An adaptive control based fluid leak automatic response emergency shut-off valve system

Abstract

The application relates to the technical field of fluid control and pipeline safety protection, and discloses a fluid leakage automatic response emergency closing valve system based on adaptive control, which comprises a main valve, an actuator, a bypass, a pilot valve, a pressure sensor and a calculation node. The calculation node controls the actuator to drive the main valve to perform a primary cut-off action; the pilot valve is driven to generate a quadrature encoding transient pressure wave and inject the downstream pipeline through the bypass; a transient pressure time sequence signal is acquired, and adaptive wave speed data is output by calling a correction model when phase change of fluid is determined; a net echo signal is extracted based on cepstrum analysis, and the actual physical distance of a leakage point and an equivalent leakage hole cross-sectional area are calculated in combination with the adaptive wave speed data; and then the actuator is controlled to adjust a secondary speed reduction closing trajectory of the main valve. The application can effectively compensate wave speed attenuation caused by phase change to improve ranging accuracy, and adaptively match a fast closing mode or a slow closing mode, so that leakage blocking and water hammer protection are considered.

Description

Technical Field

[0001] This invention relates to the field of fluid control and pipeline safety protection technology, specifically to an automatic emergency shut-off valve system for fluid leakage based on adaptive control. Background Technology

[0002] During the operation of pressurized fluid pipeline networks, fluid leakage caused by pipeline rupture can lead to internal pressure imbalances and media loss. To control the leak source, emergency shut-off valve systems are typically deployed in the main pipeline to physically shut off the flow when abnormal conditions are detected. However, existing fluid leak detection and emergency shut-off technologies still have limitations under complex operating conditions.

[0003] Current transient pressure wave ranging methods mostly rely on pre-set fixed single-phase fluid sound velocities for physical distance estimation. When a pipeline leaks, the high pressure difference between the inside and outside causes a local pressure drop in the pipe section. If the local static pressure is lower than the saturated vapor pressure of the fluid medium at the current operating temperature, the fluid will undergo thermodynamic flash evaporation and cross the phase transition critical point to precipitate a gas phase, forming a gas-liquid two-phase mixed flow. The introduction of the gas phase changes the overall density and elastic modulus of the fluid medium, causing a nonlinear attenuation of the pressure wave propagation velocity within the pipeline network. Traditional fixed sound velocity positioning models, lacking real-time sensing of the medium's phase transition state and a wave velocity compensation mechanism, will produce ranging errors under such conditions, making it difficult to accurately locate the spatial position of the leak source.

[0004] Meanwhile, industrial pipeline networks are typically accompanied by high background turbulence noise, and structures such as bends and variable cross-sections within the network can trigger multipath reflections of sound waves. Existing pressure wave detection schemes mainly employ direct threshold comparison or basic time-domain filtering, which struggles to separate and extract weak echo features under strong background noise. This conventional signal processing approach easily leads to convolution and aliasing between the initial excitation source signal and the spatial multipath impulse response signal, thereby affecting the accuracy of extracting echo delay time and reflected wave amplitude attenuation characteristics, resulting in misjudgments of the physical distance to the leak and the extent of damage.

[0005] During the leakage containment control phase, existing emergency valves typically employ a pre-set single-rate or fixed-section curve for closing. While simple rapid valve closure can shorten the containment time and reduce media loss, the instantaneous fluid interruption causes a rapid conversion of kinetic energy into pressure potential energy, triggering transient water hammer effects. The resulting localized overpressure can easily cause secondary structural damage to the pipeline network walls and along the pipeline. Conversely, while conventional slow valve closure can buffer fluid kinetic energy dissipation to prevent water hammer overpressure, the prolonged throttling time allows for continuous and substantial fluid leakage, increasing the risk of secondary disasters. Existing systems, unable to automatically optimize based on the actual physical distance and rupture scale of the leak point, struggle to achieve adaptive dynamic adjustment of the valve closing trajectory between controlling the total leakage and preventing water hammer impact. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides an automatic emergency shut-off valve system for fluid leakage based on adaptive control. This system solves the problem that when a fluid pipeline leaks, the leakage of high-pressure differential fluid causes a severe local pressure drop, making it easy for the fluid medium to cross the phase transition critical point and precipitate a gas phase. This results in serious ranging errors in traditional fixed sound velocity ranging models due to the severe attenuation of wave velocity in two-phase flow media.

[0007] To address the above problems, the present invention provides the following technical solution:

[0008] The first aspect of this invention provides an automatic response emergency shut-off valve system for fluid leakage based on adaptive control, comprising:

[0009] The main valve is installed in the main branch of the fluid pipeline network;

[0010] The actuator is connected to the main valve via a drive mechanism;

[0011] A bypass is provided, connecting the fluid pipelines upstream and downstream of the main valve, and the bypass forms an injection port in the fluid pipeline downstream of the main valve.

[0012] A pilot valve is located on the bypass and is connected to the actuator via a transmission connection;

[0013] A pressure sensor is installed in the fluid pipeline downstream of the injection port;

[0014] The computing nodes are communicatively connected to the actuator, the pilot valve, and the pressure sensor, respectively.

[0015] The computing node controls the actuator to drive the main valve to perform a primary flow throttling action to determine the primary flow throttling boundary position;

[0016] The computing node drives the pilot valve to generate an orthogonally encoded transient pressure wave, which is injected into the single-phase pure liquid fluid downstream of the main valve via the bypass.

[0017] The computing node acquires the transient pressure time-series signal collected by the pressure sensor, and when it determines that a local phase change has occurred in the fluid, it calls the fluid phase change boundary correction model to output adaptive wave velocity data.

[0018] The computing node extracts the net echo signal from the transient pressure time series signal based on cepstral analysis, and calculates the actual physical distance of the leak point and the equivalent cross-sectional area of ​​the leak hole by combining the adaptive wave velocity data.

[0019] The computing node controls the actuator to adjust the secondary deceleration and valve closing trajectory of the main valve based on the actual physical distance and the equivalent leakage cross-sectional area.

[0020] In a preferred embodiment of the present invention, the axial physical distance between the injection port and the central axis of the main valve satisfies the spatial constraint condition: the axial physical distance is greater than the maximum extension length boundary of the cavitation collapse zone of the main valve calculated based on fluid dynamics, and less than or equal to the preset upper limit of the effective detection distance. The technical mechanism of the above setting is that when the main valve performs the primary throttling action, a cavitation unsteady flow field and bubble collapse noise are generated behind the valve port due to local throttling. By setting the bypass injection port outside the maximum extension length boundary of the cavitation collapse zone, the orthogonally encoded transient pressure wave avoids the cavitation excitation interference zone behind the main valve in physical space and is injected into a relatively stable single-phase pure liquid fluid to maintain the waveform integrity of the initial excitation signal.

[0021] In a preferred embodiment of the present invention, the fluid phase change boundary correction model invoked by the computing node to determine local phase change in the fluid specifically includes: when the real-time static pressure data collected by the pressure sensor is continuously less than or equal to the standard saturated vapor pressure of the fluid medium at the current operating temperature within a preset continuous sampling period, it is determined that the local fluid in the pipeline network has crossed the critical point of phase change; the computing node deduces the dynamic gas content based on the fluid medium's equation of state: the difference between the standard saturated vapor pressure and the real-time static pressure data is divided by the standard saturated vapor pressure, the result of the phase division is multiplied by the gas phase precipitation expansion coefficient, and finally the initial dissolved gas volume free rate is added to obtain the dynamic gas content. The physical significance of this correction model is that, for fluid density changes caused by thermodynamic flash evaporation, the system dynamically calculates the gas phase volume occurrence rate using the real-time pressure drop amplitude, thereby providing real-time physical property parameters for quantitative correction of sound wave propagation velocity.

[0022] In a preferred embodiment of the present invention, the numerical correction logic for the adaptive wave velocity data output by the computing node is as follows: The first term is obtained by dividing the dynamic gas content by the product of the density of the precipitated gas phase and the square of the gas phase sound velocity; the second term is obtained by dividing the difference between the dynamic gas content and the density of the single-phase pure liquid by the product of the square of the liquid phase sound velocity; the first and second terms are added together, multiplied by the comprehensive density of the mixed medium calculated from the dynamic gas content, and the reciprocal of the square root of the multiplication result is taken to obtain the adaptive wave velocity data. The technical effect of the above correction logic is that it utilizes a two-phase fluid dynamics sound velocity model to couple the nonlinear influence of gas phase precipitation on the comprehensive elastic modulus and density of the mixed fluid in real time, thereby compensating for the wave velocity calculation deviation introduced by the unsteady phase change of the fluid.

[0023] In a preferred embodiment of the present invention, the specific process of the computing node calculating the actual physical distance of the leak point includes: the computing node performing cross-correlation integration on the transient pressure time-series signal and the original transmission sequence to extract the net echo signal; sequentially performing fast Fourier transform, logarithmic calculation, and inverse Fourier transform on the net echo signal to generate a cepstrum sequence, and extracting the round-trip main phase delay time corresponding to the main peak greater than the validity judgment threshold; and solving the actual physical distance by combining the adaptive wave velocity data: performing time integration on the adaptive wave velocity data within the time interval from the initial trigger time to the sum of the initial trigger time and the round-trip main phase delay time, and dividing the integration result by two to obtain the actual physical distance. The technical mechanism of this ranging process is that the cross-correlation integration utilizes the autocorrelation characteristics of orthogonal coding to smooth out broadband turbulent noise of the fluid, and uses cepstrum analysis to homomorphically separate the nonlinear convolutional aliased acoustic excitation source from the multipath impulse response of the pipeline space to extract the main phase delay time; subsequently, by performing continuous time integration on the adaptive wave velocity, the interference of phase transition along the path on spatial positioning is compensated.

[0024] In a preferred embodiment of the present invention, the specific process of the computing node calculating the equivalent leakage cross-sectional area includes: extracting the absolute amplitude of the main peak of the net echo signal, and performing natural damping compensation calculation in conjunction with the actual physical distance to obtain the reflected wave amplitude attenuation coefficient; using the reflected wave amplitude attenuation coefficient and the current steady-state static pressure value of the pipeline network as input indices, substituting them into the preloaded leakage cross-sectional area mapping matrix, and matching the equivalent leakage cross-sectional area under the current operating condition through two-dimensional linear interpolation. The technical mechanism of this calculation process lies in that, based on acoustic impedance theory, the acoustic pressure energy reflection characteristics caused by the acoustic impedance discontinuity at the pipeline rupture point are converted into equivalent physical damage parameters, thereby realizing the quantitative calculation of the degree of damage.

[0025] In a preferred embodiment of the present invention, the specific process of the computing node adjusting the secondary deceleration closing trajectory of the main valve includes: the computing node using the calculated actual physical distance and equivalent leakage cross-sectional area as coordinate variables and inputting them into a two-dimensional fluid dynamics control decision matrix; when the actual physical distance is not greater than a preset near-end distance threshold and the equivalent leakage cross-sectional area is not less than a preset severe rupture area threshold, a nonlinear emergency fast-closing mode is triggered; when the above concurrent conditions are not met, an exponential water hammer protection slow-closing mode is triggered. The technical mechanism of this decision association lies in establishing a multi-dimensional boundary constraint decision logic, weighing the risk of medium leakage and the risk of water hammer impact based on the calculated physical distance and leakage area, and outputting a corresponding graded interception control mode.

[0026] In a preferred embodiment of the present invention, in the nonlinear emergency fast-closing mode: the computing node controls the main valve to close with maximum tangential acceleration. The calculation logic for its drive trajectory is as follows: calculate the time difference between the real-time time and the valve closing start time; multiply half of the maximum allowable mechanical acceleration by the square of the time difference to obtain the displacement attenuation; subtract the displacement attenuation from the primary throttling boundary position to obtain the theoretical displacement stroke; take the maximum value between the theoretical displacement stroke and zero as the real-time displacement stroke of the main valve; simultaneously, the computing node sends a forced closure signal to the pilot valve to cut off the bypass. The technical mechanism of this control method is that, for the near-end severe rupture condition, the maximum mechanical operation constraint is used to implement the blocking, and the displacement control command is prevented from exceeding the limit through the extreme value function; the bypass is isolated synchronously to cut off the interference path of the pilot valve caused by the flow field oscillation, thus completing the physical locking of the leakage source.

[0027] In a preferred embodiment of the present invention, under the exponential water hammer protection slow-closing mode: the computational node controls the main valve to close according to an exponential decay trajectory. The calculation logic is as follows: multiply the time difference between the real-time time and the valve closing start time by the dynamically solved decay control parameter, take the opposite number as the exponent, and raise the power with the natural constant as the base; multiply the power result by the primary throttling boundary position to obtain the real-time displacement stroke of the main valve. The technical mechanism of this trajectory control is that, for mid-to-far end or moderately severe damage conditions, the compressibility of the fluid medium is utilized to make the throttling rate of the main valve asymptotically approach zero in the near-fully closed range based on the exponential decay trajectory, guiding the fluid kinetic energy to be converted and dissipated into elastic potential energy and internal energy.

[0028] In a preferred embodiment of the present invention, the real-time calculation logic of the attenuation control parameter satisfies the following: the fluid buffer damping coefficient is divided by the product of the actual physical distance and the adaptive wave velocity data to obtain the attenuation control parameter; and when the real-time displacement stroke of the main valve is less than or equal to the preset mechanical shut-off dead zone threshold, the calculation node exits the exponential attenuation calculation and issues a zero-set shut-off command to the actuator. The technical mechanism of this optimization scheme lies in establishing a mathematical relationship between the valve speed reduction and closure curvature, the leakage physical depth, and the real-time wave velocity, and using the mechanical shut-off dead zone threshold to cut off the infinite asymptotic output of the exponential function, thereby eliminating the fine-tuning dead zone state of the actuator at the fully closed boundary and ensuring the discrete and accurate execution of the flow-blocking action.

[0029] A second aspect of the present invention provides an automatic response emergency shut-off valve control method for fluid leakage based on adaptive control, applied to the computing node of the above-mentioned system, comprising:

[0030] The control actuator drives the main valve to perform a primary throttling action to determine the primary throttling boundary position;

[0031] The pilot valve is driven to generate an orthogonally encoded transient pressure wave, which is then injected into the single-phase pure liquid fluid downstream of the main valve via a bypass.

[0032] Acquire transient pressure time-series signals collected by pressure sensors, and when it is determined that a local phase change has occurred in the fluid, call the fluid phase change boundary correction model to output adaptive wave velocity data;

[0033] Based on cepstral analysis, the net echo signal is extracted from the transient pressure time series signal, and the actual physical distance of the leak point and the equivalent cross-sectional area of ​​the leak hole are calculated by combining the adaptive wave velocity data.

[0034] Based on the actual physical distance and the equivalent leakage cross-sectional area, the actuator is controlled to adjust the secondary deceleration valve closing trajectory of the main valve.

[0035] This invention provides an automatic emergency shut-off valve system for fluid leakage based on adaptive control. It offers the following advantages:

[0036] 1. This invention improves the accuracy of locating the actual physical distance of the leak point under two-phase flow conditions by calling the fluid phase change boundary correction model when a local phase change of the fluid is determined, dynamically outputting adaptive wave velocity data, and quantitatively compensating for the sound velocity attenuation caused by the precipitation of gas phase when the fluid medium crosses the phase change critical point due to local pressure drop in the pipeline network.

[0037] 2. This invention utilizes the autocorrelation characteristics of orthogonally encoded transient pressure waves, combined with cross-correlation integrals and cepstral analysis, to separate the nonlinear convolutional aliasing acoustic excitation source and spatial multipath impulse response from the transient pressure time series signal. This suppresses the masking of signal features by broadband turbulent noise in the fluid, thereby ensuring the extraction accuracy of echo delay time and equivalent pore cross-sectional area.

[0038] 3. Based on a two-dimensional fluid dynamics control decision matrix, this invention uses the actual physical distance and equivalent leakage cross-sectional area to optimize the matching between the nonlinear emergency fast-closing mode and the exponential water hammer protection slow-closing mode. It also introduces a mechanical shut-off dead zone threshold to eliminate the fine-tuning dead zone of the shut-off action near the fully closed boundary, thereby taking into account both pipeline leakage control and transient water hammer pressure protection. Attached Figure Description

[0039] Figure 1 This is a topology diagram of a fluid leakage emergency interception system according to an embodiment of the present invention;

[0040] Figure 2 This is a flowchart of an emergency flow control method for fluid leakage according to an embodiment of the present invention;

[0041] Figure 3 This is a timing diagram for the initial assessment of suspected leakage status and the triggering of emergency response according to an embodiment of the present invention;

[0042] Figure 4 This is a logic diagram for aligning the decoupling action and control timing of the main and auxiliary valves according to an embodiment of the present invention;

[0043] Figure 5 This is a schematic diagram of the high-frequency pseudo-random coding excitation and spatial bypass injection principle according to an embodiment of the present invention;

[0044] Figure 6 This is a logic diagram for real-time adaptive micro-compensation of wave velocity under the residual phase transition boundary according to an embodiment of the present invention;

[0045] Figure 7 This is a logic diagram for echo signal decoding and leakage quantization based on cepstral analysis according to an embodiment of the present invention;

[0046] Figure 8 This is a secondary adaptive variable rate deceleration valve control logic diagram according to an embodiment of the present invention.

[0047] Among them, 10 is the computing node; 20 is the main circuit; 21 is the main valve; 22 is the actuator; 30 is the bypass; 31 is the pilot valve; 40 is the pressure sensor; and 50 is the control system. Detailed Implementation

[0048] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0049] See attached document Figure 1 , Figure 1 This is a topology diagram of a fluid leakage emergency shut-off system according to an embodiment of the present invention. The present invention provides an automatic response emergency shut-off valve system for fluid leakage based on adaptive control, which may include: a computing node 10, a main circuit 20, a bypass 30, a pressure sensor 40, and a temperature acquisition unit. The temperature acquisition unit is a temperature sensor installed at a measuring point in the fluid pipeline network, or a pipeline temperature transmitter connected through a control system 50, used to provide the computing node 10 with the real-time operating temperature of the fluid medium.

[0050] The main circuit 20 is located in the fluid pipeline trunk line. The main circuit 20 is equipped with a main valve 21 and an actuator 22 that drives the main valve 21. The actuator 22 is a pneumatic actuator or an electric actuator, used to provide the mechanical thrust required for fluid throttling.

[0051] Bypass 30 is connected in parallel to main circuit 20. Bypass 30 is equipped with pilot valve 31. The response frequency of pilot valve 31 is greater than or equal to 100Hz. The pressure tap of bypass 30 is located upstream of main valve 21. The injection port of bypass 30 is located downstream of main valve 21. Bypass 30 is also equipped with an anti-backflow unit, which includes at least one of a one-way check valve, a normally closed shut-off valve, or a fail-safe closing mechanism interlocked with pilot valve 31. This unit is used to cut off the fluid connection of bypass 30 when main valve 21 performs a complete shut-off action or when pilot valve 31 experiences a power failure, gas failure, or communication interruption, preventing the fluid medium from bypassing main valve 21 and continuing to leak downstream via bypass 30.

[0052] The physical distance between the injection inlet and the main valve 21 is set to The maximum cavitation collapse zone length generated when the main valve 21 performs a flow-blocking operation under rated conditions is set to... The spatial location of the injection port of bypass 30 satisfies the constraints. This spatial constraint relationship is used to allow the fluid medium to bypass the two-phase flow scattering zone behind the main valve 21.

[0053] In this embodiment, the injection port of bypass 30 is located downstream of main valve 21. Therefore, the pseudo-random coded transient pressure wave injected by bypass 30 is mainly used to detect leakage points in the downstream pipe section of main valve 21. For leakage monitoring of the upstream pipe section of main valve 21, a corresponding bypass injection unit and pressure acquisition unit can be set up separately on the upstream side of main valve 21, or independent leakage determination can be performed through the upstream pressure instrument and flow instrument connected to control system 50.

[0054] Pressure sensor 40 is positioned at the same axial measurement section as the injection port of bypass 30 or downstream of the injection port. The sampling rate of pressure sensor 40 is greater than or equal to 10 kHz. Pressure sensor 40 is used to acquire transient pressure timing signals within the fluid pipeline network in real time. The axial distance between pressure sensor 40 and the injection port of bypass 30 is set as follows: When the pressure sensor 40 and the bypass 30 injection port are located on the same axial measuring section, When there is an axial misalignment between the pressure sensor 40 and the bypass 30 injection port, The calibrated structural parameters are pre-stored in computing node 10 and the path length is corrected during the calculation of the physical distance to the leak point.

[0055] Computational node 10 establishes communication connections with actuator 22, pilot valve 31, pressure sensor 40, and temperature acquisition unit. When the temperature acquisition unit is connected through control system 50, computational node 10 obtains the real-time operating temperature of the fluid medium through control system 50. Computational node 10 has independent real-time control and signal processing threads internally. Computational node 10 establishes data communication with control system 50 through an industrial bus. Control system 50 is used to acquire fluid dynamic parameters of the upstream and downstream of the fluid pipeline and send them to computational node 10.

[0056] See attached document Figure 2 , Figure 2 This is a flowchart of a fluid leakage emergency shut-off method according to an embodiment of the present invention. The present invention provides an automatic emergency valve shut-off method for fluid leakage based on adaptive control, comprising the following steps:

[0057] S10, the computing node 10 obtains the static pressure gradient and volumetric flow rate change rate of the fluid pipeline network in real time through the control system 50 and performs filtering processing. When the absolute deviation of the static pressure gradient from the reference static pressure gradient exceeds the preset pressure gradient safety boundary threshold, or the absolute value of the volumetric flow rate change rate exceeds the preset flow change rate safety boundary threshold, the emergency flow cut-off procedure is triggered.

[0058] S20, computing node 10 sends a primary throttling command to actuator 22, driving the valve core of main valve 21 to move towards the preset primary throttling boundary, and sends a control signal to pilot valve 31 within the synchronous time window of main valve 21 action.

[0059] S30, the pilot valve 31 performs a high-frequency throttling action according to a preset pseudo-random displacement sequence, and outputs a pseudo-random encoded transient pressure wave with autocorrelation peak characteristics and low cross-correlation characteristics, which is introduced into the fluid medium downstream of the main valve 21 through the injection port of the bypass 30.

[0060] S40, pressure sensor 40 continuously collects transient pressure time-series signals in the fluid pipeline network. When the transient pressure drop at the measurement point obtained by computing node 10 reaches the boundary value of the saturated vapor pressure of the medium, the preloaded fluid phase change boundary correction model is called to compensate for the pressure wave propagation constant and output adaptive wave velocity data.

[0061] S50, the computing node 10 performs cross-correlation processing and cepstral analysis on the transient pressure time-series signal collected by the pressure sensor 40 and the pseudo-random displacement sequence, extracts the phase delay time and amplitude attenuation coefficient of the reflected wave, and calculates the actual physical distance from the injection port to the pipeline leakage point and the equivalent leakage hole cross-sectional area.

[0062] S60, the computing node 10 inputs the actual physical distance and equivalent leakage cross-sectional area into the fluid dynamics control decision matrix to generate the adaptive closing curve of the main valve 21. The actuator 22 drives the main valve 21 to the fully closed position, and the pilot valve 31 performs the closing action simultaneously, and the anti-backflow unit of the bypass 30 enters the cut-off state to block the bypass flow channel after the main valve 21 is closed. When the computing node 10 confirms that the main valve 21 has reached the fully closed position, the pilot valve 31 is in the closed state and the bypass 30 is in the cut-off state, it latches and uploads the solution data to the control system 50.

[0063] See attached document Figure 3 , Figure 3 This is a timing diagram of the initial judgment of suspected leakage and the triggering of emergency response according to an embodiment of the present invention. In the fluid leakage emergency interception method provided by the present invention, the computing node 10 obtains the static pressure gradient and volumetric flow rate change rate of the fluid pipeline in real time through the control system 50 and performs filtering processing. The computing node 10 and the control system 50 establish a data transmission channel to obtain multi-source fluid dynamic parameters on the main fluid pipeline in real time and perform filtering and smoothing processing.

[0064] Specifically, the data channel of computing node 10 is connected to control system 50 based on the industrial Ethernet protocol. Control system 50 periodically collects data from pressure and flow meters located upstream and downstream of the pipeline trunk line. From a physical perspective, when a leak occurs in the fluid pipeline trunk line, the fluid loss at the leak point causes a sudden drop in local static pressure, thus forming an abnormal pressure gradient between upstream and downstream measuring points. Simultaneously, due to the formation of new pressure drop funnels before and after the leak point, fluid converges towards the leak outlet, causing a sudden change in flow velocity, which in turn leads to drastic fluctuations in the time domain of the volumetric flow rate change rate. Let the one-dimensional flow direction of the fluid in the pipeline be the positive x-axis. The raw characteristic data acquired by computing node 10 includes the static pressure gradient within the trunk line span and the volumetric flow rate change rate at a specific cross-section. The static pressure gradient is obtained by calculating the difference between the upstream and downstream real-time pressures and dividing it by the axial distance between the upstream and downstream meters.

[0065] Industrial environments are characterized by pump operation disturbances and localized fluid turbulence, which introduce high-frequency broadband noise into the signals acquired by sensors. Computation node 10 is equipped with a digital filtering module to perform data preprocessing. The time-series data of the static pressure gradient and the time-series data of the volumetric flow rate change are respectively input into this digital filtering module. For the specific algorithm of the digital filtering, those skilled in the art can use a moving average filtering algorithm or a Kalman filtering algorithm to smooth the original time-series signal. The specific implementation of filtering out broadband noise from the pipeline network background is well-known in the art and will not be elaborated here. After processing by the digital filtering module, computation node 10 outputs the static pressure gradient, reflecting the fluid pipeline transport status. and volumetric flow rate change rate .

[0066] Compute node 10 executes the security boundary threshold determination logic for suspected leakage states, establishes the switching boundary of the system state machine, and triggers the emergency interception procedure.

[0067] The storage cells of compute node 10 are pre-loaded with pressure gradient safety boundary thresholds. and the safety boundary threshold for traffic mutation rate In this embodiment, the pressure gradient safety boundary threshold... Safety boundary threshold for flow mutation rate The calibration benchmark is determined by statistically analyzing the historical maximum fluctuation of the fluid pipeline network under rated transport conditions and multiplying it by an amplification factor of 1.2 to 1.5. This avoids normal operation fluctuations caused by routine load adjustments or valve switching in the pipeline network and enables accurate identification of substantial physical rupture leaks.

[0068] The real-time control thread allocated to compute node 10 continuously scans the static pressure gradient on the input side according to the set scan cycle. With volumetric flow rate change Execution condition comparison. To avoid the judgment algorithm entering a logical dead zone due to occasional packet loss, disconnection, or abnormal zero values ​​in the sensor communication link, computing node 10 performs data integrity verification before comparison. If the number of consecutive lost frames exceeds the preset frame loss tolerance threshold, the calculation state of the previous scan cycle remains unchanged and a fault alarm is issued to control system 50. When data communication returns to normal within the preset recovery time window, computing node 10 re-executes filter initialization and continues to determine the safety boundary threshold; when data communication does not return to normal within the preset recovery time window and the main valve 21 has entered the emergency throttling mode, computing node 10 controls the main valve 21 to enter a conservative slow-closing state according to the preset safety degradation strategy, and closes the pilot valve 31 and bypass 30; after the data integrity verification is passed, the system establishes the boundary condition for triggering the emergency throttling procedure as satisfying any of the following inequalities, namely, the absolute deviation of the static pressure gradient from the reference static pressure gradient exceeds the pressure gradient safety boundary threshold, or the absolute value of the volumetric flow rate change rate exceeds the flow rate change rate safety boundary threshold:

[0069] ;

[0070] ;

[0071] In the formula, This represents the real-time static pressure gradient of the pipeline network after filtering. This represents the reference static pressure gradient of the pipeline network under rated transport conditions or current steady-state reference conditions. This represents the preset pressure gradient safety boundary threshold. This represents the real-time volumetric flow rate change after filtering. This represents the preset safety boundary threshold for the traffic mutation rate.

[0072] If any of the above inequalities hold, computing node 10 determines that a suspected leak has occurred in the pipeline network. The state machine inside computing node 10 switches from steady-state monitoring mode to emergency closure mode. After the state machine switch, the command distribution module starts working, and computing node 10 begins generating closure command sequences to actuator 22 and pilot valve 31 in the subsequent operation chain. Through the above trigger condition determination and state machine switching logic, the system can utilize the time-domain abrupt change characteristics of pipeline fluid dynamic parameters for fault identification and status early warning, providing a timing control basis for subsequent precise closure operations.

[0073] See attached document Figure 4 , Figure 4This is a logic diagram for aligning the decoupling action and control timing of the main and auxiliary valves according to an embodiment of the present invention. In the fluid leakage emergency shut-off method provided by the present invention, the computing node 10 sends a primary shut-off command to the actuator 22, driving the valve core of the main valve 21 to move towards a preset primary shut-off boundary. The computing node 10 determines the primary shut-off boundary of the main valve 21 and the corresponding macroscopic shut-off control trajectory, and issues an action command to the actuator 22.

[0074] In industrial settings, the main valve 21 and its actuator 22, constrained by their large diameter and heavy-duty structure, exhibit significant mechanical inertia. If a full-stroke hard shut-off action is directly executed at the initial stage of an emergency response, the kinetic energy conversion of the fluid within the pipeline network may lead to water hammer and overpressure shock waves, increasing the risk of secondary damage to the pipeline. Based on this physical characteristic, computation node 10 divides the flow-closing process of the main valve 21 into a primary stage and a secondary stage.

[0075] Let the maximum mechanical stroke of the main valve 21 be... The stroke coordinates of the main valve 21 when it is in the fully open position are: The stroke coordinate of main valve 21 when it is in the fully closed position is 0. The storage unit of compute node 10 is preloaded with the primary cutoff constant ratio. Primary cutoff constant ratio The value range is set to 60% to 75%. This proportion is determined based on the boundary of the linear adjustment zone in the inherent flow characteristic curve of the main valve 21, and is used to cut off most of the leakage flow while avoiding extreme fluid overpressure in the pipeline network due to complete physical blockage. The coordinates of the primary cutoff boundary position of the main valve 21 are set as follows: ,in , This indicates the closing ratio of the main valve 21 relative to its maximum mechanical stroke during the initial throttling stage.

[0076] In triggering the emergency cutoff procedure At a certain moment, computing node 10 sends a primary throttling command to actuator 22. Actuator 22 drives the valve core of main valve 21 to perform a macroscopic throttling action at a preset rated maximum operating rate. In this embodiment, it is assumed that the main valve 21 is at... The real-time displacement travel at time t is In the initial interception stage, the displacement command generated by compute node 10 causes... Gradually approaching the target coordinates During the execution of this macroscopic flow throttling action, the computing node 10 only issues smooth displacement commands and does not superimpose high-frequency disturbances into the control signal of the main valve 21, thereby avoiding the physical limitations of the low-frequency response characteristics of the heavy-duty actuator 22.

[0077] Compute node 10 allocates isolated independent control threads and executes a timing alignment algorithm to establish a synchronization time window for the coordinated action of main valve 21 and pilot valve 31.

[0078] To prevent timing misalignment between the main valve 21 and the pilot valve 31 when receiving trigger commands due to differences in communication transmission nodes or device startup dead zones, the computing node 10 pre-constructs isolated macroscopic execution threads and microscopic excitation threads within its underlying operating system architecture. The macroscopic execution thread is dedicated to managing the calculation and distribution of control signals for the actuator 22, while the microscopic excitation thread is dedicated to managing the calculation and distribution of high-frequency control signals for the pilot valve 31. These two threads are allocated independent CPU computing power and cache queues to prevent high-frequency computation tasks from blocking the execution of macroscopic control signals.

[0079] Computing node 10 obtains the first response delay time of actuator 22 and the second response delay time of pilot valve 31 In this embodiment, the aforementioned first response delay time With the second response delay time The first response delay time can be obtained and stored in compute node 10 through the no-load calibration test during the system initialization phase. Includes control system bus communication delay and start-up dead time for heavy-duty mechanical components to overcome static friction; second response delay time. This includes the electrical response delay of the pilot control loop on bypass 30 and the drive delay of the piezoelectric crystal.

[0080] Compute node 10 executes timing alignment logic based on the acquired latency parameters, comparing the first response latency time. With the second response delay time .when At that time, the macroscopic execution thread of compute node 10 is in Immediately issue a macro-level flow control command to the execution mechanism 22, while the micro-level excitation thread undergoes... After the compensation waiting time, a control signal is sent to the pilot valve 31; when At that time, the micro-excitation thread is in Immediately send a control signal to pilot valve 31, and the macro execution thread experiences... After the compensation waiting period, a macro-level interception order is issued to the executing agency 22; when At that time, the macroscopic execution thread and the microscopic excitation thread are in The corresponding control signals are sent synchronously at all times.

[0081] Through the aforementioned time delay difference compensation mechanism, computation node 10 ensures that the moment when the main valve 21 core generates physical displacement coincides with the moment when the pilot valve 31 begins high-frequency oscillation. This mechanism provides stable upstream high-pressure flow field boundary conditions for the fluid detection signals subsequently generated by bypass 30, helping to reduce the systematic errors introduced by mechanical asynchrony into the analysis of fluid dynamic characteristics.

[0082] See attached document Figure 5 , Figure 5 This is a schematic diagram of high-frequency pseudo-random coding excitation and spatial bypass injection according to an embodiment of the present invention. In the fluid leakage emergency interception method provided by the present invention, the pilot valve 31 performs a high-frequency throttling action according to a preset pseudo-random displacement sequence, outputs a pseudo-random coded transient pressure wave, and the high-frequency pseudo-random coding sequence is generated inside the calculation node 10, which drives the pilot valve 31 to perform a dynamic displacement throttling action.

[0083] The micro-excitation thread of computing node 10 is equipped with a sequence generation module. This sequence generation module is used to generate pseudo-random shift sequences with specific code lengths and clock frequencies. For the specific generation algorithm of the pseudo-random shift sequences, those skilled in the art can use m-sequences or Gold sequences generated based on linear feedback shift registers, which have good autocorrelation and low cross-correlation, and are used to suppress background noise in subsequent signal processing. The specific implementation of its sequence generation is well-known in the art and will not be described in detail here.

[0084] Computation node 10 maps the generated discrete digital sequence into a continuous dynamic displacement drive signal. Specifically, computation node 10 converts the discrete digital sequence into a continuous analog voltage signal through its internal digital-to-analog converter circuit, and drives the piezoelectric crystal inside pilot valve 31 to deform via a power amplifier. At the trigger time... The dynamic displacement function of pilot valve 31 is expressed as:

[0085] ;

[0086] In the formula, Indicates that pilot valve 31 is in Real-time displacement distance at any given moment. This represents a high-frequency pseudo-random encoded sequence that has undergone amplitude mapping. Represents the unit step function. This indicates the synchronous triggering moment when the main valve 21 and the pilot valve 31 work together.

[0087] To ensure that the pilot valve 31 operates within its linear regulation range, the modulation amplitude of the high-frequency pseudo-random coded sequence is limited to between 10% and 20% of the maximum rated stroke of the pilot valve 31. This amplitude range is limited to prevent excessive physical amplitude from causing nonlinear distortion of the piezoelectric crystal or deep flow blockage of the bypass 30 fluid.

[0088] The high-frequency mechanical oscillation displacement of the pilot valve 31 couples with the fluid kinetic energy, outputting pseudo-random coded transient pressure waves into the fluid pipeline network.

[0089] The pressure tap of bypass 30 is located in the high-pressure stable region upstream of main valve 21, and the fluid medium entering bypass 30 has a high initial static pressure and fluid kinetic energy. When pilot valve 31 adjusts according to the dynamic displacement function... When performing high-frequency throttling, the flow cross-sectional area of ​​the valve seat inside bypass 30 undergoes high-frequency periodic changes. During this physical process, the stable upstream high-pressure fluid is disturbed by the high-frequency micro-displacement of the pilot valve 31, and the local kinetic energy and pressure energy of the fluid undergo high-frequency alternating conversion.

[0090] During this process, the high-frequency changes in the orifice area of ​​pilot valve 31 cause periodic changes in local fluid flow resistance. According to the principle of fluid momentum conservation, the originally stable static pressure is modulated by this mechanical throttling effect, thereby generating transient pressure waves carrying pseudo-random coding characteristics within the pipeline network. These pseudo-random coded transient pressure waves contain elements related to the dynamic displacement function in the time domain. Based on the frequency and phase sequence characteristics of the same source, the system uses this transient pressure wave as an acoustic carrier for subsequent detection of the physical boundary and leakage source characteristics of the fluid pipeline network.

[0091] The computation node 10 establishes spatial coordinate constraints based on the fluid phase change characteristics, so that the pseudo-random encoded transient pressure wave bypasses the cavitation collapse zone behind the main valve 21 and is injected into the main fluid medium.

[0092] When the main valve 21 performs its primary macroscopic flow-closing action, the flow area at the valve core of the main valve 21 decreases sharply, causing a surge in the flow velocity of the fluid medium at this point, leading to a drop in local static pressure. When the local static pressure falls below the saturated vapor pressure of the fluid medium, gas phase precipitates out from the liquid phase, forming a cavitation cavitation zone. This cavitation cavitation zone moves downstream with the fluid and collapses upon pressure recovery, forming a cavitation collapse zone filled with a gas-liquid two-phase mixture downstream of the main valve 21. Due to the acoustic impedance difference between the gas-liquid two-phase mixture, this region will scatter and attenuate high-frequency sound waves and transient pressure waves. If a pseudo-random coded transient pressure wave directly passes through this region, its coding characteristics will be distorted.

[0093] To maintain the energy penetration and waveform integrity of the pseudo-random encoded transient pressure wave, the injection port of bypass 30 is physically located downstream of main valve 21. Based on the Reynolds number of the fluid pipeline network, the pressure difference before and after throttling by main valve 21, and the saturated vapor pressure of the medium, calculation node 10 calculates the maximum cavitation influence zone length generated by main valve 21 when it performs a throttling action under rated operating conditions. In this embodiment, The calculation can be obtained by combining the cavitation bubble dynamics model in fluid dynamics. As one feasible method, the calculation node 10 takes the rated throttling pressure difference of the main valve 21, the saturated vapor pressure of the fluid medium, the pipe inner diameter, the primary throttling velocity of the main valve 21, and the Reynolds number of the medium as input parameters, and substitutes them into a preset empirical model of cavitation length to obtain the initial cavitation influence zone length. As another feasible method, during the system installation and commissioning phase, the main valve 21 is made to perform the primary throttling action according to the rated operating conditions, and the pressure wave attenuation intensity or gas-liquid two-phase scattering intensity is collected at different axial positions downstream of the main valve 21. The position where the pressure wave attenuation intensity recovers to the single-phase flow healthy baseline range is determined as the boundary of the cavitation collapse zone, thereby obtaining the maximum cavitation influence zone length. And it is fixed to the storage unit of computing node 10. The physical coordinate distance between the injection inlet and the central axis of the main valve 21 is set to The system satisfies the following spatial constraints in terms of hardware topology layout:

[0094] ;

[0095] In the formula, This indicates the axial physical distance between the bypass 30 injection port and the main valve 21. This represents the boundary of the maximum extension length of the cavitation collapse zone obtained from hydrodynamic calculations. This indicates the preset upper limit of the effective detection distance, used to prevent the detection signal from naturally attenuating after traveling too long a distance inside the pipeline network.

[0096] By constraining the physical space conditions described above, the pseudo-random coded transient pressure wave generated by bypass 30 bypasses the two-phase flow region and is directly injected into the stable region of the single-phase pure liquid fluid downstream of the main valve 21. This injection mechanism physically bypasses the two-phase flow scattering region generated by the throttling of the main valve 21, reduces the nonlinear interference of the phase transition process on the high-frequency detection signal, and enables the pseudo-random coded transient pressure wave to propagate towards the leak point under relatively stable single-phase physical field boundary conditions.

[0097] See attached document Figure 6 , Figure 6This is a logic diagram for real-time adaptive micro-compensation of wave velocity under the residual phase change boundary according to an embodiment of the present invention. In the fluid leakage emergency interception method provided by the present invention, the pressure sensor 40 continuously collects the transient pressure time-series signal in the fluid pipeline network. When the transient pressure drop at the measuring point obtained by the computing node 10 reaches the saturated vapor pressure boundary value of the medium, the computing node 10 monitors the transient pressure drop of the fluid downstream of the injection port in real time through the pressure sensor 40 and executes the logic for determining the local phase change critical point.

[0098] Pressure sensor 40 continuously acquires real-time static pressure data within the pipeline network and transmits it to computing node 10. Simultaneously, computing node 10 acquires the real-time operating temperature of the fluid medium via a temperature acquisition unit or a pipeline temperature transmitter connected to control system 50, and, in conjunction with a pre-loaded fluid property table, interpolates and extracts the standard saturated vapor pressure parameter corresponding to that temperature in real time. In transient conditions such as fluid leakage or rapid valve closure, a pressure drop wave propagating along the pipeline is generated within the network, causing a localized transient pressure drop at the measuring point. The signal processing thread of computing node 10 compares the real-time static pressure data with the standard saturated vapor pressure parameter in real time according to a set sampling clock.

[0099] To avoid frequent state switching near the phase transition critical point caused by fluid turbulence noise or high-frequency sensor jitter, a continuous sampling confirmation mechanism is introduced. The mathematical triggering condition for determining whether a fluid transitions from a single-phase liquid state to a gas-liquid two-phase flow state is defined as:

[0100] ;

[0101] In the formula, This indicates the real-time static pressure data continuously collected by pressure sensor 40 at the measuring point. This represents the standard saturated vapor pressure of the dynamically extracted fluid medium at the current operating temperature. When the above inequality holds true for N consecutive preset sampling periods, computation node 10 determines that the local fluid in the pipeline network has crossed the phase transition critical point, and gas phase precipitation occurs inside the fluid. Computation node 10 then activates its internally pre-loaded fluid phase transition boundary correction model to address the distortion of ranging parameters under gas-liquid two-phase flow conditions.

[0102] Computation node 10 calls the fluid phase change boundary correction model and quantifies the local transient pressure drop into transient dynamic gas content based on the medium property state equation.

[0103] In a gas-liquid two-phase flow state, the volume percentage of free gas inside the fluid increases as the static pressure decreases. Calculation node 10, based on the fluid medium's equation of state, quantifies and extrapolates the real-time static pressure data into a dynamic gas content. The specific calculation process satisfies the following mapping relationship:

[0104] ;

[0105] In the formula, This indicates that compute node 10 is in The local dynamic gas content of the pipeline network is calculated through continuous simulation. This represents the initial volumetric free gas fraction of a fluid under normal pressure and steady-state conditions. This represents the gas phase evolution expansion coefficient, which is related to fluid properties.

[0106] In an embodiment of the present invention, The value is usually determined based on routine laboratory indicators of the medium transported in the pipeline network, and its range is generally between 0.1% and 2%. The coefficient of thermal expansion can be obtained by fitting an offline constant-volume pressure-reducing calibration experiment on the target fluid medium. Specifically, in the offline calibration stage, the target fluid medium is placed in a constant-volume calibration container, the static pressure inside the container is changed sequentially, and the corresponding changes in the volume fraction of the evolved gas are recorded. The gas phase evolution expansion coefficient is obtained by least-squares fitting. and will , The applicable temperature and pressure ranges are stored in computing node 10. When the real-time pressure or temperature exceeds the calibrated applicable range, computing node 10 uses boundary parameters for conservative compensation and uploads an out-of-range alarm for physical property parameters to the control system 50. Computing node 10 converts the local transient pressure change amplitude into microscopic gas content data, providing physical property input for subsequent wave velocity correction.

[0107] Computation node 10 constructs a dynamic wave velocity compensation equation based on the dynamic gas content and the comprehensive density of the gas-liquid two-phase mixed medium, and outputs adaptive wave velocity data.

[0108] For obtaining basic fluid properties such as liquid density and gas density, those skilled in the art can consult standard fluid property databases or pre-calibrate them using conventional sensor arrays. The acquisition and configuration of these basic parameters are well-known technologies in the field and will not be elaborated here.

[0109] Computation node 10 extracts the pre-configured liquid phase medium density. Liquid phase sound velocity and the density of the gas phase medium Gas phase sound velocity The dynamic gas content calculated using node 10 is combined with the data. Calculate the overall density of the gas-liquid two-phase mixture. The calculation formula is as follows:

[0110] ;

[0111] From a physical perspective, the compressibility of a gaseous medium is much greater than that of a liquid medium. Therefore, even the precipitation of trace amounts of gas in a fluid will cause a sharp decrease in the overall bulk elastic modulus of the mixed fluid, leading to a nonlinear attenuation of the macroscopic sound wave propagation velocity. To address this, computation node 10 executes a dynamic wave velocity compensation equation based on the compressibility principle of the mixed fluid. The specific numerical correction logic used to reduce the impact of gas precipitation on the distortion of the sound wave propagation constant is as follows:

[0112] ;

[0113] In the formula, This represents the real-time adaptive wave velocity data output after phase transition correction calculation. This represents the calculated overall density of the mixed medium. and These represent the basic values ​​of density and sound wave propagation speed for a single-phase pure liquid fluid, respectively. and These represent the basic values ​​of the density of the precipitated gas phase and the sound wave propagation speed, respectively.

[0114] By executing the aforementioned self-correcting mathematical model, computation node 10 facilitates the dynamic calibration of acoustic ranging parameters under unsteady phase change environments. This numerical correction logic mitigates the abrupt change in pressure wave propagation constant caused by local cavitation, enabling the system to obtain a relatively stable and accurate compensated wave velocity even under residual phase change boundaries, thereby improving the benchmark accuracy for subsequent calculations of the spatial location of pipeline leakage sources.

[0115] See attached document Figure 7 , Figure 7 This is a logic diagram for echo signal decoding and leakage quantization based on cepstral analysis according to an embodiment of the present invention. In the fluid leakage emergency interception method provided by the present invention, the computing node 10 performs signal decoding on the acquired transient pressure time-series signal to extract the net echo. The computing node 10 performs matched filtering and cross-correlation operations to extract the net echo signal reflecting the changes in the physical boundary of the pipeline network from the background noise.

[0116] In actual operation, the pipeline network experiences fluid turbulence noise, valve throttling noise, and external mechanical vibration. These environmental disturbances are superimposed on weak echo signals. Computation node 10 acquires the transient pressure time-series signal synchronously collected by pressure sensor 40. To eliminate broadband fluid noise, the signal processing thread of computation node 10 performs cross-correlation integration on the transient pressure time-series signal and the original high-frequency pseudo-random encoded sequence generated in step S30. The specific digital processing logic adopts the following cross-correlation integration formula:

[0117] ;

[0118] In the formula, This represents the calculated cross-correlation function value. Indicates the time parameter of the integral translation. This indicates the maximum signal monitoring time window length set by the system. This indicates that the transient pressure timing signal received and converted by the pressure sensor 40 This indicates the high-frequency pseudo-random coding sequence used by the pilot valve 31 to operate. In this embodiment, the maximum signal listening time window length... The setting is based on twice the ratio of the maximum physical length of the pipeline to the minimum sound velocity of the fluid, to ensure that the system has the time-domain tolerance to receive the echo from the farthest boundary.

[0119] From a mathematical perspective, the pre-defined high-frequency pseudo-random coding sequence possesses good autocorrelation and low cross-correlation. When the time parameter is shifted... When aligned with the delay time of the true echo, the cross-correlation function The output shows a significant energy correlation peak; while fluid turbulence noise and background clutter in the pipeline network that are unrelated to the transmitted sequence are mitigated during integration. Through the above cross-correlation noise suppression mechanism, computing node 10 effectively reduces the interference of random broadband environmental noise and extracts a net echo signal with time-domain characteristics.

[0120] Computation node 10 performs cepstral analysis on the net echo signal to extract time delay features, and combines this with an adaptive wave velocity model to quantitatively calculate the spatial physical distance. Before performing cepstral peak identification, computation node 10 calls the pre-loaded pipeline health baseline reflection library, which records the axial position, echo delay time, and echo amplitude range of known structural boundaries such as flanges, elbows, reducers, branch pipes, and fixed valves under leak-free healthy operating conditions.

[0121] Computation node 10 compares the current net echo signal with the healthy baseline reflection library. It performs masking, cancellation, or weight reduction processing on inherent reflection peaks matching known structural boundaries, only considering reflection peaks newly added relative to the healthy baseline or whose amplitude changes exceed a preset leakage increment threshold as candidate leakage echo peaks. The pipeline healthy baseline reflection library is established during the system installation and commissioning phase. Under leak-free healthy operating conditions, computation node 10 drives pilot valve 31 to inject pseudo-random coded transient pressure waves, collecting inherent reflection peaks corresponding to flanges, elbows, reducers, branch pipes, and fixed valve components. This data is then combined with pipeline construction drawings or measured axial distances to generate the healthy baseline reflection library.

[0122] Within fluid pipeline networks, transient pressure waves encounter variable cross-section structures such as flanges, elbows, or necking points during propagation, leading to multipath reflection interference. Directly searching for echo peaks in the time-domain waveform can easily result in misjudgments due to time delays caused by multipath effects. Node 10 performs cepstral analysis on the net echo signal output after cross-correlation processing. From a signal processing perspective, cepstral analysis can separate the nonlinearly convolved and aliased acoustic excitation source from the spatial impulse response of the pipeline network.

[0123] Specifically, computing node 10 sequentially performs Fast Fourier Transform, logarithmic transformation, and Inverse Fourier Transform on the net echo signal, converting the frequency domain convolutional quantity into an additive component in the cepstrum domain. The underlying homomorphic signal processing algorithm for the cepstrum transform can be implemented using conventional digital signal processing library functions, which is well-known in the field and will not be elaborated upon here.

[0124] Computation node 10 has a pre-set threshold for determining the validity of cepstral peaks. From the candidate leakage echo peak set after masking, cancellation, or weighting by the healthy baseline reflection library, computation node 10 retrieves the first valid main peak on the amplitude attenuation envelope. If the absolute amplitude of this main peak is greater than the aforementioned validity threshold, computation node 10 determines it as a valid leakage echo, and the inverse frequency corresponding to this main peak represents the round-trip main phase delay time of the transient pressure wave propagating from the bypass 30 injection port to the leakage point and reflecting back. .

[0125] If the absolute amplitude of the main peak is not greater than the validity judgment threshold, it is determined that there is no abnormal leakage in the current pipeline section. The computing node 10 stops the high-frequency excitation of the pilot valve 31 and controls the main valve 21 to return to the pre-trigger opening position along the preset reset curve. When the main valve 21 is reset and the static pressure gradient and volume flow rate change rate are restored to within the safety boundary threshold, the computing node 10 switches the state machine from the emergency throttling mode back to the steady-state monitoring mode.

[0126] After acquiring the delay time, computing node 10 calls the adaptive wave velocity data updated in real time in step S40, and calculates the actual physical distance from the injection port to the leak point through time integration. When the pressure sensor 40 and the bypass 30 injection port are at the same axial measurement section, computing node 10 calculates the distance according to the round-trip path of the transient pressure wave propagating from the injection port to the leak point and reflecting back to the measurement point; when there is an axial offset between the pressure sensor 40 and the bypass 30 injection port, computing node 10 calculates the distance based on the pre-calibrated... The echo propagation path is subtracted or compensated to avoid incorrectly including sensor installation offset in the leakage distance. The calculation equation is as follows:

[0127] ;

[0128] In the formula, This represents the calculated actual physical distance from the injection port to the leak point. This indicates the initial triggering time of the transient pressure wave injected into the fluid medium. This represents the round-trip principal phase delay time extracted based on cepstral analysis. This represents the real-time adaptive wave velocity output by the fluid phase transition boundary correction model. When When the value is not 0, the computing node 10 performs an equivalent correction to the aforementioned round-trip propagation time or propagation path based on the installation direction of the pressure sensor 40 relative to the injection port of the bypass 30, so that... The value always represents the axial physical distance from the bypass 30 injection port to the leak point. This equation, by continuously integrating the wave velocity considering the dynamic effects of gas phase precipitation, largely compensates for the ranging error caused by the unsteady phase change of the fluid.

[0129] Node 10 calculates the amplitude attenuation coefficient of the reflected wave and uses the preloaded mapping matrix model to quantitatively solve the equivalent hole cross-sectional area.

[0130] At the fluid dynamics and acoustic coupling surface, a rupture or leak in the pipe network wall causes a sudden change in the acoustic impedance in that local area. When the incident transient pressure wave passes through this impedance discontinuity, part of the acoustic pressure energy is reflected. The larger the geometric size of the leak orifice, the higher the rate of acoustic impedance abrupt change, and the corresponding echo amplitude characteristics also exhibit regular variations. Node 10 extracts the absolute amplitude of the main peak in the net echo signal and, combined with the propagation distance, performs natural damping compensation of the pipe network to extract the reflected wave amplitude attenuation coefficient. The calculation method is expressed as follows:

[0131] ;

[0132] In the formula, This represents the calculated attenuation coefficient of the reflected wave amplitude. This represents the absolute amplitude of the main peak of the net echo signal extracted by computing node 10. This indicates the amplitude of the pre-calibrated reference incident wave source. This represents the natural attenuation constant of acoustic waves in the target fluid medium within the pipe network structure. This represents the leakage distance obtained through integration. The above reference incident wave source amplitude... With natural decay constant The test wave can be injected by manually driving the pilot valve 31 to inject the test wave and measure the friction loss amplitude of the multi-level sensor nodes under the leak-free and healthy operating conditions in the early stage of pipeline network deployment. The test wave can then be stored in the storage unit of the computing node 10.

[0133] The storage unit of compute node 10 pre-loads a leakage cross-sectional area mapping matrix. This mapping matrix is ​​a multi-dimensional data table constructed by collecting reflected acoustic data under different apertures through previous physical experiments on leakage purging under multiple operating conditions for pipe networks of the same specification. Specifically, during the calibration phase, multiple standard leakage holes with known areas are set on pipe sections of the same specification, and pseudo-random coded transient pressure waves are injected by driving pilot valve 31 under multiple steady-state pressure levels. The echo amplitude, echo delay time, and propagation attenuation characteristics corresponding to each standard leakage hole are collected, thereby establishing a mapping matrix based on the reflected wave amplitude attenuation coefficient. and the current steady-state static pressure value of the pipeline network For the input index, use the equivalent orifice cross-sectional area. This is a two-dimensional mapping matrix for the output elements.

[0134] The calculation node 10 will calculate the attenuation coefficient. The current steady-state static pressure value of the pipeline network is used as an input index and substituted into the preloaded mapping matrix. The system matches the corresponding leakage physical quantities through two-dimensional linear interpolation, thereby quantitatively calculating the equivalent leakage cross-sectional area under the current operating conditions. When the input index is located between adjacent calibration nodes, computation node 10 uses a two-dimensional linear interpolation method to obtain the index. When the input index exceeds the calibration boundary of the mapping matrix, computation node 10 outputs the corresponding boundary node. It also uploads out-of-range alarm information to the control system 50.

[0135] This mapping analysis process converts acoustic energy reflection characteristics into intuitive physical damage parameters, providing quantitative data support for the system to determine the pipeline damage level and output control decisions.

[0136] See attached document Figure 8 , Figure 8 This is a secondary adaptive variable rate deceleration valve control logic diagram according to an embodiment of the present invention. In the fluid leakage emergency interception method provided by the present invention, the computing node 10 performs an optimization decision between preventing fluid water hammer overpressure and reducing the amount of medium leakage based on the spatial physical distance of the leakage point and the equivalent leakage orifice cross-sectional area. The computing node 10 inputs the spatial physical distance and the equivalent leakage orifice cross-sectional area into the fluid dynamics control decision matrix to establish the input correlation between leakage characteristics and control mode.

[0137] The storage unit of compute node 10 is pre-loaded with a two-dimensional hydrodynamic control decision matrix. The horizontal axis parameter of this matrix represents the physical distance boundary, and the vertical axis parameter represents the damage area boundary. Computation node 10 extracts the actual physical distance of the leak point calculated in step S50. and equivalent leakage cross-sectional area Compute node 10 will and As input coordinate variables, they are substituted into the two-dimensional fluid dynamics control decision matrix for region matching.

[0138] Computation node 10 has a near-end distance threshold set internally. With the threshold of severe fracture area The basis for obtaining the above thresholds is the near-end distance threshold. Based on the propagation velocity of pipeline pressure waves, the critical reflection period of pipeline water hammer waves, and the shortest mechanical response time of actuator 22, this is comprehensively determined to define the high-pressure abrupt change region where water hammer shock waves have not yet interfered with each other. In a physical sense, This indicates the critical spatial range within which the pressure wave can propagate between the leak point and the main valve 21 and potentially form a superimposed overpressure during the shortest mechanical response time of the actuator 22; severe rupture area threshold. Typically set to 30% to 40% of the standard cross-sectional area of ​​the main pipeline, this threshold is used to define whether the fluid faces a large-flow-rate pressure loss ejection condition. Computation node 10 maps the continuous leakage state at the physical level to discrete control mode commands by comparing the currently calculated physical quantities with the aforementioned threshold parameters. At the equivalent implementation level of the patent, the logic of this decision matrix is ​​equivalent to a multi-dimensional lookup table control algorithm with boundary constraints.

[0139] Specifically, when and When the conditions are not met, computing node 10 outputs a nonlinear emergency fast shutdown mode; if the conditions are not met... and Under concurrent conditions, the computing node 10 outputs an exponential water hammer protection slow-closing mode, causing the main valve 21 to gradually reduce its opening degree according to the principle of prioritizing kinetic energy dissipation.

[0140] When the near-end severe rupture boundary condition is met, the computation node 10 triggers the nonlinear emergency fast-closing mode, controlling the main valve 21 to close according to the maximum mechanical constraint and simultaneously cutting off the fluid interaction of the bypass 30.

[0141] When computing node 10 determines that the input parameters satisfy and Under concurrent conditions, the system identifies a near-end severe rupture and ejection condition. Under this condition, the safety hazard caused by the large-volume fluid leakage is assessed as higher than the water hammer risk generated by the system. The macroscopic execution thread of computing node 10 issues a nonlinear emergency fast-closing command to actuator 22.

[0142] Let the start time of the secondary valve closing procedure be... Main valve 21 in The position coordinates at that moment are the positions of the primary interception boundary. In the nonlinear emergency fast-closing mode, computation node 10 controls the main valve 21 to close with a set maximum tangential acceleration. The mathematical model of the drive trajectory of the main valve 21 is expressed as:

[0143] ;

[0144] In the formula, Indicates that the main valve 21 is in Real-time displacement distance at any given moment. This indicates the target coordinates reached by the main valve 21 in the initial stage. This indicates the maximum permissible mechanical acceleration of actuator 22 under current hydraulic or electrical power conditions. The function is used to define the mechanical zero point of the main valve 21 to prevent position commands from exceeding limits due to over-calculation. When When the calculated value decays to 0, the main valve 21 achieves mechanical blocking.

[0145] During this process, the near-end fluid ejection alters the high-pressure flow field distribution in the upstream stable region. To prevent structural damage to the pilot valve 31 due to severe cavitation oscillations in the flow field, the micro-excitation thread of computation node 10 is... A forced closure signal is sent to the pilot valve 31 at all times. The actuator 22 overcomes the hydrodynamic pressure to forcibly cut off the flow area of ​​the main valve 21, and the bypass 30 is in a physically isolated state. The system reduces the risk of continued leakage of the medium through this force boundary control logic.

[0146] In dissatisfaction and When the near-end severe rupture boundary condition is met, the calculation node 10 triggers the exponential water hammer protection slow closure mode dominated by kinetic energy dissipation, and dynamically solves the attenuation parameters to complete the slow closure of the main valve 21.

[0147] When computing node 10 determines that the input parameters do not meet the concurrent conditions for a severe near-end rupture, the system identifies it as a non-severe near-end rupture condition. Non-severe near-end rupture conditions include severe mid-to-far-end rupture conditions, mid-to-far-end leakage conditions, and moderate to minor damage conditions. Under this condition, the pipeline network has sufficient spatial depth to buffer fluid kinetic energy, providing the physical prerequisite for executing the water hammer suppression algorithm. Computing node 10 triggers the exponential water hammer protection slow-closing mode.

[0148] Based on the principle of fluid compressibility and the acoustic characteristics of the pipeline network, computation node 10 constructs a mathematical model for the exponential decay trajectory. The drive trajectory of the main valve 21 in this mode is represented as follows:

[0149] ;

[0150] In the formula, This represents the attenuation control parameter dynamically solved by computation node 10, used to adjust the deceleration and valve closing curvature of actuator 22 in the near-zero opening range. Regarding this attenuation control parameter... Node 10 performs real-time calculations based on the principle of fluid kinetic energy conversion and dissipation ratio, and its mathematical relationship satisfies:

[0151] ;

[0152] In the formula, This represents the fluid buffer damping coefficient, which is related to the kinematic viscosity of the fluid medium and the roughness of the pipe wall. For real-time calculation of physical distance, For adaptive wave velocity, this formula dynamically links the remaining depth of the pipeline network to the wave velocity; the farther the distance and the higher the wave velocity, the gentler the curvature when the valve cuts off the flow.

[0153] Considering the asymptotic properties of the exponential function itself, and to prevent the actuator 22 from falling into an inefficient infinite fine-tuning dead zone, a mechanical shutdown dead zone threshold is preset inside the computation node 10. This threshold is typically set to 1% to 3% of the maximum mechanical stroke of the main valve 21. When the system determines... At that time, computing node 10 exits the exponential decay calculation and directly issues a zero-set shutdown command to the execution mechanism 22.

[0154] The exponential water hammer protection slow-closing mode utilizes the dynamic compressibility properties of the fluid medium. Computation node 10 dynamically distributes parameters... The control command causes the main valve 21 to gradually reduce its throttling velocity as it approaches zero opening. The residual fluid kinetic energy within the pipeline system is converted into elastic deformation energy of the pipeline wall and internal fluid energy within a relatively long throttling time window. This control process helps to mitigate the amplitude of local overpressure shock waves caused by diffuse water hammer kinetic energy, allowing the system to complete the final valve shut-off action while protecting the overall structure of the pipeline network.

[0155] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An automatic response emergency shut-off valve system for fluid leakage based on adaptive control, characterized in that, include: The main valve is installed in the main branch of the fluid pipeline network; The actuator is connected to the main valve via a drive mechanism; A bypass is provided, connecting the fluid pipelines upstream and downstream of the main valve, and the bypass forms an injection port in the fluid pipeline downstream of the main valve. A pilot valve is located on the bypass and is connected to the actuator via a transmission connection; A pressure sensor is installed in the fluid pipeline downstream of the injection port; The computing nodes are communicatively connected to the actuator, the pilot valve, and the pressure sensor, respectively. The computing node controls the actuator to drive the main valve to perform a primary flow throttling action to determine the primary flow throttling boundary position; The computing node drives the pilot valve to generate an orthogonally encoded transient pressure wave, which is injected into the single-phase pure liquid fluid downstream of the main valve via the bypass. The computing node acquires the transient pressure time-series signal collected by the pressure sensor, and when it determines that a local phase change has occurred in the fluid, it calls the fluid phase change boundary correction model to output adaptive wave velocity data. The computing node extracts the net echo signal from the transient pressure time series signal based on cepstral analysis, and calculates the actual physical distance of the leak point and the equivalent cross-sectional area of ​​the leak hole by combining the adaptive wave velocity data. The computing node controls the actuator to adjust the secondary deceleration and valve closing trajectory of the main valve based on the actual physical distance and the equivalent leakage cross-sectional area.

2. The automatic response emergency shut-off valve system for fluid leakage based on adaptive control according to claim 1, characterized in that, The axial physical distance between the injection port and the central axis of the main valve satisfies the spatial constraint condition: The axial physical distance is greater than the maximum extension length boundary of the main valve cavitation collapse zone calculated based on fluid dynamics, and less than or equal to the preset upper limit of the effective detection distance. Through this spatial constraint, the orthogonal encoded transient pressure wave is injected spatially, bypassing the cavitation collapse zone behind the main valve.

3. A fluid leakage automatic response emergency shut-off valve system based on adaptive control according to claim 1, characterized in that, The fluid phase transition boundary correction model that the computing node determines and invokes to identify a local phase transition in the fluid specifically includes: When the real-time static pressure data collected by the pressure sensor is continuously less than or equal to the standard saturated vapor pressure of the fluid medium at the current operating temperature within a preset continuous sampling period, it is determined that the local fluid in the pipeline network has crossed the critical point of phase change. The computing node derives the dynamic gas content based on the fluid medium's equation of state: the difference between the standard saturated vapor pressure and the real-time static pressure data is divided by the standard saturated vapor pressure, the result is multiplied by the gas phase precipitation expansion coefficient, and finally the initial dissolved gas volume free rate is added to obtain the dynamic gas content.

4. A fluid leakage automatic response emergency shut-off valve system based on adaptive control according to claim 3, characterized in that, The numerical correction logic for the adaptive wave velocity data output by the computing node is as follows: The first component is obtained by dividing the dynamic gas content by the product of the density of the precipitated gas phase and the square of the gas phase sound velocity; the second component is obtained by dividing the difference between the dynamic gas content and the density of the single-phase pure liquid by the product of the square of the liquid phase sound velocity; the first and second components are added together, multiplied by the comprehensive density of the mixed medium calculated from the dynamic gas content, and the reciprocal of the square root of the multiplication result is taken to obtain the adaptive wave velocity data.

5. A fluid leakage automatic response emergency shut-off valve system based on adaptive control according to claim 1, characterized in that, The specific process by which the computing node calculates the actual physical distance to the leak point includes: The computing node performs cross-correlation integration between the transient pressure time-series signal and the original transmission sequence to extract the net echo signal; The net echo signal is sequentially subjected to fast Fourier transform, logarithm, and inverse Fourier transform to generate a cepstral sequence, and the round-trip main phase delay time corresponding to the main peak that is greater than the validity determination threshold is extracted. The actual physical distance is calculated by combining the adaptive wave velocity data: the adaptive wave velocity data is integrated over a time interval from the initial trigger time to the sum of the initial trigger time and the round-trip main phase delay time, and the integration result is divided by two to obtain the actual physical distance.

6. A fluid leakage automatic response emergency shut-off valve system based on adaptive control according to claim 5, characterized in that, The specific process by which the computing node calculates the equivalent orifice cross-sectional area includes: Extract the absolute amplitude of the main peak of the net echo signal and perform natural damping compensation calculation in combination with the actual physical distance to obtain the reflected wave amplitude attenuation coefficient; Using the reflected wave amplitude attenuation coefficient and the current steady-state static pressure value of the pipeline as input indices, and substituting them into the preloaded leakage cross-sectional area mapping matrix, the equivalent leakage cross-sectional area under the current operating condition is obtained through two-dimensional linear interpolation.

7. A fluid leakage automatic response emergency shut-off valve system based on adaptive control according to claim 1, characterized in that, The specific process by which the computing node adjusts the secondary deceleration valve closing trajectory of the main valve includes: The computing node uses the calculated actual physical distance and equivalent leak cross-sectional area as coordinate variables and inputs them into the two-dimensional fluid dynamics control decision matrix. When the actual physical distance is not greater than the preset near-end distance threshold and the equivalent leak cross-sectional area is not less than the preset severe rupture area threshold, the nonlinear emergency fast closure mode is triggered; if the above concurrent conditions are not met, the exponential water hammer protection slow closure mode is triggered.

8. A fluid leakage automatic response emergency shut-off valve system based on adaptive control according to claim 7, characterized in that, In the nonlinear emergency fast shutdown mode: The computing node controls the main valve to close with maximum tangential acceleration. The calculation logic for its drive trajectory is as follows: calculate the time difference between the real-time time and the valve closing start time; multiply half of the maximum allowable mechanical running acceleration by the square of the time difference to obtain the displacement attenuation; subtract the displacement attenuation from the primary throttling boundary position to obtain the theoretical displacement stroke; take the maximum value between the theoretical displacement stroke and zero as the real-time displacement stroke of the main valve. Simultaneously, the computing node sends a forced shutdown signal to the pilot valve to cut off the bypass.

9. A fluid leakage automatic response emergency shut-off valve system based on adaptive control according to claim 7, characterized in that, In the exponential water hammer protection slow-closing mode: The computing node controls the main valve to close according to an exponential decay trajectory. Its calculation logic is as follows: multiply the time difference between the real time and the valve closing start time by the dynamically solved decay control parameter, take the opposite number as the exponent, and raise it to the power of the natural constant; multiply the power result by the position of the primary throttling boundary to obtain the real-time displacement stroke of the main valve.

10. A fluid leakage automatic response emergency shut-off valve system based on adaptive control according to claim 9, characterized in that, The real-time calculation logic of the attenuation control parameters satisfies: The attenuation control parameter is obtained by dividing the fluid buffer damping coefficient by the product of the actual physical distance and the adaptive wave velocity data. Furthermore, when the real-time displacement stroke of the main valve is less than or equal to the preset mechanical shut-off dead zone threshold, the computing node exits the exponential decay calculation and issues a zero-set shut-off command to the actuator.

Images