Natural gas pipeline coaxial on-off valve capable of carrying flow prediction module

Abstract

The invention relates to the technical field of structure testing, in particular to a natural gas pipeline coaxial on-off valve capable of carrying a flow prediction module, and the flow prediction module easily forms a backflow area guided by inertial airflow based on geometric changes of airflow in a pipeline and a corner of a valve cavity, and the backflow area is guided by the flow prediction module along with mutual reaction of low-pressure vortex and mainstream airflow. According to the characteristic that the mass flow of the pipeline is dynamically lagged or the vibration is distorted due to the fact that the mass flow of the pipeline is vibrated, the lagging effect and vibration distortion characteristics of fluid in the pipeline are dynamically analyzed through an empirical mode decomposition and fast Fourier transform algorithm, and a flow fluctuation coefficient, a lagging characteristic coefficient and a vibration abnormal coefficient are extracted; according to the method, a time sequence dynamic lag curve and a time sequence vibration distortion curve are constructed, a dynamic correction model is established based on the correlation of the time sequence dynamic lag curve and the time sequence vibration distortion curve, and nonlinear distortion of a flow-valve opening transfer function is compensated in real time, so that a more accurate predicted flow value after correction is obtained, and the accuracy of valve opening adjustment according to the predicted flow value is higher.

Description

Technical Field

[0001] The present invention relates to the technical field of structural testing, and in particular to a coaxial on-off valve for a natural gas pipeline capable of carrying a flow prediction module. Background Art

[0002] Traditional natural gas pipeline valve control relies on set pressure and flow data, typically based on experience or preset conditions. However, with the development of intelligent technology, the addition of flow prediction modules can predict future flow trends in the pipeline system by collecting and analyzing flow, pressure, temperature, and other information in real time. Compared with traditional natural gas pipeline valve control methods, this method has higher real-time performance.

[0003] Currently, coaxial on-off valves in natural gas pipelines generally adopt a split-type start-up actuator and valve body assembly design. When a flow prediction module is introduced, the interface between the split valve body and the actuator has geometric mutations (such as right-angle edges and reduced diameter sections). When the valve is opened and closed quickly, the fluid cannot instantly follow the valve movement due to inertia, resulting in local backflow or flow separation, forming a backflow zone. The interaction between the backflow zone and the mainstream will cause slow flow deviation and micro-oscillation, which in turn triggers nonlinear distortion of the flow-valve opening transfer function based on the prediction model, resulting in low accuracy of the predicted flow value, making the accuracy of valve opening adjustment based on the predicted flow value poor. Summary of the Invention

[0004] In order to solve the technical problem that the traditional split valve has low accuracy in the predicted flow value obtained based on the prediction model due to the oscillation of the recirculation zone caused by the geometric mutation, the purpose of this application is to provide a natural gas pipeline coaxial on-off valve that can be equipped with a flow prediction module. The technical solution adopted is as follows:

[0005] The present application provides a natural gas pipeline coaxial on-off valve capable of being equipped with a flow prediction module, including a flow prediction module; the flow prediction module includes:

[0006] The data acquisition and preprocessing submodule is used to collect pipeline flow signals and pipeline vibration signals of the natural gas pipeline; perform empirical mode decomposition on the pipeline flow signals to obtain corresponding IMF component signals;

[0007] The first determination submodule is configured to determine, in the IMF component signal, a hysteresis characteristic coefficient at each sampling moment based on the intensity of signal fluctuations and the phase deviation of the envelope in a historical time period at each sampling moment; and determine a time series dynamic hysteresis curve based on the hysteresis characteristic coefficients at each sampling moment;

[0008] The second determination submodule is configured to determine, on the pipeline vibration signal, a vibration anomaly coefficient value at each sampling moment based on a local change of the signal within a time sequence neighborhood at each sampling moment; and determine a time sequence vibration distortion curve based on the vibration anomaly coefficient values at each sampling moment;

[0009] The valve opening adjustment submodule is used to determine the model correction coefficient based on the hysteresis characteristic coefficient, the vibration abnormality coefficient value and the curve correlation between the corresponding time-series dynamic hysteresis curve and the time-series vibration distortion curve at the current moment; correct the real-time pipeline flow data according to the model correction coefficient to determine the predicted flow value; and adjust the valve opening according to the predicted flow value.

[0010] Furthermore, the process of obtaining the hysteresis characteristic coefficient includes:

[0011] The local signal segment corresponding to the historical time period at each sampling moment in the IMF component signal is used as the historical component signal segment at each sampling moment;

[0012] Determine the flow fluctuation coefficient at each sampling moment according to the slope deviation on both sides of the extreme value point in the historical component signal and the fluctuation of the signal difference value;

[0013] Obtaining an upper envelope and a lower envelope of the historical component signal segment; using a phase difference between frequency domain data after fast Fourier transform of the upper envelope and frequency domain data after fast Fourier transform of the lower envelope as a flow hysteresis coefficient at a sampling moment corresponding to the historical component signal segment;

[0014] The hysteresis characteristic coefficient at each sampling moment is determined according to the product of the flow hysteresis coefficient and the flow fluctuation coefficient.

[0015] Furthermore, the process of obtaining the flow fluctuation coefficient includes:

[0016] In the historical component signal segment, the difference between the tangent slope at the first sampling moment before each extreme value point and the tangent slope at the first sampling moment after each extreme value point is used as the slope change value for determining each extreme value point; and the degree of extreme value change at each sampling moment is determined based on the average of the slope change values of all extreme value points in the historical component signal segment;

[0017] In the historical component signal segment, the difference between the signal value at each moment and the signal value at the previous moment is used as the instantaneous signal difference at each moment; and the degree of timing fluctuation at the sampling moment corresponding to the historical component signal segment is determined based on the mean value of the instantaneous signal differences at all moments in the historical component signal segment;

[0018] The flow fluctuation coefficient at each sampling moment is determined according to the product of the extreme value change degree and the time series fluctuation degree.

[0019] Furthermore, the process of obtaining the time series dynamic hysteresis curve includes:

[0020] The hysteresis characteristic coefficients of all sampling moments are arranged in time sequence and then curve fitting is performed to obtain the time series dynamic hysteresis curve.

[0021] Furthermore, the process of obtaining the vibration abnormality coefficient value includes:

[0022] In the pipeline vibration signal, the pipeline vibration signal is divided into at least two initial window sub-signal segments with the current time as the starting point and a preset number of data points as the step size; and the corresponding data fluctuation coefficient is determined according to the standard deviation of all signal values in each initial window sub-signal segment;

[0023] Merge windows based on the similarity of data fluctuation coefficients between adjacent initial window sub-signal segments to obtain all merged window sub-signal segments;

[0024] Determine the corresponding combined fluctuation coefficient based on the average of the data fluctuation coefficients of all initial window sub-signal segments in each combined window sub-signal segment; normalize the difference between the combined fluctuation coefficient of each combined window sub-signal segment and the combined fluctuation coefficient of the previous combined window sub-signal segment to determine the variation anomaly coefficient of each combined window sub-signal segment; and determine the corresponding overall anomaly coefficient based on the product of the normalized value of the combined fluctuation coefficient and the variation anomaly coefficient;

[0025] The overall abnormal coefficient of the merged window sub-signal segment at each sampling moment is taken as the corresponding vibration abnormal coefficient value.

[0026] Furthermore, the process of acquiring the merged window sub-signal segments includes:

[0027] In chronological order, the adjacent fluctuation difference that characterizes the difference between the data fluctuation coefficient of each initial window sub-signal segment and the data fluctuation coefficient of the previous initial window sub-signal segment is calculated; two initial window sub-signal segments whose normalized value of the adjacent fluctuation difference is less than a preset fluctuation threshold are merged; after traversing all initial window sub-segments, all merged window sub-signal segments are obtained.

[0028] Furthermore, the process of obtaining the time series vibration distortion curve includes:

[0029] The vibration anomaly coefficient values at all sampling moments are arranged in chronological order and then curve fitting is performed to determine the time series vibration distortion curve.

[0030] Furthermore, the process of obtaining the model correction coefficient includes:

[0031] Normalize the product of the hysteresis characteristic coefficient and the vibration abnormality coefficient value at the current moment to determine the reference correction coefficient at the current moment;

[0032] Calculating the Pearson correlation coefficient between the time series dynamic hysteresis curve and the time series vibration distortion curve;

[0033] The model correction coefficient at the current moment is determined according to the positive correlation mapping value of the sum value between the reference correction coefficient and the Pearson correlation coefficient.

[0034] Furthermore, the process of obtaining the predicted flow value includes:

[0035] The predicted flow value at the current moment is determined according to the product between the predicted flow value and the pipeline flow data at the current moment.

[0036] Furthermore, the process of adjusting the valve opening according to the predicted flow value includes:

[0037] Based on the PID controller, the valve opening is adjusted through the proportional term, the integral term and the differential term according to the difference between the predicted flow value and the prior target flow value.

[0038] This application has the following beneficial effects:

[0039] The flow prediction module of the present application is based on the fact that the geometric changes of the air flow in the pipeline and the corner of the valve cavity easily form a backflow zone dominated by inertial airflow, accompanied by the mutual reaction of low-pressure vortex and mainstream airflow, which leads to dynamic lag or vibration distortion of the pipeline mass flow. Through empirical mode decomposition and fast Fourier transform algorithm, the lag effect and vibration distortion characteristics of the fluid in the pipeline are dynamically analyzed, the flow fluctuation coefficient, lag characteristic coefficient and vibration abnormality coefficient are extracted, and the time-series dynamic lag curve and time-series vibration distortion curve are constructed. Based on the correlation between the two, a dynamic correction model is established to compensate for the nonlinear distortion of the flow-valve opening transfer function in real time, so as to obtain a more accurate predicted flow value after correction, so that the accuracy of valve opening adjustment based on the predicted flow value is higher. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0041] Figure 1A schematic diagram of the overall structure of a coaxial on-off valve for a natural gas pipeline equipped with a flow prediction module provided by one embodiment of the present invention;

[0042] Figure 2 A structural diagram of a traffic prediction module provided by one embodiment of the present invention;

[0043] Figure 1 The numbers in the figure are: 1-valve seat; 2-valve seat sealing ring; 3-valve cover; 4-valve position high temperature resistant feedback proximity switch; 5-flow prediction module. DETAILED DESCRIPTION

[0044] In order to further illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the following is a detailed description of the specific implementation method, structure, features and effects of a natural gas pipeline coaxial on-off valve that can be equipped with a flow prediction module proposed by the present invention, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment, and the specific features, structures or characteristics in one or more embodiments may be combined in any suitable form. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as implying or suggesting relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features.

[0045] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0046] The following describes in detail a specific solution of a coaxial on-off valve for a natural gas pipeline equipped with a flow prediction module provided by the present invention with reference to the accompanying drawings.

[0047] This application provides a natural gas pipeline coaxial on-off valve that can be equipped with a flow prediction module. Figure 1 , which shows a schematic diagram of the overall structure of a natural gas pipeline coaxial on-off valve equipped with a flow prediction module provided by an embodiment of the present invention, including: a valve seat 1, a valve seat sealing ring 2, a valve cover 3, a valve position high-temperature resistant feedback proximity switch 4 and a flow prediction module 5. For the flow prediction module 5, please refer to Figure 2 , which shows a structural diagram of a flow prediction module provided by an embodiment of the present invention, including: a data acquisition and preprocessing submodule 201, a first determination submodule 202, a second determination submodule 203 and a valve opening adjustment submodule 204.

[0048] The data acquisition and preprocessing submodule 201 is used to acquire pipeline flow signals and pipeline vibration signals of the natural gas pipeline; perform empirical mode decomposition on the pipeline flow signals to obtain corresponding IMF component signals.

[0049] In a specific implementation of an embodiment of the present invention, a Coriolis mass flowmeter is used to collect pipeline flow data for monitoring the time series changes of the flow in the pipeline, and a three-axis MEMS accelerometer is used to collect pipeline vibration data for monitoring the time series changes of the vibration in the pipeline; the collected pipeline flow data are then arranged in time sequence and curve fitting is performed to obtain a pipeline flow signal, and the collected pipeline vibration data are arranged in time sequence and curve fitting is performed to obtain a pipeline vibration signal; it should be noted that the pipeline flow signal and the pipeline vibration signal are both signals of the same pipeline point, and curve fitting is a technical means well known to those skilled in the art, which will not be further limited or elaborated here. In a specific implementation of an embodiment of the present invention, the time range of the pipeline flow signal and the pipeline vibration signal are both set to within 30 minutes before the current moment, and the sampling frequency is set to collect once per second. The implementer can adjust it according to the specific implementation environment, and will not be further elaborated here.

[0050] By analyzing the intrinsic mode function (IMF) component signal after performing empirical mode decomposition on the pipeline flow signal, it is possible to better cope with the non-stationary, nonlinear, and multi-scale characteristics of the signal, thereby more accurately extracting features. In a specific implementation of an embodiment of the present invention, the IMF component signal analyzed is the IMF1 signal, which can be adjusted by the implementer according to the specific implementation environment. It should be noted that empirical mode decomposition is a technical means well known to those skilled in the art and will not be further defined or elaborated on here.

[0051] The first determination submodule 202 is used to determine the hysteresis characteristic coefficient of each sampling moment in the IMF component signal based on the signal fluctuation intensity and the phase deviation of the envelope in the historical time period of each sampling moment; and determine the time series dynamic hysteresis curve based on the hysteresis characteristic coefficient of each sampling moment.

[0052] When the valve opens and closes quickly, the geometric changes in the airflow in the pipeline and the corners of the valve cavity easily form a recirculation zone dominated by inertial airflow, accompanied by the mutual reaction of low-pressure vortices and mainstream airflow, resulting in dynamic lag or flow prediction deviation in the pipeline mass flow. The generation of the recirculation zone in the pipeline is due to frequent and violent valve transient operations such as valve opening and closing. This phenomenon is the result of the dynamic flow regulation requirements of the natural gas transmission and distribution system, such as peak and valley load fluctuations and multi-gas source switching. For flow data, the more drastic the flow change in a short period of time, the more likely it is to correspond to frequent valve transient operations of the pipeline on-off valve. In addition, frequent valve transient operations will trigger a large number of recirculation zones and fluid vortices in the pipeline, causing the mainstream fluid to be violently disturbed in the reverse direction, resulting in flow lag. Especially during valve operation, pipeline flow changes present irregular and complex transient characteristics, which further aggravates the flow hysteresis phenomenon. Therefore, when the valve is frequently adjusted, the flow prediction data often deviates as a whole, resulting in an inaccurate reflection of flow changes, which causes system response delays or improper switch adjustments. Therefore, the hysteresis characteristics are further characterized based on the flow fluctuations and the phase deviation of the envelope line that characterizes the overall deviation of the flow data.

[0053] Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the hysteresis characteristic coefficient includes:

[0054] The local signal segment corresponding to the historical time period of each sampling moment in the IMF component signal is used as the historical component signal segment of each sampling moment; in a specific implementation method of an embodiment of the present invention, the historical time period of each sampling moment is set to include each sampling moment and the time period corresponding to the previous 135 seconds, which can be adjusted according to the specific implementation environment; here, the lag characteristics are analyzed by the changes in the IMF component signal of the historical data corresponding to the sampling moment, which can prevent the influence of the observation limitations of the instantaneous data and can analyze the subsequent real-time lag characteristic coefficients.

[0055] The flow fluctuation coefficient at each sampling moment is determined based on the slope deviation on both sides of the extreme point in the historical component signal and the fluctuation of the signal differential value. Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the flow fluctuation coefficient includes:

[0056] In the historical component signal segment, the difference between the tangent slope at the first sampling moment before each extreme point and the tangent slope at the first sampling moment after each extreme point is used as the slope change value for determining each extreme point; the degree of extreme value change at each sampling moment is determined based on the mean of the slope change values of all extreme points in the historical component signal segment; in the historical component signal segment, the difference between the signal value at each moment and the signal value at the previous moment is used as the instantaneous signal difference at each moment; the degree of time series fluctuation at the sampling moment corresponding to the historical component signal segment is determined based on the mean of the instantaneous signal differences at all moments in the historical component signal segment; the flow fluctuation coefficient at each sampling moment is determined based on the product of the degree of extreme value change and the degree of time series fluctuation.

[0057] For each extreme point, the greater the difference in the tangent slopes between the two adjacent sampling moments on both sides, the greater the change in the signal change trend at the corresponding extreme point position, and the more significant the signal fluctuation characteristics are. Therefore, the greater the degree of extreme value change, the stronger the signal fluctuation degree in the corresponding historical time period. In addition, the greater the instantaneous signal difference, the more drastic the signal value change at the corresponding moment, that is, the instantaneous signal difference characteristic characterizes the significant characteristics of the instantaneous signal change at each sampling moment. Therefore, the degree of temporal fluctuation obtained by combining the instantaneous signal difference of all moments as a whole is further combined to characterize the degree of signal fluctuation as a whole. The greater the corresponding degree of temporal fluctuation, the more significant the signal fluctuation characteristics are. Finally, the flow fluctuation coefficient is jointly characterized based on the degree of extreme value change and the degree of temporal fluctuation, so that when the flow fluctuation coefficient is larger, the temporal fluctuation of the flow is stronger, the more drastic the change is, and the more likely it is to correspond to the frequent transient operation of the pipeline on-off valve.

[0058] In a specific implementation of the embodiment of the present invention, the process of obtaining the flow fluctuation coefficient is expressed by the formula: Among them, R k is the flow fluctuation coefficient at the kth sampling moment; N k is the number of extreme points in the historical component signal segment at the kth sampling moment; F′ k,i is the slope of the tangent line at the first sampling moment before the i-th extreme point in the historical component signal segment at the k-th sampling moment; F″ k,i is the tangent slope of the first sampling moment after the i-th extreme point in the historical component signal segment at the k-th sampling moment; |F′ k,i -F″ k,i | is the slope change value of the i-th extreme point in the historical component signal segment at the k-th sampling moment; is the degree of extreme value change at the kth sampling moment; M kis the number of moments in the historical component signal segment at the kth sampling moment. The moment here corresponds to the sampling moment. The purpose of using the moment here to express it is to avoid confusion of nouns. L k,m is the signal value at the mth moment in the historical component signal segment at the kth sampling moment; L k,m-1 is the signal value at the m-1th moment in the historical component signal segment at the kth sampling moment; |L k,m -L k,m-1 | is the instantaneous signal difference at the mth moment in the historical component signal segment at the kth sampling moment; is the degree of timing fluctuation at the kth sampling moment; || is the absolute value sign; it should be noted that when the extreme point is located at the first sampling moment or the last sampling moment of the historical component signal segment, the analysis is performed in combination with the tangent slope of the corresponding previous sampling moment or next sampling moment in the pipeline flow signal. When the previous sampling moment or next sampling moment does not exist in the pipeline flow signal, the slope change value of the corresponding extreme point is set to 0.

[0059] Obtain the upper envelope and lower envelope of the historical component signal segment; use the phase difference between the frequency domain data after fast Fourier transform of the upper envelope and the frequency domain data after fast Fourier transform of the lower envelope as the flow lag coefficient at the sampling moment corresponding to the historical component signal segment. It should be noted that fast Fourier transform and phase difference are technical terms well known to those skilled in the art, and will not be further defined or elaborated here. Based on the nature of frequency domain data, the larger the phase difference, the more consistent it is with the overall offset of the flow data; therefore, the larger the flow lag coefficient, the more significant the flow lag phenomenon in the pipeline, the poor coupling between the fluid response and the pipeline operation, resulting in an increase in the flow prediction error, which in turn affects the response speed and accuracy of the valve control; and the larger the flow lag coefficient, the more significant the delay in pipeline flow regulation, and the need to optimize the flow prediction and regulation control strategy.

[0060] In natural gas pipelines, frequent and large fluctuations in flow within a short period of time can exacerbate airflow hysteresis. Therefore, the flow fluctuation coefficient and the flow hysteresis coefficient are combined to comprehensively characterize the hysteresis characteristic coefficient at each sampling moment. This allows the hysteresis characteristic coefficient to more accurately represent the flow hysteresis characteristics within the pipeline. This embodiment of the present invention determines the hysteresis characteristic coefficient at each sampling moment based on the product of the flow hysteresis coefficient and the flow fluctuation coefficient.

[0061] In a specific implementation of the embodiment of the present invention, the process of obtaining the hysteresis characteristic coefficient is expressed by the formula: k =R k ×ΔX k Among them, W kis the hysteresis characteristic coefficient at the kth sampling moment; R k is the flow fluctuation coefficient at the kth sampling moment; ΔX k It is the phase difference between the frequency domain data after fast Fourier transform of the upper envelope of the historical component signal segment at the kth sampling moment and the frequency domain data after fast Fourier transform of the lower envelope, that is, the flow hysteresis coefficient at the kth sampling moment.

[0062] A time series dynamic hysteresis curve is further determined based on the flow hysteresis coefficient at each sampling moment to more intuitively observe the flow hysteresis and facilitate subsequent comparative analysis. Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the time series dynamic hysteresis curve includes: arranging the hysteresis characteristic coefficients of all sampling moments in chronological order and performing curve fitting to obtain the time series dynamic hysteresis curve.

[0063] The second determination submodule 203 is used to determine the vibration anomaly coefficient value at each sampling moment based on the local signal change within the time sequence neighborhood of each sampling moment on the pipeline vibration signal; and determine the time sequence vibration distortion curve based on the vibration anomaly coefficient value at each sampling moment.

[0064] In natural gas pipeline systems, the dynamic hysteresis curve of airflow primarily reflects overall pipeline flow variations, particularly the hysteresis effect of reverse flow on the mainstream caused by frequent valve operation. However, it is difficult to accurately measure local flow distortion within the pipeline. When the airflow velocity in the pipeline is high, especially when it exceeds a critical value, commonly known as the Reynolds number threshold, the flow transitions from laminar to turbulent. In natural gas pipeline systems, sudden changes in airflow velocity due to factors such as pipe wall fouling, pipe bends, or transient valve operation can cause localized changes in flow from laminar to turbulent. This transition triggers localized flow inhomogeneities, which in turn lead to unstable flow distribution within the pipeline. When the flow transitions from laminar to turbulent, localized flow inhomogeneities lead to unstable flow distribution within the pipeline, generating pressure waves that propagate along the pipeline. Pressure waves of varying intensity are applied to the pipeline wall at different locations, causing vibrations to vibrate, resulting in anomalies in the monitored vibration data. In other words, in the presence of turbulent or unstable flow, the vibration data fluctuates abnormally compared to the stable vibration data under normal conditions.

[0065] Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the vibration abnormality coefficient value includes:

[0066] In the pipeline vibration signal, with the current moment as the starting point and a preset number of data points as the step size, the pipeline vibration signal is divided into at least two initial window sub-signal segments; based on the standard deviation of all signal values in each initial window sub-signal segment, the corresponding data fluctuation coefficient is determined; in a specific implementation of an embodiment of the present invention, the preset number is set to 9, and when the number of data points does not meet the preset number at the end of traversal, the remaining data points are used as an initial window sub-signal segment for analysis, and the preset number can be adjusted according to the specific implementation environment, which will not be further elaborated here.

[0067] Window merging is performed based on similarity of data fluctuation coefficients between adjacent initial window sub-signal segments to obtain all merged window sub-signal segments. Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the merged window sub-signal segments includes:

[0068] In time sequence, the adjacent fluctuation difference that characterizes the difference between the data fluctuation coefficient of each initial window sub-signal segment and the data fluctuation coefficient of the previous initial window sub-signal segment is calculated; the two initial window sub-signal segments whose normalized values of the adjacent fluctuation difference are less than the preset fluctuation threshold are merged; after traversing all the initial window sub-segments, all the merged window sub-signal segments are obtained. In a specific implementation of an embodiment of the present invention, the preset fluctuation threshold is set to 0.3, and the normalization method of the adjacent fluctuation difference adopts linear normalization, which can be adjusted by itself. It should be noted that the difference between the two numerical values represents the absolute value of the difference between the two numerical values.

[0069] First, turbulence can cause unusual fluctuations in vibration data, while laminar flow is relatively stable. Therefore, for two adjacent initial window sub-signal segments, a small difference in adjacent fluctuations indicates that the two initial window sub-signal segments belong to the same data vibration mode: the vibration mode corresponding to laminar flow and the vibration mode corresponding to turbulence. Therefore, merging the window sub-signal segments allows for a more accurate determination of the influence of turbulence.

[0070] According to the mean of the data fluctuation coefficients of all initial window sub-signal segments in each merged window sub-signal segment, the corresponding merged fluctuation coefficient is determined; the difference between the merged fluctuation coefficient of each merged window sub-signal segment and the merged fluctuation coefficient of the previous merged window sub-signal segment is normalized to determine the variation anomaly coefficient of each merged window sub-signal segment; the corresponding overall anomaly coefficient is determined based on the product of the normalized value of the merged fluctuation coefficient and the variation anomaly coefficient. First, the presence of turbulence will cause abnormal fluctuations in vibration data, while the vibration data corresponding to laminar flow is relatively stable, so the data fluctuation coefficient under the laminar flow mode is smaller; therefore, for each merged window sub-signal segment, the larger the corresponding merged fluctuation coefficient, the more it indicates that the vibration data fluctuation characteristics of the corresponding merged window sub-signal segment are abnormal, and the higher the possibility that it corresponds to a turbulent vibration mode, that is, the greater the influence of turbulence. In addition, considering that the influence of turbulence usually appears independently, while laminar flow is the norm, the combined fluctuation coefficient of the combined window sub-signal segment corresponding to the laminar flow influence is usually much smaller than the combined fluctuation coefficient of the combined window sub-signal segment corresponding to the turbulent flow. That is, the greater the change in the corresponding combined fluctuation coefficient from the combined window sub-signal segment of laminar flow to the combined window sub-signal segment corresponding to turbulence, the greater the change in the corresponding combined fluctuation coefficient. Therefore, for each combined window sub-signal segment, the larger the change abnormality coefficient, the more consistent the corresponding combined window sub-signal segment is with the vibration mode of turbulence.

[0071] In some possible implementations of the present invention, the process of obtaining the overall anomaly coefficient includes: h =Norm(μ h -μ h-1 )×μ h Among them, Y h is the overall abnormal coefficient of the h-th merged window sub-signal segment; μ h is the combined fluctuation coefficient of the h-th combined window sub-signal segment; μ h-1 is the combined fluctuation coefficient of the h-1th combined window sub-signal segment; Norm() is a linear normalization function used to avoid the influence of negative numbers on the calculation; it should be noted that when the h-1th combined window sub-signal segment does not exist, the overall anomaly coefficient of the corresponding hth combined window sub-signal segment is set to 0.

[0072] Finally, based on the overall anomaly coefficient of each merged window sub-signal segment, the vibration anomaly coefficient value at each moment is replicated. Specifically, the overall anomaly coefficient of the merged window sub-signal segment at each sampling moment is used as the corresponding vibration anomaly coefficient value. A larger vibration anomaly coefficient value indicates that the local historical data at the corresponding sampling moment is more likely to correspond to flow distortion, that is, the presence of turbulence or flow instability.

[0073] A time-series vibration distortion curve is further determined based on the vibration anomaly coefficient values at each sampling moment, providing a more intuitive view of flow distortion and facilitating subsequent comparative analysis. Preferably, in some possible implementations of the present invention, the process of obtaining the time-series vibration distortion curve includes arranging the vibration anomaly coefficient values at all sampling moments in chronological order and performing curve fitting to obtain the time-series vibration distortion curve.

[0074] The valve opening adjustment submodule 204 is used to determine the model correction coefficient based on the hysteresis characteristic coefficient, the vibration abnormality coefficient value and the curve correlation between the corresponding time-series dynamic hysteresis curve and the time-series vibration distortion curve at the current moment; correct the real-time pipeline flow data according to the model correction coefficient to determine the predicted flow value; and adjust the valve opening according to the predicted flow value.

[0075] First, the larger the hysteresis characteristic coefficient at the corresponding sampling moment, the greater the delay in pipeline flow regulation and the greater the flow prediction error. The larger the vibration anomaly coefficient value, the greater the abnormal flow state at the corresponding sampling moment, which affects the accuracy of the prediction and the greater the corresponding prediction error. In addition, if the time series vibration distortion curve and the time series dynamic hysteresis curve are highly correlated, it means that the deviation between the pipeline flow prediction and the valve opening transmission is caused by the airflow hysteresis effect and the intensified vibration distortion caused by local turbulence, thereby aggravating the flow prediction deviation. Therefore, preferably, the process of obtaining the model correction coefficient includes:

[0076] Normalize the product of the current hysteresis characteristic coefficient and the vibration anomaly coefficient to determine the current reference correction coefficient. Calculate the Pearson correlation coefficient between the time-series dynamic hysteresis curve and the time-series vibration distortion curve. Determine the current model correction coefficient based on the positive correlation mapping between the sum of the reference correction coefficient and the Pearson correlation coefficient. Larger reference correction coefficients and larger Pearson correlation coefficients between the time-series dynamic hysteresis curve and the time-series vibration distortion curve indicate greater flow prediction deviations. Therefore, a larger model correction coefficient is required to compensate for flow prediction deviations caused by pipeline structure.

[0077] In some possible implementations of the present invention, the process of obtaining the model correction coefficient is expressed as follows: b =1+Norm((W b ×Y b ′ )+ρ); where L b is the model correction coefficient at the current moment b; W b is the hysteresis characteristic coefficient of the current moment b; Y b ′is the vibration anomaly coefficient value at the current moment b; ρ is the Pearson correlation coefficient between the time series dynamic hysteresis curve and the time series vibration distortion curve.

[0078] Since the flow prediction deviation caused by the pipeline structure is greater when the model correction coefficient is larger, the pipeline flow data at the current moment is further weighted by the model correction coefficient having a value greater than or equal to 1. Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the predicted flow value includes:

[0079] The predicted flow value at the current moment is determined based on the product between the predicted flow value and the pipeline flow data at the current moment. When the model correction coefficient is large, it will correspond to a larger flow prediction value, so that there will be a larger deviation from the set flow target as a whole, so that the valve opening can be adjusted more accurately based on the deviation combined with PID control. In another specific implementation of the embodiment of the present invention, the product between the model correction coefficient and the set flow target can also be used as the flow prediction value, which will not be further elaborated here. Finally, the process of adjusting the valve opening according to the predicted flow value includes: based on the PID controller, the valve opening is adjusted through the proportional term, the integral term and the differential term according to the difference between the predicted flow value and the prior target flow value. The proportional term adjusts the valve opening in real time according to the size of the deviation, the integral term accumulates past deviations to help eliminate long-term errors, and the differential term predicts the trend of the deviation and makes corresponding adjustments in advance to improve the response speed and stability. The method of adjusting the valve opening using a PID controller based on the difference between the predicted flow value and the a priori target flow value can determine a larger target deviation, i.e., the difference between the predicted flow value and the a priori target flow value, when the model correction coefficient is large. This allows for a faster response to valve opening adjustment based on the PID controller, thereby improving the accuracy of valve opening adjustment. It should be noted that setting the flow target needs to be specifically determined based on the pipeline specifications in the specific implementation environment. In one specific implementation of the present invention, the flow target is set to 1500 cubic meters per hour and is adjustable.

[0080] To sum up, in a coaxial on-off valve for a natural gas pipeline that can be equipped with a flow prediction module proposed in this application, the flow prediction module is based on the geometric changes of the airflow in the pipeline and the corner of the valve cavity, which easily forms a backflow zone dominated by inertial airflow, accompanied by the mutual reaction of the low-pressure vortex and the mainstream airflow, resulting in dynamic lag or vibration distortion of the pipeline mass flow. Through empirical mode decomposition and fast Fourier transform algorithm, the fluid lag effect and vibration distortion characteristics in the pipeline are dynamically analyzed, the flow fluctuation coefficient, lag characteristic coefficient and vibration abnormality coefficient are extracted, and the time-series dynamic lag curve and time-series vibration distortion curve are constructed. Based on the correlation between the two, a dynamic correction model is established to compensate for the nonlinear distortion of the flow-valve opening transfer function in real time, thereby obtaining a more accurate predicted flow value after correction, so that the accuracy of valve opening adjustment based on the predicted flow value is higher.

[0081] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0082] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. A coaxial on-off valve for a natural gas pipeline capable of carrying a flow prediction module, characterized in that: Including traffic prediction module; the traffic prediction module includes: The data acquisition and preprocessing submodule is used to collect pipeline flow signals and pipeline vibration signals of the natural gas pipeline; perform empirical mode decomposition on the pipeline flow signals to obtain corresponding IMF component signals; The first determination submodule is configured to determine, in the IMF component signal, a hysteresis characteristic coefficient at each sampling moment based on the intensity of signal fluctuations and the phase deviation of the envelope in a historical time period at each sampling moment; and determine a time series dynamic hysteresis curve based on the hysteresis characteristic coefficients at each sampling moment; The second determination submodule is configured to determine, on the pipeline vibration signal, a vibration anomaly coefficient value at each sampling moment based on a local change of the signal within a time sequence neighborhood at each sampling moment; and determine a time sequence vibration distortion curve based on the vibration anomaly coefficient values at each sampling moment; The valve opening adjustment submodule is used to determine the model correction coefficient based on the hysteresis characteristic coefficient, the vibration abnormality coefficient value and the curve correlation between the corresponding time-series dynamic hysteresis curve and the time-series vibration distortion curve at the current moment; correct the real-time pipeline flow data according to the model correction coefficient to determine the predicted flow value; and adjust the valve opening according to the predicted flow value.

2. A natural gas pipeline coaxial on-off valve capable of carrying a flow prediction module according to claim 1, characterized in that: The process of obtaining the hysteresis characteristic coefficient includes: The local signal segment corresponding to the historical time period at each sampling moment in the IMF component signal is used as the historical component signal segment at each sampling moment; Determine the flow fluctuation coefficient at each sampling moment according to the slope deviation on both sides of the extreme value point in the historical component signal and the fluctuation of the signal difference value; Obtaining an upper envelope and a lower envelope of the historical component signal segment; using a phase difference between frequency domain data after fast Fourier transform of the upper envelope and frequency domain data after fast Fourier transform of the lower envelope as a flow hysteresis coefficient at a sampling moment corresponding to the historical component signal segment; The hysteresis characteristic coefficient at each sampling moment is determined according to the product of the flow hysteresis coefficient and the flow fluctuation coefficient.

3. A natural gas pipeline coaxial on-off valve capable of carrying a flow prediction module according to claim 2, characterized in that: The process of obtaining the flow fluctuation coefficient includes: In the historical component signal segment, the difference between the tangent slope at the first sampling moment before each extreme value point and the tangent slope at the first sampling moment after each extreme value point is used as the slope change value for determining each extreme value point; and the degree of extreme value change at each sampling moment is determined based on the average of the slope change values of all extreme value points in the historical component signal segment; In the historical component signal segment, the difference between the signal value at each moment and the signal value at the previous moment is used as the instantaneous signal difference at each moment; and the degree of timing fluctuation at the sampling moment corresponding to the historical component signal segment is determined based on the mean value of the instantaneous signal differences at all moments in the historical component signal segment; The flow fluctuation coefficient at each sampling moment is determined according to the product of the extreme value change degree and the time series fluctuation degree.

4. A natural gas pipeline coaxial on-off valve capable of carrying a flow prediction module according to claim 1, characterized in that: The process of obtaining the time series dynamic hysteresis curve includes: The hysteresis characteristic coefficients of all sampling moments are arranged in time sequence and then curve fitting is performed to obtain the time series dynamic hysteresis curve.

5. The natural gas pipeline coaxial on-off valve capable of carrying a flow prediction module according to claim 1, characterized in that: The process of obtaining the vibration abnormality coefficient value includes: In the pipeline vibration signal, the pipeline vibration signal is divided into at least two initial window sub-signal segments with the current time as the starting point and a preset number of data points as the step size; and the corresponding data fluctuation coefficient is determined according to the standard deviation of all signal values in each initial window sub-signal segment; Merge windows based on the similarity of data fluctuation coefficients between adjacent initial window sub-signal segments to obtain all merged window sub-signal segments; Determine the corresponding combined fluctuation coefficient based on the average of the data fluctuation coefficients of all initial window sub-signal segments in each combined window sub-signal segment; normalize the difference between the combined fluctuation coefficient of each combined window sub-signal segment and the combined fluctuation coefficient of the previous combined window sub-signal segment to determine the variation anomaly coefficient of each combined window sub-signal segment; and determine the corresponding overall anomaly coefficient based on the product of the normalized value of the combined fluctuation coefficient and the variation anomaly coefficient; The overall abnormal coefficient of the merged window sub-signal segment at each sampling moment is taken as the corresponding vibration abnormal coefficient value.

6. A natural gas pipeline coaxial on-off valve capable of carrying a flow prediction module according to claim 5, characterized in that: The process of acquiring the merged window sub-signal segments includes: In chronological order, the adjacent fluctuation difference that characterizes the difference between the data fluctuation coefficient of each initial window sub-signal segment and the data fluctuation coefficient of the previous initial window sub-signal segment is calculated; two initial window sub-signal segments whose normalized value of the adjacent fluctuation difference is less than a preset fluctuation threshold are merged; after traversing all initial window sub-segments, all merged window sub-signal segments are obtained.

7. A natural gas pipeline coaxial on-off valve capable of carrying a flow prediction module according to claim 1, characterized in that: The process of obtaining the time series vibration distortion curve includes: The vibration anomaly coefficient values at all sampling moments are arranged in chronological order and then curve fitting is performed to determine the time series vibration distortion curve.

8. The natural gas pipeline coaxial on-off valve capable of carrying a flow prediction module according to claim 1, characterized in that: The process of obtaining the model correction coefficient includes: Normalize the product of the hysteresis characteristic coefficient and the vibration abnormality coefficient value at the current moment to determine the reference correction coefficient at the current moment; Calculating the Pearson correlation coefficient between the time series dynamic hysteresis curve and the time series vibration distortion curve; The model correction coefficient at the current moment is determined according to the positive correlation mapping value of the sum value between the reference correction coefficient and the Pearson correlation coefficient.

9. The natural gas pipeline coaxial on-off valve capable of carrying a flow prediction module according to claim 1, characterized in that: The process of obtaining the predicted flow value includes: The predicted flow value at the current moment is determined according to the product between the predicted flow value and the pipeline flow data at the current moment.

10. A natural gas pipeline coaxial on-off valve capable of carrying a flow prediction module according to claim 1, characterized in that: The process of adjusting the valve opening according to the predicted flow value includes: Based on the PID controller, the valve opening is adjusted through the proportional term, the integral term and the differential term according to the difference between the predicted flow value and the prior target flow value.

Images