attenuation and phase lag. This dynamic response lag has little impact under steady-state conditions, but under dynamic conditions such as rapid valve opening and closing or sudden load changes, it can introduce significant measurement errors.

[0054] To address this problem, this invention describes the frequency response relationship between the sensor output and the actual pressure by constructing a sensor transfer function. Transfer Function This reflects the sensor's response characteristics to input signals of different frequencies: in the low-frequency range, the sensor output can follow the input signal well; as the frequency increases and approaches the natural frequency, the output amplitude attenuates and the phase lags. The inverse filter is theoretically the reciprocal of the transfer function; applying it to the sensor output can recover the true pressure signal.

[0055] However, directly using an inverse filter presents a problem: the sensor transfer function tends to decay at high frequencies, while its inverse tends to amplify at high frequencies. Since high-frequency noise always exists in actual measurement signals, this noise is excessively amplified after passing through the inverse filter, causing severe oscillations in the deconvolution result, rendering it unusable. Therefore, this invention introduces regularization processing, determining the regularization parameter (using...) based on the energy magnitude of the high-frequency components in the initial deconvolution sequence. (This indicates that a larger value is used when the high-frequency noise energy is large). To suppress noise amplification, a smaller value is used when the high-frequency noise energy is low. To ensure deconvolution accuracy, the Tikhonov regularization method introduces regularization into the denominator. This approach achieves a balance between deconvolution accuracy and noise suppression.

[0056] The dynamic response compensation is obtained by subtracting the deconvolutioned pressure estimate sequence from the original sensor output sequence. When the sensor response is fast enough and the output accurately reflects the true pressure, the deconvolutioned estimate is close to the output sequence, and the compensation approaches zero. When the sensor response is lagging and the output deviates from the true pressure, the compensation is positive or negative, used to correct for amplitude attenuation and phase lag in the sensor output. This compensation is subsequently added to the calibration parameters, enabling the calibration benchmark to track changes in true pressure under dynamic operating conditions.

[0057] In practical applications, when a valve closes or opens rapidly, the pressure inside the valve chamber changes dramatically within a very short time. The sensor output often lags behind the actual pressure, resulting in measured values ​​that are either too low or too high. The dynamic response compensation obtained in this embodiment can be used to correct the measured value, making the calibrated pressure value closer to the actual pressure. This deconvolution compensation method based on the sensor's physical characteristics does not rely on external reference signals; it calculates entirely based on the sensor's inherent parameters and the actual output signal. It is applicable to various types and specifications of pressure sensors, exhibiting strong versatility and engineering practicality.

[0058] In one embodiment, the step of obtaining the flow-related error compensation amount based on the multidimensional operating condition parameters and the operating condition classification result includes: S401. Obtain the specific heat ratio and current pressure measurement of the fluid; S402. Obtain the pressure change rate based on the pre-valve pressure sequence and the post-valve pressure sequence; S403. Obtain the estimated value of adiabatic temperature rise based on the pressure change rate and specific heat ratio; S404. Obtain the temperature sensitivity coefficient of the sensor, and obtain the fluid adiabatic effect compensation component based on the estimated adiabatic temperature rise and the temperature sensitivity coefficient. S405. Obtain the sensor deviation compensation component based on the multi-dimensional operating condition parameters; S406. Obtain the flow-related error compensation amount based on the fluid adiabatic effect compensation component and the sensor deviation compensation component.

[0059] As described in steps S401-S406 above, the present invention obtains the specific heat ratio and current pressure measurement value of the fluid, obtains the pressure change rate according to the pressure sequence before and after the valve, obtains the adiabatic temperature rise estimate according to the pressure change rate and specific heat ratio, obtains the temperature sensitivity coefficient of the sensor, obtains the fluid adiabatic effect compensation component according to the adiabatic temperature rise estimate and temperature sensitivity coefficient, obtains the inter-sensor deviation compensation component according to the multi-dimensional operating condition parameters, and obtains the flow state related error compensation amount according to the fluid adiabatic effect compensation component and the inter-sensor deviation compensation component.

[0060] When the high-pressure regulating valve closes rapidly, the fluid inside the valve chamber is quickly compressed, causing a sharp increase in pressure. This process occurs in a very short time, and heat cannot be exchanged with the outside environment quickly enough; it is an adiabatic compression process, and the fluid temperature rises accordingly. Conversely, when the valve opens rapidly, the fluid expands quickly, causing a sharp drop in pressure; this is an adiabatic expansion process, and the fluid temperature decreases accordingly. This transient temperature change caused by fluid thermodynamic processes affects the measurement accuracy through heat conduction on the pressure sensor diaphragm. Traditional externally mounted temperature sensors cannot detect the instantaneous temperature field of the fluid inside the valve chamber, leading to this error source being neglected for a long time.

[0061] To address this issue, this invention calculates the estimated adiabatic temperature rise based on the ideal gas adiabatic process equation. When the valve actuates, the pressure values ​​before and after the actuation are acquired, the pressure change rate is calculated, and combined with the fluid's specific heat ratio and the current absolute temperature, the temperature change generated by the fluid during adiabatic compression or expansion is estimated. When the pressure increases, the estimated adiabatic temperature rise is positive, indicating a rise in fluid temperature; when the pressure decreases, the estimated adiabatic temperature rise is negative, indicating a decrease in fluid temperature. The temperature sensitivity coefficient of the pressure sensor is obtained, reflecting the sensor's output sensitivity to temperature changes. The estimated adiabatic temperature rise is multiplied by the temperature sensitivity coefficient and the current pressure measurement value to obtain the fluid adiabatic effect compensation component. When the valve closes rapidly, causing the fluid temperature to rise, the compensation component is positive, used to correct the measurement deviation caused by the temperature rise; when the valve opens rapidly, causing the fluid temperature to fall, the compensation component is negative, used to correct the measurement deviation caused by the temperature fall.

[0062] When multiple pressure sensors operate simultaneously, deviations exist between them, including uniformity deviation components and differential deviation components. The uniformity deviation component originates from common disturbances experienced by all sensors, such as common-mode vibration, common-mode temperature drift, and shared installation effects, manifesting as a shift in the measured values ​​of each sensor in the same direction. The differential deviation component originates from the aging drift and installation differences unique to each sensor, manifesting as relative differences in the measured values ​​between sensors. Traditional methods typically average or weightedly fuse the measured values ​​of multiple sensors. This approach only reduces random noise and cannot eliminate systematic uniformity deviations. To address this issue, this invention acquires multiple pressure measurements from multiple sensors under the same operating conditions, constructs a deviation matrix, and performs eigenvalue decomposition on the deviation covariance matrix. The eigenvalues ​​reflect the variance of the deviation in each direction. The direction of the eigenvector corresponding to the largest eigenvalue is the direction of the largest variance among the sensors, i.e., the direction of the systematic uniformity deviation. The projection of the deviation vector in this direction is taken as the uniformity deviation component and removed; the remaining deviation is the differential deviation component, which is used as the inter-sensor deviation compensation component. This approach eliminates the systematic bias caused by common-mode interference from multiple sensors, compensating only for the unique individual differences of each sensor.

[0063] The fluid adiabatic effect compensation component is added to the sensor inter-sensor deviation compensation component to obtain the flow-related error compensation amount. This compensation amount simultaneously considers the transient temperature effects caused by the fluid thermodynamic process and the systematic deviation effects between sensors. It will then be superimposed on the calibration parameters, enabling the calibration results to adapt to various error sources caused by flow state changes. In practical applications, when the valve is in a steady-state condition without rapid action, the estimated adiabatic temperature rise approaches zero, and the fluid adiabatic effect compensation component can be ignored. When the valve actuates rapidly, the system automatically calculates the compensation amount to correct for transient temperature effects. When multiple sensors are simultaneously subjected to external vibration interference, the sensor inter-sensor deviation compensation component can eliminate common-mode interference, ensuring the calibration results are unaffected. This compensation method, which comprehensively considers the fluid thermodynamic characteristics and sensor array characteristics, enables the system to maintain the accuracy of pressure measurement under various complex operating conditions.

[0064] In one embodiment, the step of obtaining the inter-sensor deviation compensation component based on the multi-dimensional operating condition parameters includes: S4051. Acquire multiple measurement values ​​from multiple sensors under the same operating conditions; S4052. Obtain the deviation matrix between the measurement values ​​based on the plurality of measurement values; S4053. Obtain the mean deviation vector and the covariance deviation matrix based on the deviation matrix; S4054. Perform eigenvalue decomposition on the deviation covariance matrix to obtain eigenvalues ​​and eigenvectors; S4055. Project the deviation along the eigenvector direction corresponding to the largest eigenvalue to obtain the consistent deviation component; S4056. Obtain the difference deviation component from the remaining deviation after removing the consistency deviation component; S4057. Obtain the sensor-to-sensor deviation compensation component from the difference deviation component.

[0065] As described in steps S4051-S4057 above, the present invention acquires multiple measurement values ​​from multiple sensors under the same operating conditions, obtains a deviation matrix between the measurement values ​​based on the multiple measurement values, obtains a mean deviation vector and a deviation covariance matrix based on the deviation matrix, performs eigenvalue decomposition on the deviation covariance matrix to obtain eigenvalues ​​and eigenvectors, projects the deviation in the direction of the eigenvector corresponding to the largest eigenvalue to obtain a consistent deviation component, obtains a differential deviation component from the remaining deviation after removing the consistent deviation component, and uses the differential deviation component as a deviation compensation component between sensors.

[0066] In pressure measurement systems for high-pressure regulating valves, multiple pressure sensors are typically installed to improve measurement reliability and provide redundancy. These sensors are susceptible to various common sources of interference during operation, such as common-mode vibrations in the piping system, common drift in ambient temperature, and common deformation of the mounting base. When these common-mode interferences occur, the measurements from all sensors shift in the same direction, resulting in a consistent bias. Traditional multi-sensor fusion methods usually employ arithmetic averaging or weighted averaging. While these methods can reduce the impact of random noise, they are ineffective against systematic consistent biases because when all sensors drift in the same direction, the average value also drifts, making it impossible to identify and eliminate this bias.

[0067] This invention introduces a deviation matrix analysis method. First, measurements from multiple sensors are collected under the same operating conditions. Since the conditions are identical, theoretically, the measurements from each sensor should be consistent; the actual differences represent the deviations between the sensors. After constructing the deviation matrix, covariance analysis is performed. The covariance matrix reflects the correlation between the deviations of each sensor. When all sensors are subjected to common interference, the deviation vector has the largest variance in a certain direction. This direction represents the main trend of deviation changes for each sensor, which is the direction of common-mode interference.

[0068] By performing eigenvalue decomposition on the deviation covariance matrix, the eigenvalues ​​reflect the magnitude of the variance of the deviation in each direction. The eigenvector corresponding to the largest eigenvalue points in the direction of the largest deviation variance, i.e., the direction of common-mode interference. Projecting the deviation vector onto this direction yields the uniform deviation component, which is the systematic deviation commonly experienced by all sensors. Subtracting this component from the original deviation leaves the sensor-specific differential deviation component, which originates from factors such as aging drift, installation differences, and individual sensitivity differences within each sensor.

[0069] The differential deviation component is used as the inter-sensor deviation compensation component for subsequent calibration calculations. The advantage of this approach is that when external common-mode interference occurs, the uniform deviation component is identified and removed, preventing it from entering the compensation stage; while the individual differences of each sensor are retained as the object requiring compensation. In practical applications, when a pipeline system experiences severe vibration, all pressure sensors are simultaneously affected, causing a shared shift in measured values. Using the method described in this embodiment, the system can identify and eliminate this uniform deviation component, compensating only for the inherent individual differences between sensors, thus avoiding the erroneous compensation behavior of misjudging common-mode interference as sensor deviation. When a sensor gradually drifts due to aging, its deviation from other sensors gradually increases; this deviation is retained in the differential deviation component, which the system can effectively compensate for.

[0070] In one embodiment, the step of adaptively updating the basic calibration parameters based on the calibration deviation between the calibrated pressure value and the pressure reference standard includes: S601. Obtain the residual based on the calibrated pressure value and the pressure reference standard; S602. Determine whether the absolute value of the residual is greater than a preset update threshold; If the absolute value of the residual is greater than the preset update threshold, then the corresponding element in the flow-coefficient mapping relationship is updated according to the residual and the preset learning rate; If the absolute value of the residual is not greater than the preset update threshold, then obtain the consecutive non-updated count and determine whether the consecutive non-updated count is greater than the preset forgetting threshold. If the consecutive unupdated count is greater than the preset forgetting threshold, then the corresponding element in the flow-coefficient mapping relationship is decayed and updated according to the preset forgetting factor; In this embodiment, the pressure reference standard does not rely on an external high-precision standard source, but is constructed in the following way: When there are multiple pressure sensors in the system, the pressure values ​​of each calibrated sensor are weighted and fused to obtain a fused estimate as a reference benchmark. When the valve is in a preset steady-state condition (both the differential pressure gradient and the rate of temperature change are below the threshold), the current calibrated pressure value is recorded as the historical steady-state benchmark. In the adaptive update phase, the weighted deviation between the fused estimate and the historical steady-state benchmark is taken as the calibration deviation.

[0071] In this way, the system can still achieve adaptive evolution of basic calibration parameters without relying on external reference signals, thus avoiding the problem of errors circulating and failing to converge between sensors.

[0072] As described in steps S601-S602 above, the present invention determines whether the absolute value of the residual between the calibrated pressure value and the pressure reference standard is greater than a preset update threshold. If it is greater, the corresponding element in the flow-coefficient mapping relationship is updated according to the residual and the preset learning rate. If it is not greater, the continuous unupdated count is obtained and it is determined whether it is greater than a preset forgetting threshold. If it is greater, the corresponding element in the flow-coefficient mapping relationship is updated by decay according to the preset forgetting factor.

[0073] During the long-term operation of a high-pressure regulating valve, its characteristics will slowly change due to factors such as valve core wear, spring fatigue, and seal aging. Simultaneously, the operating conditions will change with process requirements. If the calibration parameters remain unchanged over a long period, the originally accurate calibration parameters will gradually deviate as the valve characteristics drift. Furthermore, if updates are made each time based on a single calibration deviation (where the multi-sensor fusion estimate is obtained by weighted summation of measurements from multiple pressure sensors, and the historical steady-state reference value is taken from the pressure statistical characteristics of the valve body under preset steady-state conditions), the parameters will be affected by measurement noise and random disturbances, leading to frequent parameter fluctuations. Therefore, an adaptive update mechanism is needed that can both track long-term changes and suppress short-term disturbances.

[0074] This invention determines whether calibration parameters need updating based on residuals. The residual is the difference between the calibrated pressure value and the pressure reference standard, reflecting the accuracy of the current calibration parameters. When the absolute value of the residual exceeds a preset update threshold, it indicates a significant deviation between the current calibration parameters and the actual pressure reference. This deviation exceeds the normal fluctuation range and requires immediate correction. At this point, the corresponding elements in the flow regime-coefficient mapping relationship are updated based on the residual and a preset learning rate. The learning rate determines the step size of each correction; a larger residual results in a larger correction, enabling the calibration parameters to quickly converge to the value required for the current state. After the update, the count of continuously unupdated parameters is reset to zero, indicating that the calibration parameters for that flow regime have just been activated and updated.

[0075] When the absolute value of the residual is not greater than the preset update threshold, it indicates that the deviation between the current calibration parameter and the pressure reference is within the allowable range and does not require immediate updating. At this time, the consecutive unupdated count is incremented by 1. This counter records the number of consecutive times the calibration parameter under the current flow condition has not been updated. When the consecutive unupdated count exceeds the preset forgetting threshold, it indicates that the calibration parameter under this flow condition has not been used or verified for a long time and may be outdated. For example, a certain flow condition may have rarely occurred in past operations, and its corresponding calibration parameter may still be a historical value. However, after a long period of valve degradation, this historical value may no longer be applicable. In this case, a decay update is performed according to the preset forgetting factor, causing the reference coefficient to gradually regress to the default calibration coefficient. The forgetting factor determines the decay rate; the closer the value is to 1, the slower the decay, and the closer the value is to 0, the faster the decay. After the decay update, the counter is also reset to zero to ensure that the parameter does not decay indefinitely.

[0076] When the calibration deviation is large, the system can respond quickly, using incremental updates to rapidly converge the parameters to accurate values. When the calibration deviation is small, the system maintains parameter stability, avoiding frequent adjustments that introduce additional noise. When a certain flow condition has not appeared for a long time, the system automatically decays its corresponding calibration parameters, preventing outdated parameters from being misused under new operating conditions. In practical applications, when a pressure regulating valve is newly put into use, the calibration parameters corresponding to various flow conditions are initialized to default values. As deviations occur during operation and trigger updates, the parameters gradually adapt to the actual characteristics of the valve. When the valve wears out after operating for a period of time, the calibration parameters automatically track this change. When a certain operating condition reappears after a long period of absence, the parameters have decayed back to near their default values, avoiding the use of outdated historical parameters.

[0077] This invention also discloses an adaptive pressure calibration system for a high-pressure regulating valve, comprising: The data acquisition module is used to acquire multi-dimensional operating condition parameters and pressure measurement values ​​of the valve body under operating conditions, and to acquire pressure pulsation characteristics based on the multi-dimensional operating condition parameters. The multi-dimensional operating condition parameters include the valve inlet pressure sequence, valve outlet pressure sequence, valve body temperature sequence, valve body vibration signal sequence, and valve core position signal sequence. The parameter acquisition module is used to obtain the working condition classification result based on the pressure pulsation characteristics, and to obtain the basic calibration parameters based on the working condition classification result and the pressure measurement value. The basic calibration module is used to obtain the dynamic response compensation amount based on the pressure measurement value and multi-dimensional operating condition parameters, and to obtain the benchmark calibration value based on the dynamic response compensation amount and basic calibration parameters. The benchmark acquisition module is used to obtain the flow-related error compensation amount based on the multi-dimensional working condition parameters and working condition classification results, and to obtain the comprehensive calibration coefficient based on the benchmark calibration value and the flow-related error compensation amount. The calibration output module is used to calibrate and correct the pressure measurement value according to the comprehensive calibration coefficient, and obtain the calibrated pressure value. The update module is used to adaptively update the basic calibration parameters based on the calibration deviation between the calibrated pressure value and the pressure reference standard.

[0078] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in this application and in the embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual-speed SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0079] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0080] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent results or equivalent process transformations made based on the content of the present invention specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for adaptive pressure calibration of a high-pressure regulating valve under specific operating conditions, characterized in that, include: The system acquires multi-dimensional operating parameters and pressure measurement values ​​of the valve body under operating conditions, and obtains pressure pulsation characteristics based on the multi-dimensional operating parameters. The multi-dimensional operating parameters include the valve inlet pressure sequence, valve outlet pressure sequence, valve body temperature sequence, valve body vibration signal sequence, and valve core position signal sequence. Based on the pressure pulsation characteristics, the working condition classification result is obtained, the preset flow state-coefficient mapping relationship is obtained, and the benchmark coefficient under the corresponding flow state is obtained by querying the flow state-coefficient mapping relationship based on the working condition classification result. The sensor signal quality assessment value is obtained based on the pressure measurement value and the reference coefficient; The sensor confidence weight is obtained based on the sensor signal quality assessment value; The basic calibration parameters are obtained based on the benchmark coefficients and sensor confidence weights; The dynamic response compensation amount is obtained based on the pressure measurement value and multi-dimensional operating condition parameters, and the benchmark calibration value is obtained based on the dynamic response compensation amount and the basic calibration parameters. Obtain the specific heat ratio and current pressure measurement of the fluid; The pressure change rate is obtained based on the pre-valve pressure sequence and the post-valve pressure sequence. The estimated adiabatic temperature rise is obtained based on the pressure change rate and specific heat ratio. The temperature sensitivity coefficient of the sensor is obtained, and the fluid adiabatic effect compensation component is obtained based on the estimated adiabatic temperature rise and the temperature sensitivity coefficient. The sensor deviation compensation component is obtained based on the multi-dimensional operating parameters. The flow-related error compensation amount is obtained based on the fluid adiabatic effect compensation component and the sensor deviation compensation component, and the comprehensive calibration coefficient is obtained based on the benchmark calibration value and the flow-related error compensation amount. The pressure measurement value is calibrated and corrected according to the comprehensive calibration coefficient to obtain the calibrated pressure value; The basic calibration parameters are adaptively updated based on the calibration deviation between the calibrated pressure value and the pressure reference standard.

2. The adaptive pressure calibration method for a high-pressure regulating valve according to claim 1, characterized in that, The step of obtaining pressure pulsation characteristics based on the multidimensional operating condition parameters includes: The differential pressure sequence is obtained based on the inlet and outlet pressure sequences, and the differential pressure gradient characteristics are obtained based on the differential pressure sequence. The temperature change rate characteristics are obtained based on the valve body temperature sequence; Vibration energy characteristics are obtained based on the valve body vibration signal sequence; Pressure pulsation characteristics are obtained based on the pre-valve pressure sequence and post-valve pressure sequence.

3. The adaptive pressure calibration method for a high-pressure regulating valve according to claim 2, characterized in that, The step of obtaining pressure pulsation characteristics based on the pre-valve pressure sequence and the post-valve pressure sequence includes: The pressure pulsation components before and after the valve are obtained based on the pressure sequence before and after the valve, respectively. The pressure pulsation sequence is obtained based on the pre-valve pressure pulsation component and the post-valve pressure pulsation component. The pressure pulsation sequence is decomposed into wavelet multi-scale values ​​to obtain wavelet coefficient energies at multiple scales. The scaling exponent and multifractal spectrum width are obtained based on the wavelet coefficient energy at the multiple scales. The pulsation frequency characteristics are obtained based on the scaling index, and the pressure pulsation characteristics are obtained based on the multifractal spectrum width.

4. The adaptive pressure calibration method for a high-pressure regulating valve according to claim 1, characterized in that, The step of obtaining the operating condition classification result based on the pressure pulsation characteristics includes: Obtain the pulsation feature vector based on the pressure pulsation characteristics; Obtain template feature vectors for multiple preset flow state categories, wherein the flow state categories include laminar steady-state category, transitional flow state category, and turbulent flow state category; A distance set is obtained based on the distance between the pulsating feature vector and each of the template feature vectors; Obtain the working condition classification result by identifying the working condition flow state category corresponding to the minimum distance in the distance set.

5. The adaptive pressure calibration method for a high-pressure regulating valve according to claim 1, characterized in that, The step of obtaining the dynamic response compensation amount based on the pressure measurement value and multi-dimensional operating condition parameters includes: The sensor output sequence is obtained based on the pressure measurement value; Obtain the sensor's natural frequency and damping ratio, and construct the sensor transfer function based on the natural frequency and damping ratio; The inverse filter is obtained based on the sensor transfer function; A preliminary deconvolution sequence is obtained based on the inverse filter and the sensor output sequence; The regularization parameter is obtained from the initial deconvolution sequence, and the initial deconvolution sequence is regularized according to the regularization parameter to obtain the pressure estimation sequence after deconvolution. The dynamic response compensation amount is obtained based on the deconvolutioned pressure estimation sequence and the sensor output sequence.

6. The adaptive pressure calibration method for a high-pressure regulating valve according to claim 1, characterized in that, The step of obtaining the inter-sensor deviation compensation component based on the multi-dimensional operating condition parameters includes: Acquire multiple measurement values ​​from multiple sensors under the same operating conditions; Obtain the deviation matrix between the measurement values ​​based on the multiple measurement values; The mean deviation vector and the covariance deviation matrix are obtained from the deviation matrix. The deviation covariance matrix is ​​decomposed into eigenvalues ​​to obtain eigenvalues ​​and eigenvectors. The consistent deviation component is obtained by projecting the deviation along the direction of the eigenvector corresponding to the largest eigenvalue. The difference deviation component is obtained by removing the uniform deviation component from the remaining deviation. The difference deviation component is used to obtain the inter-sensor deviation compensation component.

7. The adaptive pressure calibration method for a high-pressure regulating valve according to claim 1, characterized in that, The step of adaptively updating the basic calibration parameters based on the calibration deviation between the calibrated pressure value and the pressure reference standard includes: The residual is obtained based on the calibrated pressure value and the pressure reference standard; Determine whether the absolute value of the residual is greater than a preset update threshold; If the absolute value of the residual is greater than the preset update threshold, then the corresponding element in the flow-coefficient mapping relationship is updated according to the residual and the preset learning rate; If the absolute value of the residual is not greater than the preset update threshold, then obtain the consecutive non-updated count and determine whether the consecutive non-updated count is greater than the preset forgetting threshold. If the number of consecutive unupdated values ​​is greater than the preset forgetting threshold, then the corresponding elements in the flow-coefficient mapping relationship are decayed and updated according to the preset forgetting factor.

8. A condition-adaptive pressure calibration system for a high-pressure regulating valve, characterized in that, include: The data acquisition module is used to acquire multi-dimensional operating condition parameters and pressure measurement values ​​of the valve body under operating conditions, and to acquire pressure pulsation characteristics based on the multi-dimensional operating condition parameters. The multi-dimensional operating condition parameters include the valve inlet pressure sequence, valve outlet pressure sequence, valve body temperature sequence, valve body vibration signal sequence, and valve core position signal sequence. The parameter acquisition module is used to obtain the working condition classification result based on the pressure pulsation characteristics, obtain the preset flow state-coefficient mapping relationship, and query the flow state-coefficient mapping relationship based on the working condition classification result to obtain the reference coefficient under the corresponding flow state. The sensor signal quality assessment value is obtained based on the pressure measurement value and the reference coefficient; The sensor confidence weight is obtained based on the sensor signal quality assessment value; The basic calibration parameters are obtained based on the benchmark coefficients and sensor confidence weights; The basic calibration module is used to obtain the dynamic response compensation amount based on the pressure measurement value and multi-dimensional operating condition parameters, and to obtain the benchmark calibration value based on the dynamic response compensation amount and basic calibration parameters. The reference acquisition module is used to acquire the specific heat ratio and current pressure measurement value of the fluid; The pressure change rate is obtained based on the pre-valve pressure sequence and the post-valve pressure sequence. The estimated adiabatic temperature rise is obtained based on the pressure change rate and specific heat ratio. The temperature sensitivity coefficient of the sensor is obtained, and the fluid adiabatic effect compensation component is obtained based on the estimated adiabatic temperature rise and the temperature sensitivity coefficient. The sensor deviation compensation component is obtained based on the multi-dimensional operating parameters. The flow-related error compensation amount is obtained based on the fluid adiabatic effect compensation component and the sensor deviation compensation component, and the comprehensive calibration coefficient is obtained based on the benchmark calibration value and the flow-related error compensation amount. The calibration output module is used to calibrate and correct the pressure measurement value according to the comprehensive calibration coefficient, and obtain the calibrated pressure value. The update module is used to adaptively update the basic calibration parameters based on the calibration deviation between the calibrated pressure value and the pressure reference standard.

Images