A differential pressure balance control method and system for a shutoff guide device

Abstract

The application relates to the technical field of industrial control, and provides a differential pressure balance control method and system for a plugging and draining device, which comprises the following steps: obtaining a pressure signal of the plugging and draining device, and performing segmented processing on the pressure signal; extracting characteristic parameters reflecting the fluid state from the pressure signal after the segmented processing; judging a flow mode in which the fluid is located according to the characteristic parameters; and adjusting a differential pressure balance control strategy according to the flow mode, wherein: when the flow mode is a stable state, a conventional differential pressure feedback adjustment mode is adopted; when the flow mode is an unstable state, a trend estimation is performed on the pressure signal to obtain an overall differential pressure change trend, and the opening of a control valve is adjusted according to the overall differential pressure change trend, while the adjustment speed of the control valve is limited, and the minimum opening of the control valve is maintained. The application can avoid misjudgment and excessive adjustment caused by instantaneous fluctuation of a traditional control system.

Description

Technical Field

[0001] This application relates to the field of industrial control technology, and more specifically, to a differential pressure balance control method and system for a sealing and drainage device. Background Technology

[0002] In underground engineering, sealing and drainage devices are crucial equipment for responding to sudden water inrushes. They discharge gushing water through a central drainage pipe while simultaneously using a sealing body to prevent the water flow from spreading. Maintaining pressure balance inside and outside the drainage pipe is essential to ensure the stability and sealing effectiveness of the device. Traditional pressure balance control systems typically rely on real-time pressure sensor data to adjust valve openings for controlled water discharge. However, in complex underground environments, especially when the water flow contains dissolved gases, traditional control methods face significant challenges.

[0003] In deep underground engineering, high-pressure water inrush accidents are often accompanied by groundwater rich in dissolved gases. When this gas-bearing water is discharged through the central drain pipe of the sealed drainage device, dissolved gases are released as the pressure decreases, forming a complex liquid-gas two-phase flow. This two-phase flow is highly unstable, with drastic changes in fluid density and compressibility, causing violent and nonlinear fluctuations in the flow velocity and local pressure within the drain pipe in a very short time, exhibiting unstable flow states such as pulsating flow and slug flow.

[0004] Traditional pressure sensors, operating in a two-phase flow environment, no longer capture steadily changing static or dynamic pressure. Instead, they collect high-frequency noise caused by bubble bursts, gas-liquid interface impacts, drastic flow regime changes, and the alternating passage of gas and liquid clumps. The amplitude of this noise can be very large, even far exceeding the actual overall water pressure variation trend. For example, when a large bubble passes through the sensor, the gas density is much lower than the liquid density, causing a sharp drop in the instantaneous pressure reading. The subsequent liquid clump then causes the pressure to rise rapidly, creating a sharp pulse signal. This signal is characterized by high frequency, strong randomness, and large instantaneous amplitude, completely different from the actual, controllable overall pressure difference variation trend.

[0005] Faced with such high-frequency, drastically fluctuating pressure signals, traditional differential pressure balance control systems cannot effectively distinguish between these instantaneous pressure fluctuations caused by air bubbles and the actual, overall water pressure changes on both sides of the sealing body. The control system misinterprets the instantaneous pressure drop caused by air bubbles as an excessively low differential pressure, thus instructing the valve to close urgently; conversely, when the air bubble passes through and the pressure rapidly rises, it is misinterpreted as an excessively high differential pressure, causing the valve to open urgently. This misinterpretation and overreaction cause the control system to completely lose effective control over the actual differential pressure, falling into a state of high-frequency, repeated over-opening and over-closing oscillations. This frequent valve adjustment not only fails to stabilize the differential pressure but also exacerbates the turbulence of the fluid in the guide pipe, generating powerful hydraulic shock waves that cause continuous, high-intensity mechanical stress on the entire sealing device, accelerating the fatigue and aging of the sealing material, and may even lead to sealing failure and secondary water inrush disasters. Summary of the Invention

[0006] This application provides a differential pressure balance control method and system for a sealing and drainage device to solve at least one of the above-mentioned technical problems.

[0007] To achieve the above objectives, this application adopts the following technical solution: In a first aspect, this application discloses a differential pressure balance control method for a sealing and drainage device, comprising: Acquire the pressure signal of the sealing and drainage device, and process the pressure signal in segments; Extract characteristic parameters reflecting the fluid state from the segmented pressure signal; Determine the flow pattern of the fluid based on the characteristic parameters; Adjust the differential pressure balance control strategy according to the flow pattern, where: When the flow pattern is in a steady state, the conventional differential pressure feedback regulation method is adopted; When the flow mode is unstable, the pressure signal is trend estimated to obtain the overall differential pressure change trend. Based on the overall differential pressure change trend, the opening of the control valve is adjusted, while the adjustment speed of the control valve is limited and the minimum opening of the control valve is maintained.

[0008] Secondly, this application also discloses a differential pressure balance control system for a sealing and drainage device, comprising: The pressure signal acquisition module is used to acquire the pressure signal of the sealing and guiding device and to process the pressure signal in segments. The feature parameter extraction module is used to extract feature parameters reflecting the fluid state from the segmented pressure signal; The flow mode determination module is used to determine the flow mode of the fluid based on characteristic parameters. The control strategy adjustment module is used to adjust the differential pressure balance control strategy according to the flow pattern, wherein: When the flow pattern is in a steady state, the conventional differential pressure feedback regulation method is adopted; When the flow mode is unstable, the pressure signal is trend estimated to obtain the overall differential pressure change trend. Based on the overall differential pressure change trend, the opening of the control valve is adjusted, while the adjustment speed of the control valve is limited and the minimum opening of the control valve is maintained.

[0009] Compared with the prior art, this application has at least the following beneficial effects: This application enables stable control of pressure differential, ensuring the safety and reliability of the sealing and drainage device in complex underground environments and effectively preventing secondary water inrush disasters. Attached Figure Description

[0010] Figure 1 A schematic flowchart of a differential pressure balance control method for a plugging and draining device provided in this application; Figure 2 This is a schematic diagram of the differential pressure balance control system of a sealing and drainage device provided in this application. Detailed Implementation

[0011] The technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0012] like Figure 1 As shown in the figure, this application proposes a differential pressure balance control method for a plugging and draining device, including: Acquire the pressure signal of the sealing and drainage device, and process the pressure signal in segments; Extract characteristic parameters reflecting the fluid state from the segmented pressure signal; Determine the flow pattern of the fluid based on the characteristic parameters; Adjust the differential pressure balance control strategy according to the flow pattern, where: When the flow pattern is in a steady state, the conventional differential pressure feedback regulation method is adopted; When the flow mode is unstable, the pressure signal is trend estimated to obtain the overall differential pressure change trend. Based on the overall differential pressure change trend, the opening of the control valve is adjusted, while the adjustment speed of the control valve is limited and the minimum opening of the control valve is maintained.

[0013] This application effectively solves the problem of control method failure in liquid-gas two-phase flow environment by segmenting the pressure signal, extracting fluid state characteristic parameters, judging the flow mode, and adjusting the differential pressure balance control strategy according to different flow modes. In particular, it adopts trend estimation and limiting valve adjustment speed in unstable state, which significantly improves the stability and safety of the sealing and drainage device.

[0014] The sealing and drainage device refers to a device used in underground engineering to prevent the spread of water flow in the event of a water inrush accident, and to discharge the gushing water through a central drainage pipe. Its core function is to maintain pressure balance inside and outside the drainage pipe to ensure the stability of the device and the sealing effect. The pressure signal refers to the fluid pressure data collected in real time by pressure sensors installed on the sealing and drainage device; this data is the basis for subsequent analysis and control. Fluid state characteristic parameters are key indicators extracted from the original pressure signal that reflect the physical properties and flow behavior of the fluid (especially liquid-gas two-phase flow), such as fluctuation degree, frequency composition, skewness, and kurtosis. The flow mode refers to the macroscopic flow pattern exhibited by the fluid within the drainage pipe, such as stable single-phase flow, pulsating flow, slug flow, and other unstable two-phase flows. The differential pressure balance control strategy refers to the specific methods and rules for adjusting the opening of control valves to maintain the differential pressure balance inside and outside the drainage pipe according to different fluid flow modes.

[0015] The differential pressure balance control method for the sealing and drainage device proposed in this application will be described in further detail below: Specifically, this method first requires acquiring the pressure signal from the sealing and drainage device and then segmenting the pressure signal. The pressure signal can be acquired by installing high-precision pressure sensors inside and outside the drainage pipe; these sensors can monitor fluid pressure changes in real time. The purpose of segmentation is to divide the continuous pressure signal into several time periods to facilitate subsequent feature extraction and analysis. For example, 1000 pressure data points collected per second can be divided into 10 time periods, each containing 100 data points. This segmentation can be achieved using a sliding window method, which sets a fixed-length time window and slides it across the entire pressure signal sequence in steps, forming a new data segment with each slide. Another approach is event-based segmentation; for example, when a statistical characteristic of the pressure signal (such as the mean or variance) changes significantly, it is used as the starting point of a new time period.

[0016] After processing the segmented pressure signal, it is necessary to extract characteristic parameters reflecting the fluid state. These characteristic parameters are key to identifying fluid flow patterns. For example, statistical characteristics of the pressure signal within each segment can be calculated, such as mean, variance, standard deviation, skewness, and kurtosis. The mean reflects the average pressure level, while variance and standard deviation reflect the degree of pressure fluctuation. Skewness measures the symmetry of the pressure distribution, and kurtosis reflects the sharpness of the pressure distribution. Furthermore, frequency domain analysis methods such as Fourier transform can be used to extract the main frequency components of the pressure signal and their corresponding energy intensities to identify periodic fluctuations or noise at specific frequencies. For example, if the variance of the pressure signal is large and there are multiple significant frequency peaks within a certain segment, it may indicate that the fluid is in an unstable state.

[0017] Subsequently, based on the extracted feature parameters, the flow mode of the fluid is determined. This step is the intelligent core of the entire control method. For example, a rule-based expert system can be established, with preset thresholds and combination rules for a series of feature parameters. When the extracted feature parameters meet specific rules, the corresponding flow mode is determined. For instance, if the fluctuation of the pressure signal (such as the standard deviation) is below a certain threshold, and the main frequency components are concentrated in the low-frequency region, it can be judged as a stable state. Conversely, if the fluctuation is high and there are multiple high-frequency components, it may be judged as an unstable state. Another approach is to use machine learning models, such as support vector machines (SVM) or neural networks. By collecting a large amount of pressure signal data under different flow modes and labeling their corresponding flow modes, a classifier is trained. In practical applications, the extracted feature parameters are input into the trained model, and the model can output the current flow mode of the fluid.

[0018] Finally, the differential pressure balance control strategy is adjusted based on the determined flow pattern. This is a crucial step in achieving intelligent control. When the fluid flow pattern is determined to be stable, a conventional differential pressure feedback adjustment method can be used. In this mode, the control system directly adjusts the opening of the control valve based on the deviation between the real-time differential pressure inside and outside the guide tube and the set target differential pressure, using classic control algorithms such as PID controllers to eliminate the differential pressure deviation. For example, if the differential pressure inside the guide tube is higher than the target value, the valve opening is increased to reduce the internal pressure; conversely, the valve opening is decreased. This method provides precise and rapid response control when the fluid state is stable.

[0019] When the fluid flow pattern is determined to be unstable, traditional differential pressure feedback control methods are no longer applicable and may even exacerbate system instability. In this case, this application proposes a more robust control strategy: trend estimation of the pressure signal to obtain the overall differential pressure change trend, and adjustment of the control valve opening based on this trend, while limiting the valve's adjustment speed and maintaining its minimum opening. Trend estimation can be achieved through various methods, such as moving average filtering, Kalman filtering, or wavelet analysis. Moving average filtering can smooth out high-frequency noise and reveal the underlying slow-changing trend. Kalman filtering can combine system models and measurement data to make optimal estimates of the system state, thereby more accurately predicting the future trend of the differential pressure. Wavelet analysis can decompose the pressure signal at different scales, separating the low-frequency components representing the overall trend. These methods effectively filter out instantaneous and drastic fluctuations caused by unstable factors such as bubbles, obtaining a trend that truly reflects the overall differential pressure change across the sealing body. For example, even if the instantaneous pressure drops sharply due to the passage of air bubbles, the control system will not mistakenly interpret the pressure difference as too low and shut off the valve in an emergency if the trend estimate shows that the overall pressure difference is still rising.

[0020] When adjusting the opening of the control valve according to the overall differential pressure change trend, it is also necessary to limit the adjustment speed of the control valve. This means that the valve cannot be adjusted drastically in a short period of time to avoid causing new instability due to overreaction. For example, the maximum opening change per second can be set to ensure smooth valve operation. In addition, it is necessary to maintain the minimum opening of the control valve. Even in unstable conditions, to prevent complete blockage of the guide pipe or sudden pressure rise, the valve should always maintain a certain opening, allowing a small amount of fluid to pass through, thereby providing a certain buffer and pressure relief capacity for the system. For example, the minimum valve opening can be set to 5%, so that even in extreme cases, the valve will not be completely closed.

[0021] The differential pressure balance control method for the sealing and drainage device proposed in this application effectively solves the problem of failure in complex liquid-gas two-phase flow environments through the synergistic effect of the aforementioned technical features. Firstly, by segmenting the pressure signal and extracting characteristic parameters, the system can identify key information truly reflecting the fluid state from the noisy raw signal. For example, in gas-containing water flow, traditional methods misjudge instantaneous pressure fluctuations caused by bubbles as overall differential pressure changes, leading to frequent valve malfunctions. However, this application, by extracting characteristic parameters such as the degree of fluctuation and frequency components, can distinguish between such instantaneous fluctuations and the true overall differential pressure changes.

[0022] Secondly, determining the flow pattern of the fluid based on the extracted characteristic parameters is a crucial step in achieving intelligent control. When the system identifies the fluid as being in a stable state, the conventional differential pressure feedback adjustment method can ensure control accuracy and response speed. However, when the system identifies the fluid as being in an unstable state, it no longer blindly performs real-time feedback adjustment, but switches to a more robust control strategy.

[0023] Specifically, in unstable mode, trend estimation of the pressure signal can effectively filter out high-frequency noise and obtain the overall pressure difference change trend. For example, through moving average filtering or Kalman filtering, even if a single pressure reading fluctuates drastically due to the passage of air bubbles, the system can identify the long-term upward or downward trend of the pressure difference. Based on this trend, the opening of the control valve is adjusted to avoid overreacting to instantaneous fluctuations. At the same time, limiting the adjustment speed of the control valve prevents high-frequency oscillations of the valve and reduces mechanical shock to the device. Maintaining the minimum opening of the control valve provides a safety guarantee for the system, preventing complete blockage of the guide pipe and potential catastrophic consequences.

[0024] Overall, this application achieves adaptive control of complex fluid environments through a closed-loop control logic of perception-judgment-decision-execution. Instead of treating all pressure signals indiscriminately, it dynamically adjusts the control strategy based on the inherent characteristics of the fluid state, thereby significantly improving the operational stability and safety of the sealing and drainage device under harsh conditions while ensuring pressure differential balance.

[0025] This application constructs an intelligent sensing layer by introducing three steps: segmenting the pressure signal, extracting characteristic parameters reflecting the fluid state, and determining the flow mode of the fluid based on the characteristic parameters. This layer can deeply analyze the intrinsic structure of the pressure signal and identify whether the fluid is in a stable single-phase flow state or an unstable liquid-gas two-phase flow state (such as pulsating flow or slug flow). For example, in a traditional system, a large bubble cluster passing through the sensor may cause a sudden and sharp drop in instantaneous pressure, and the system will immediately command the valve to close. However, this application analyzes the fluctuation degree and frequency components of the pressure signal to determine that this is merely a local instability phenomenon, rather than a true change in the overall pressure difference.

[0026] Building upon this, this application further proposes a scheme for adjusting the differential pressure balance control strategy based on the flow pattern. When a steady state is identified, efficient conventional differential pressure feedback regulation is used; however, when an unstable state is identified, a more robust control strategy is adopted. This involves estimating the trend of the pressure signal to obtain the overall differential pressure change trend, and adjusting the opening of the control valve according to this trend while limiting the valve's adjustment speed and maintaining its minimum opening. This strategy effectively avoids overreaction and oscillation problems in unstable flow conditions. Through trend estimation, the system can detect instantaneous noise and grasp the true trend of the overall differential pressure; by limiting the adjustment speed and maintaining the minimum opening, stability and safety are guaranteed for the system.

[0027] In some embodiments, the step of determining the flow mode of the fluid based on characteristic parameters includes: Acquire the pressure signal of the sealing and drainage device, and perform preliminary processing on the pressure signal; The pre-processed pressure signal is decomposed into multiple scale components to obtain multiple scale components. Extracting component characteristics from multiple scale components; Determine the flow pattern of the fluid based on its component characteristics.

[0028] The acquisition and preliminary processing of pressure signals from the sealing and drainage device involves real-time acquisition of pressure data from within the device via sensors, followed by preprocessing of this raw data, such as noise reduction, filtering, or data smoothing, to eliminate measurement errors and transient interference, ensuring data quality for subsequent analysis. The purpose of this preliminary processing is to provide a cleaner and more reliable pressure signal that accurately reflects the true state of the fluid.

[0029] Multi-scale decomposition of the pre-processed pressure signal yields multiple scale components. This can be understood as decomposing the pre-processed pressure signal at different time or frequency scales, thereby revealing the hidden fluctuation patterns within the signal across different frequency ranges. For example, wavelet decomposition, empirical mode decomposition (EMD), or variational mode decomposition (VMD) can be used to decompose the original pressure signal into a series of sub-signals or modal components with different frequency characteristics. These scale components represent the dynamic behavior of the fluid at different time scales; for instance, high-frequency components may reflect instantaneous disturbances, while low-frequency components may reflect macroscopic trends.

[0030] Extracting component characteristics from multiple scale components refers to analyzing the specific properties of the fluid component represented by each scale component. For example, statistical or kinetic characteristics such as energy, frequency, amplitude, phase, and entropy can be extracted for each scale component. These component characteristics can quantify the information carried by different scale components, thereby reflecting the existence, proportion, interaction, and dynamic behavior of different components (such as gas, liquid, and solid phases) in the fluid.

[0031] Determining the flow mode of a fluid based on its component characteristics involves comprehensively analyzing the component properties extracted from various scales and using pre-established identification models or rules to categorize the current fluid state into a specific flow mode. For example, by analyzing the energy distribution, frequency characteristics, and interrelationships of different scale components, it is possible to identify whether the fluid is in a specific mode, such as laminar flow, turbulent flow, slug flow, or annular flow, within single-phase or multiphase flow (e.g., gas-liquid two-phase flow, liquid-solid two-phase flow). The purpose is to provide accurate flow state information for subsequent differential pressure balance control strategies.

[0032] This application ensures data quality by performing preliminary processing on the pressure signal from the sealing and drainage device. Subsequently, multi-scale decomposition breaks down the complex pressure signal into multiple scale components, clearly revealing the dynamic behavior of the fluid at different time scales. Extracting the component characteristics of each scale component allows for precise quantification of the properties and behaviors of each component in the fluid. Finally, based on these comprehensive component characteristics, the flow mode of the fluid can be accurately determined. This hierarchical and refined analysis method enables a more comprehensive and accurate identification of the fluid state, thus providing a more reliable basis for subsequent differential pressure balance control strategies.

[0033] The above technical solutions overcome potential limitations in complex fluid flow pattern recognition. By introducing multi-scale decomposition and component characteristic extraction, the analysis of pressure signals becomes more in-depth and detailed, effectively capturing the dynamic characteristics of different components in the fluid at different time scales. Consequently, the accuracy and robustness of fluid flow pattern judgment are significantly improved, providing a more precise decision-making basis for the differential pressure balance control of the sealing and drainage device, thereby enhancing the adaptability and stability of the control system.

[0034] In some embodiments, the step of extracting characteristic parameters reflecting the fluid state from the segmented pressure signal may include: For the segmented pressure signal, extract the fluctuation degree, main frequency components, skewness and kurtosis of the pressure signal; Cross-correlation analysis was performed on the fluctuation degree, main frequency components, skewness, and kurtosis of the extracted pressure signal; Based on the results of cross-correlation analysis, fluid component identification rules are constructed; Based on fluid component identification rules, the presence and contribution ratio of different components in the fluid are determined as characteristic parameters reflecting the fluid state.

[0035] Specifically, after acquiring and segmenting the pressure signal from the sealing and drainage device, these segmented pressure signals need to be analyzed in depth to identify the internal state of the fluid. Extracting the fluctuation degree, main frequency components, skewness, and kurtosis of the pressure signal are key steps in quantifying its characteristics. Fluctuation degree can be understood as the range or intensity of pressure signal variation over a period of time, reflecting the severity of fluid disturbance. The main frequency components reveal the dominant periodic changes in pressure fluctuations, which may be related to the vibration or mixing characteristics of specific fluid components. Skewness measures the symmetry of the pressure signal distribution; positive or negative skewness may indicate the asymmetric influence of a certain component in the fluid. Kurtosis reflects the sharpness or tail thickness of the pressure signal distribution; high kurtosis may indicate the presence of abnormal pressure pulses or transient events.

[0036] Cross-correlation analysis was performed on the fluctuation level, main frequency components, skewness, and kurtosis of the extracted pressure signals. This analysis aims to explore the interrelationships and dependencies among these different characteristic parameters. For example, by analyzing the correlation between fluctuation level and main frequency components, it can be determined whether the intensity of fluid disturbance is associated with a specific vibration mode; by analyzing the correlation between skewness and kurtosis, a more comprehensive understanding of the non-Gaussian nature of the pressure signal can be obtained. This analysis helps to reveal the complex interaction modes between fluid components.

[0037] Based on the results of cross-correlation analysis, fluid component identification rules are constructed. These rules can be combinations of statistical models, machine learning algorithms, or expert experience, used to map specific combinations of feature parameters to specific fluid component types and their states. For example, high volatility, the presence of specific low-frequency components, and a positive skewness may indicate the presence of bubble aggregation in a gas-liquid two-phase flow.

[0038] Based on the constructed fluid component identification rules, the presence and contribution ratio of different components in the fluid are determined. For example, it can identify whether the fluid is a single-phase liquid, a single-phase gas, a gas-liquid two-phase flow, or a multiphase flow containing solid particles, and quantify the proportion of each component in the overall fluid. This information is integrated as characteristic parameters reflecting the fluid state, providing accurate input for subsequent determination of the fluid's flow pattern.

[0039] This application achieves a more comprehensive and in-depth understanding of fluid dynamics by performing multi-dimensional feature extraction and cross-correlation analysis on segmented pressure signals. Basic schemes may only focus on pressure averages or simple fluctuations, failing to capture the complex pressure changes caused by different fluid components. By extracting the degree of fluctuation, main frequency components, skewness, and kurtosis, the pressure signal can be finely characterized in both the time and frequency domains. Cross-correlation analysis reveals the intrinsic relationships between these features, aiding in the identification of interactions between fluid components. Therefore, the constructed fluid component identification rules can accurately determine the presence and contribution ratio of each component in the fluid based on these comprehensive features, providing a more reliable and refined basis for subsequent flow pattern judgment and avoiding misjudgments due to insufficient feature information.

[0040] Through the above technical solution, this application can significantly improve the accuracy and robustness of fluid state identification within the sealing and drainage device. Compared to methods relying on only a single or limited number of features, this solution comprehensively analyzes the fluctuation degree, main frequency components, skewness, and kurtosis of the pressure signal and performs cross-correlation, enabling a more comprehensive capture of the complex pressure dynamics caused by different components (such as gas, liquid, and solid) in the fluid. This multi-dimensional and in-depth feature extraction and analysis makes the construction of fluid component identification rules more precise, thereby enabling a more accurate determination of the presence and contribution ratio of different components in the fluid. This provides a more refined and reliable input for subsequent flow pattern determination, effectively avoiding deviations in the differential pressure balance control strategy caused by inaccurate fluid state identification, and thus improving the stability and efficiency of differential pressure balance control in the sealing and drainage device.

[0041] In some embodiments, the step of extracting characteristic parameters reflecting the fluid state from the segmented pressure signal includes: Multi-scale energy normalization is performed on the segmented pressure signal to eliminate the differences in pressure fluctuation amplitude caused by different components. Energy intensity, fluctuation asymmetry, and local fluctuation frequency are extracted from the normalized pressure signal at different frequency scales. By comparing energy intensity at different frequency scales, the relative contribution of each component in the fluid is determined. Combined with wave asymmetry and local wave frequency, the existence and state of all components in the fluid are identified as characteristic parameters reflecting the fluid state.

[0042] Multi-scale energy normalization refers to decomposing the pressure signal into components at different time or frequency scales and standardizing the energy of these components. The aim is to eliminate the inherent amplitude differences in the pressure signal caused by different components in the fluid (e.g., gas, liquid, and solid phases), making the contributions of each component comparable across different scales and preventing a single component's large amplitude from masking the characteristics of other components. In practical applications, this processing can be achieved using signal processing techniques such as wavelet transform, empirical mode decomposition (EMD), or variational mode decomposition (VMD), by calculating and normalizing the energy of the decomposed modal components.

[0043] Extracting energy intensity, wave asymmetry, and local wave frequencies at different frequency scales involves in-depth analysis of the signal at each scale component or within a specific frequency range after normalization. Energy intensity reflects the activity or power of the signal at that frequency scale; wave asymmetry reveals the nonlinear characteristics of fluid behavior, such as the presence of transient shocks or non-uniform distributions; and local wave frequencies are used to identify the characteristic oscillation frequencies that specific fluid components may possess. These parameters can be extracted using methods such as Fourier transform, power spectral density analysis, or Hilbert transform.

[0044] By comparing energy intensities at different frequency scales, the relative contributions of each component in a fluid can be determined. For example, if the energy intensity at a certain frequency scale is significantly higher than at other scales, it may indicate that the fluid component or flow phenomenon corresponding to that frequency scale is dominant. Furthermore, combining wave asymmetry and local wave frequencies allows for a more comprehensive identification of the presence and states of all components in the fluid. For instance, specific combinations of energy intensity distributions, wave asymmetry patterns, and local wave frequencies can serve as fingerprint characteristics for identifying specific fluid components (such as bubbles, droplets, and solid particles) and their flow states (such as laminar, turbulent, and slug flow).

[0045] This application effectively addresses the problem of insufficient recognition accuracy caused by large differences in the fluctuation amplitudes of different components when processing multi-component fluid pressure signals by introducing multi-scale energy normalization processing. By extracting energy intensity, fluctuation asymmetry, and local fluctuation frequencies from the normalized signal at different frequency scales, the dynamic characteristics of each fluid component can be captured more precisely. Therefore, by comprehensively comparing these multi-scale features, the relative contribution of each component can be accurately determined, and its existence and specific state can be identified, thus providing a more reliable and comprehensive input for subsequent flow pattern determination.

[0046] The above technical solution significantly improves the accuracy and robustness of extracting fluid state characteristic parameters within the sealing and drainage device. Especially in multi-component fluids or complex flow modes, this method effectively eliminates the differences in pressure fluctuation amplitude caused by different components, allowing the characteristics of each component to be clearly presented, thus achieving accurate identification of fluid components and their states. This provides a more reliable basis for subsequent flow mode judgment, thereby making the adjustment of the differential pressure balance control strategy more precise and timely, effectively avoiding control deviations caused by misjudgment of fluid states, and improving the stability and efficiency of the entire control system.

[0047] Assuming a two-phase gas-liquid fluid exists within the sealing and drainage device, and that the gas and liquid phases exhibit different fluctuation amplitudes and frequency characteristics in the pressure signal, the following steps are taken. First, the acquired pressure signal is segmented and then decomposed into multiple scales using wavelet transform. Energy calculations and normalization are performed on the wavelet coefficients at each scale to eliminate the differences in pressure fluctuation amplitudes caused by the gas and liquid phases. Subsequently, for each normalized scale component, its energy intensity, skewness, and kurtosis (reflecting fluctuation asymmetry), as well as the dominant frequency component (reflecting local fluctuation frequencies), are calculated. For example, significant energy intensity and near-zero skewness at higher frequency scales may indicate the presence of the gas phase, while higher energy intensity and higher kurtosis at lower frequency scales may indicate the presence of the liquid phase or the formation of a liquid plug. By comparing these multi-scale characteristics, the relative contributions of the gas and liquid phases can be accurately determined, and the specific state of the current fluid—whether it is bubbly flow, slug flow, or annular flow—can be identified, thus providing refined fluid state information for the differential pressure balance control system.

[0048] In some embodiments, the step of extracting characteristic parameters reflecting the fluid state from the segmented pressure signal includes: Nonlinear characteristic transformation is performed on the segmented pressure signal to enhance the pressure fluctuation characteristics caused by nonlinear interactions between fluid components; Energy intensity, fluctuation asymmetry, and local fluctuation frequency are extracted from the pressure signal after nonlinear characteristic transformation at multiple frequency scales. By analyzing the interaction between energy intensity, wave asymmetry, and local wave frequencies at different frequency scales, the cooperative interaction modes of various components in the fluid and their contribution to the overall fluid state are identified, serving as characteristic parameters reflecting the fluid state.

[0049] Specifically, nonlinear characteristic transformation refers to applying a nonlinear function or algorithm, such as wavelet transform, empirical mode decomposition (EMD), variational mode decomposition (VMD), or nonlinear dynamic analysis methods (such as phase space reconstruction and Lyapunov exponent calculation), to the original pressure signal. Its purpose is to make the nonlinear information in the original signal explicit, making it easier to capture and quantify in subsequent analysis, thereby enhancing the pressure fluctuation characteristics caused by nonlinear interactions between fluid components. Extracting energy intensity, fluctuation asymmetry, and local fluctuation frequencies at multiple frequency scales can be understood as using frequency domain analysis (such as Fourier transform and wavelet packet decomposition) or time-frequency analysis methods on the signal after nonlinear characteristic transformation to obtain the energy distribution, waveform deviation from symmetry, and instantaneous changes of specific frequency components within different frequency ranges. Energy intensity reflects the severity of signal fluctuations at a specific frequency scale; fluctuation asymmetry reveals the asymmetry of the signal waveform along the time axis, which is usually related to nonlinear dynamic behavior; local fluctuation frequencies focus on the frequency characteristics of the signal within a local time period, helping to identify transient or locally occurring nonlinear events. In practical applications, by analyzing the interactions between energy intensity, fluctuation asymmetry, and local fluctuation frequencies at different frequency scales, the synergistic interaction patterns of various components in a fluid and their contributions to the overall fluid state can be identified. Specifically, this involves establishing correlation models between these characteristic parameters or employing machine learning algorithms, such as Support Vector Machines (SVM), neural networks, or decision trees, to comprehensively analyze these multi-scale, nonlinear characteristics. For example, when a fluid component undergoes a phase transition, it may cause significant changes in energy intensity at a specific frequency scale, accompanied by an increase in fluctuation asymmetry and a drift in local fluctuation frequencies. By capturing these interrelated change patterns, the presence of different fluid components (such as gas, liquid, and solid phases), their relative proportions, and their complex synergistic interactions can be identified, such as the coalescence and breakup of bubbles, and the formation and dispersion of droplets, thereby more accurately characterizing the overall fluid state.

[0050] This application, by introducing nonlinear characteristic transformation, effectively separates and enhances the pressure fluctuation characteristics caused by complex nonlinear interactions between fluid components from the original signal, overcoming the limitations of linear methods in processing such complex signals. Furthermore, by extracting energy intensity, fluctuation asymmetry, and local fluctuation frequencies at multiple frequency scales, the analysis of fluid states is no longer limited to a single frequency or time dimension, but can comprehensively examine fluid dynamics from both macroscopic and microscopic perspectives. Through in-depth analysis of the interactions between these multi-scale, nonlinear characteristics, more refined cooperative interaction patterns among fluid components can be revealed, such as coupled oscillations of different phases and energy transfer paths. This allows for a more accurate determination of the fluid's flow pattern, especially under complex multiphase flow or rapidly changing conditions.

[0051] Through the above technical solution, this application can significantly improve the accuracy and comprehensiveness of extracting fluid state characteristic parameters within the plugging and drainage device. By fully considering and quantifying the nonlinear interactions between fluid components, the extracted characteristic parameters more realistically reflect the intrinsic dynamic characteristics of the fluid, avoiding the loss of characteristic information due to linear assumptions. Therefore, higher reliability can be achieved in determining fluid flow patterns, especially in scenarios with complex and variable fluid states and significant nonlinear effects, such as complex conditions like gas-liquid two-phase flow and sand-carrying flow that may be encountered during oil and gas well plugging and drainage. This more accurate extraction of characteristic parameters provides a more solid data foundation for subsequent adjustments to differential pressure balance control strategies, helping to achieve more refined and robust differential pressure control, thereby effectively avoiding control instability or inefficiency caused by misjudgment of flow patterns.

[0052] The following is a specific example to illustrate this.

[0053] Assume a two-phase gas-liquid flow exists in the sealing and drainage device, and that the coalescence and breakup of bubbles have a significant nonlinear impact on the pressure signal. Linear feature extraction methods may only be able to identify the average pressure or simple fluctuation amplitude, making it difficult to distinguish the deeper effects of bubble size, distribution, and their dynamic changes on the fluid state.

[0054] When employing the scheme of this application, the acquired segmented pressure signal is first subjected to nonlinear feature transformation, such as wavelet packet decomposition, to decompose the signal into different frequency sub-bands. The energy of each sub-band is then nonlinearly mapped or reconstructed to enhance the nonlinear fluctuation characteristics associated with bubble coalescence and breakup. Subsequently, energy intensity, fluctuation asymmetry, and local fluctuation frequencies are extracted from these nonlinearly transformed signals at multiple frequency scales. For example, at the low-frequency scale, energy intensity may reflect overall flow velocity and pressure pulsation; at the mid-frequency scale, fluctuation asymmetry may be related to the periodic coalescence and breakup of bubbles; and at the high-frequency scale, changes in local fluctuation frequencies may indicate the formation or disappearance of microbubbles.

[0055] By analyzing the interactions between characteristic parameters at different frequency scales, for example, when bubbles coalesce in large numbers, a significant increase in energy intensity may be observed in a specific mid-frequency band, along with a corresponding increase in fluctuation asymmetry, while the local fluctuation frequency in the high-frequency band may experience a brief decrease. By identifying this specific combination of characteristic patterns, the system can accurately determine the relative contributions of the gas and liquid phases in the fluid, as well as the impact of bubble dynamics (such as coalescence and breakup) on the overall fluid state, thereby identifying specific unsteady flow modes such as pulsating flow and slug flow. Based on these accurately identified characteristic parameters, the system can more intelligently adjust the differential pressure balance control strategy. For example, when it detects that violent bubble coalescence leads to rapid differential pressure fluctuations, it can take measures in advance to limit the adjustment speed of the control valve to avoid over-adjustment, thereby maintaining the stable operation of the system.

[0056] In some embodiments, the step of extracting characteristic parameters reflecting the fluid state from the segmented pressure signal includes: Real-time adaptive calibration of feature parameters is performed on the segmented pressure signal to dynamically adjust the parameter thresholds or recognition rules for feature extraction. For the adaptively calibrated pressure signal, extract the fluctuation degree, main frequency components, skewness and kurtosis of the pressure signal; Based on the fluctuation degree, main frequency components, skewness, and kurtosis of the extracted pressure signal, the presence and contribution ratio of different components in the fluid are determined, which serve as characteristic parameters reflecting the fluid state.

[0057] Specifically, real-time adaptive calibration of feature parameters refers to the system's ability to automatically adjust the parameter thresholds or recognition rules used for feature extraction based on real-time monitored changes in pressure signals. Its purpose is to ensure that the feature extraction process maintains high accuracy and robustness even when fluid conditions change. For example, when fluid composition or operating conditions change, fixed parameters may not accurately capture the new fluid characteristics, while the adaptive calibration mechanism can dynamically optimize these parameters to better match the current fluid conditions.

[0058] Dynamically adjusting the thresholds or recognition rules for feature extraction can be understood as the system no longer using preset fixed parameters for feature extraction, but instead automatically updating or optimizing these parameters based on the characteristics of the real-time data stream through algorithms. For example, the thresholds or model parameters used to identify the degree of fluctuation, main frequency components, skewness, and kurtosis can be adjusted in real time based on the statistical characteristics of the pressure signal (such as changes in mean and variance) or the changing trend of the frequency spectrum.

[0059] Extracting the fluctuation degree, main frequency components, skewness, and kurtosis of an adaptively calibrated pressure signal refers to the precise quantification of these statistical characteristics from the pressure signal under dynamically adjusted parameters or rules. Fluctuation degree reflects pressure instability; main frequency components reveal periodic or quasi-periodic fluctuation patterns; skewness describes the asymmetry of the pressure distribution; and kurtosis characterizes the sharpness or tail thickness of the pressure distribution. These parameters together constitute a comprehensive description of the fluid state.

[0060] Determining the presence and contribution ratio of different components in a fluid by analyzing the fluctuations, main frequency components, skewness, and kurtosis of the extracted pressure signal involves using these adaptively calibrated feature parameters and a pre-defined identification model or algorithm to analyze the different components (e.g., gas, liquid, and solid phases) that may exist in the fluid and their relative content or influence in the overall fluid. The aim is to provide more accurate and representative input for subsequent flow pattern determination.

[0061] This application effectively solves the problem of insufficient feature extraction accuracy of fixed parameters in complex and variable fluid environments by introducing a real-time adaptive calibration mechanism for feature parameters. Specifically, when the fluid composition or operating conditions within the sealing and drainage device change, the statistical characteristics and spectral features of the pressure signal also change accordingly. Fixed parameter extraction methods may fail to accurately capture these changes, resulting in extracted feature parameters that do not truly reflect the current fluid state. This solution, however, dynamically adjusts the parameter thresholds or recognition rules for feature extraction by real-time evaluation of the pressure signal's changing trend and intensity, ensuring that the feature extraction process always maintains optimal matching with the current fluid characteristics. Therefore, both small fluctuations in fluid component ratios and drastic changes in flow patterns can be more accurately identified and quantified, providing a more reliable basis for subsequent flow pattern judgment and thus improving the adaptability and robustness of the entire differential pressure balance control method.

[0062] Through the above technical solution, this application can significantly improve the adaptability and accuracy of the differential pressure balance control method for the sealing and drainage device under complex and variable operating conditions. Since the extraction of characteristic parameters no longer relies on static preset values ​​but can be dynamically adjusted according to the real-time fluid state, the system can accurately capture key characteristics even if the fluid composition properties change drastically, avoiding misjudgments or control lags caused by parameter mismatch. This allows the differential pressure balance control strategy to respond more promptly and accurately to changes in fluid state, thereby effectively maintaining the stable operation of the sealing and drainage device, reducing operational risks, and improving the reliability of the overall control system.

[0063] Suppose that during the operation of the sealing and drainage device, the fluid medium gradually changes from a single liquid phase to a gas-liquid two-phase flow, or the gas phase content in the two-phase flow changes significantly. In this case, the fluctuation characteristics, main frequency components, skewness, and kurtosis of the pressure signal will all change significantly. If fixed feature extraction parameters are used, it may be impossible to accurately identify this change in fluid composition, leading to misjudgment of the flow mode by the system and the adoption of an inappropriate differential pressure balance control strategy.

[0064] In this embodiment, the system continuously monitors the pressure signal of the sealing and drainage device. When a significant change in the statistical characteristics of the pressure signal (e.g., variance or skewness) is detected within a short period, the real-time characteristic parameter adaptive calibration module is triggered. This module dynamically adjusts the algorithm parameters or thresholds used to calculate the fluctuation level, main frequency components, skewness, and kurtosis based on the characteristics of the current pressure signal, such as by analyzing the differences between its short-term historical data and real-time data. For example, it can increase the sensitivity to high-frequency components or adjust the calculation window for skewness and kurtosis to better capture the irregular fluctuations unique to two-phase flow.

[0065] After adaptive calibration, even if the fluid composition changes, the extracted feature parameters can accurately reflect the current true state of the fluid. For example, when the gas phase content increases, the fluctuation of the pressure signal may increase, while the energy proportion of high-frequency components increases, and the skewness and kurtosis may also exhibit stronger asymmetry and sharpness. These accurately extracted feature parameters are then used to determine the flow mode of the fluid (e.g., from a stable liquid phase flow to an unstable gas-liquid two-phase flow), enabling the control strategy adjustment module to switch in a timely manner to a differential pressure balance control strategy for unstable states. For example, it can estimate the trend of the pressure signal and adjust the opening of the control valve, while limiting the adjustment speed and maintaining a minimum opening to ensure the stable operation of the sealing and drainage device.

[0066] In some embodiments, the step of performing real-time adaptive calibration of feature parameters on the segmented pressure signal to dynamically adjust the parameter thresholds or recognition rules for feature extraction includes: The pressure signal after segmented processing is evaluated in real time to determine the degree of change in the properties of the fluid components. When the changes in the properties of fluid components exceed a preset threshold, a rapid calibration mode is triggered, which increases the update frequency of feature parameters and uses short-term historical data and real-time data difference analysis to adjust the parameter thresholds or recognition rules for feature extraction. When the drastic change in the properties of the fluid components does not exceed the preset threshold, the normal calibration mode is maintained. In this mode, a calibration algorithm based on long-term historical data trend analysis is used to adjust the parameter thresholds or recognition rules for feature extraction.

[0067] Specifically, the real-time rate of change of the segmented pressure signal is evaluated to quantify the speed and magnitude of changes in fluid component properties over time. This evaluation can be based on various indicators such as the instantaneous rate of change of the pressure signal, short-term mean changes, or energy changes of specific frequency components. Its purpose is to provide a basis for subsequent calibration mode selection. A preset threshold is a key parameter used to distinguish the degree of drastic change in fluid component properties. This threshold can be set according to actual operating conditions, historical data analysis, and the system's requirements for response speed and stability. When the evaluated drastic change exceeds this threshold, it indicates that the fluid state may have undergone a significant or abrupt change, requiring a rapid system response.

[0068] When the drastic changes in fluid component properties exceed a preset threshold, the system triggers a rapid calibration mode. In this mode, the update frequency of feature parameters is increased to ensure the system can quickly capture the latest changes in fluid state. Simultaneously, a method of analyzing the difference between short-term historical data and real-time data is used to adjust the parameter thresholds or recognition rules for feature extraction. This method focuses on comparing recent data with current data, enabling rapid identification and adaptation to sudden changes, avoiding the lag caused by long-term historical data. When the drastic changes in fluid component properties do not exceed the preset threshold, the system maintains a conventional calibration mode. In this mode, a calibration algorithm based on long-term historical data trend analysis is used to adjust the parameter thresholds or recognition rules for feature extraction. This algorithm emphasizes the long-term stability and trend changes of fluid state; by analyzing data over a longer time span, it provides a more stable and robust calibration, avoiding overreaction to minor fluctuations.

[0069] This application introduces a real-time rate-of-change assessment mechanism to dynamically sense the drastic changes in fluid component properties, thereby enabling intelligent switching of adaptive calibration strategies for feature parameters. When fluid component properties change drastically, a rapid calibration mode is activated. By increasing the update frequency and employing short-term data analysis, the parameter thresholds or recognition rules extracted for feature extraction can quickly adapt to the new fluid state, avoiding misjudgments caused by calibration lag. Conversely, when fluid component properties change gradually, a conventional calibration mode is maintained. Through trend analysis based on long-term historical data, the stability and robustness of the calibration are ensured, preventing over-adjustment to normal fluctuations. This hierarchical, adaptive calibration mechanism ensures that the extracted feature parameters always maintain a high degree of consistency with the actual fluid state, thus providing more accurate and reliable input for subsequent flow pattern judgment.

[0070] Through the above technical solution, this application can intelligently select and execute the most suitable adaptive calibration mode for characteristic parameters based on the actual severity of changes in fluid component properties. This significantly improves the real-time performance and accuracy of characteristic parameter extraction, especially when fluid states undergo abrupt changes, enabling rapid response and adjustment, avoiding the response lag or oversensitivity problems that may exist in a single calibration mode. Therefore, it ensures that the differential pressure balance control method of the plugging and draining device can obtain more accurate fluid state characteristic parameters under various complex operating conditions, thereby improving the reliability of flow pattern judgment and the effectiveness and stability of the differential pressure balance control strategy.

[0071] Suppose the sealing and drainage device is processing a multiphase fluid. During normal operation, the fluid composition properties change relatively gradually, such as minor changes in component proportions due to slow fluctuations in temperature or pressure. In this case, the system performs real-time rate-of-change assessment on the segmented pressure signal, determining that its intensity does not exceed a preset threshold, thus maintaining the conventional calibration mode. In this mode, the system uses a trend analysis algorithm based on historical data from the past few hours or even days to fine-tune the parameter thresholds or recognition rules for feature extraction, ensuring a smooth calibration process with good anti-interference capabilities. For example, when a large number of bubbles or droplets suddenly appear in the fluid, causing drastic changes in the fluid composition properties, the real-time rate-of-change assessment will detect the drastic fluctuations in the pressure signal and determine that their intensity exceeds a preset threshold. At this point, the system immediately triggers a rapid calibration mode. In this mode, the update frequency of feature parameters is significantly increased, for example, from once per minute to several times per second. Simultaneously, the calibration algorithm will primarily focus on the differences between short-term historical data (last few seconds or tens of seconds) and real-time data, rapidly adjusting the parameter thresholds or recognition rules for feature extraction to adapt to the new fluid composition state. For example, if the presence of bubbles causes a significant increase in the kurtosis of the pressure signal, the fast calibration mode will quickly adjust the recognition rules to more accurately capture this change, thereby ensuring that characteristic parameters reflecting the fluid state can be accurately extracted even when the fluid state changes drastically.

[0072] In some embodiments, the step of evaluating the real-time rate of change of the segmented pressure signal to determine the severity of the change in fluid component properties includes: The segmented pressure signal is decomposed into components at multiple time scales. For components at multiple time scales, calculate their energy intensity within a preset time window; By comparing energy intensities at different time scales, the changing trends and drastic degrees of fluid component properties at different time scales can be identified, thus serving as a measure of the drastic degree of change in fluid component properties.

[0073] Specifically, multi-scale decomposition of the segmented pressure signal refers to using signal processing techniques, such as wavelet transform, empirical mode decomposition (EMD), or variational mode decomposition (VMD), to decompose the original pressure signal into components with different frequency characteristics at different time scales. These components can respectively reflect the changes in fluid component properties across different time dimensions, such as rapid transients, moderate fluctuations, and slow trends. The time-scale components can be understood as sub-signals obtained after decomposition, representing a specific frequency range or time resolution.

[0074] Calculating the energy intensity of components at multiple time scales within a preset time window means quantifying the activity or fluctuation intensity of each component within a specific time window by calculating its sum of squares or energy spectral density, for each decomposed time scale component, within its corresponding local time period. The length of the preset time window can be set according to the actual application scenario and the characteristics of fluid changes to ensure effective capture of energy changes at different time scales.

[0075] Comparing energy intensities at different time scales aims to comprehensively assess the drastic changes in fluid component properties by analyzing the relative magnitudes and variation patterns of energy intensities at each scale. For example, a significant increase in the energy intensity of high-frequency components may indicate rapid and drastic changes in the fluid composition; while changes in the energy intensity of low-frequency components may indicate longer-term trends. This comparison allows for a more comprehensive and accurate identification of the trends and drastic changes in fluid component properties across different time scales.

[0076] This application employs multi-scale decomposition of the segmented pressure signal to separate information on changes in fluid component properties into different time-scale components. This decomposition allows the system to independently analyze the dynamic behavior of the fluid across different time dimensions, including instantaneous, short-term, and long-term timeframes. By calculating the energy intensity of these time-scale components within a preset time window, the activity level of fluid component changes at each time scale can be quantified. Furthermore, by comparing the energy intensity at different time scales, this approach can identify the trends and drastic changes in fluid component properties across different time scales. For example, drastic changes in high-frequency component energy may indicate rapid abrupt changes in component properties, while slow changes in low-frequency component energy may reflect gradual evolution of component properties. This multi-scale analysis method avoids information loss or misjudgment that may result from single-time-scale assessment, thus enabling a more comprehensive and accurate capture of the true drastic changes in fluid component properties.

[0077] The above technical solution overcomes the limitations of single-timescale assessment methods in handling complex fluid composition changes. This solution, through multi-scale decomposition and energy intensity comparison, enables a more refined and accurate assessment of the severity of fluid composition property changes, especially when fluid composition changes exhibit multi-timescale characteristics, effectively identifying and distinguishing change patterns across different timescales. This significantly improves the robustness and accuracy of real-time rate of change assessment, thereby optimizing the performance of adaptive calibration of feature parameters. It ensures that under complex and variable fluid conditions, the parameter thresholds or identification rules for feature extraction can be dynamically adjusted more promptly and accurately, thus enhancing the adaptability and control effectiveness of the overall differential pressure balance control method for the sealing and drainage device.

[0078] When evaluating the real-time rate of change of the segmented pressure signal, wavelet packet decomposition can be used as a multi-scale decomposition method. Specifically, the acquired segmented pressure signal is first decomposed into three levels of wavelet packets, resulting in eight wavelet packet components with different frequency ranges, each representing a specific time scale. Then, for each wavelet packet component, its energy intensity is calculated within a sliding time window of one second. For example, the energy intensity of high-frequency components may reflect rapid changes such as bubble bursting or instantaneous turbulence in the fluid; the energy intensity of mid-frequency components may be related to moderate fluctuations in droplet size or component mixing in the fluid; while low-frequency components may reflect slow changes in the overall density or viscosity of the fluid. By monitoring and comparing the energy intensity changes of these eight components within their respective time windows in real time, for example, when the energy of the high-frequency component rises sharply in a short period of time while the energy of the low-frequency component remains relatively stable, the system can determine that a drastic instantaneous change has occurred in the fluid composition properties. Conversely, if the energy intensity of all components shows a slow upward or downward trend, it is judged as a gradual change. Based on this multi-scale energy intensity comparison, the system can more accurately determine the severity of changes in fluid component properties and trigger a rapid calibration mode or maintain a regular calibration mode accordingly, so as to dynamically adjust the parameter thresholds or recognition rules for feature extraction.

[0079] In some embodiments, the step of evaluating the real-time rate of change of the segmented pressure signal to determine the severity of the change in fluid component properties includes: The segmented pressure signal is denoised to eliminate the effects of instantaneous noise and outliers. For the denoised pressure signal, calculate its variation amplitude and frequency within a continuous time window; The degree of drastic change in the properties of fluid components can be determined based on the magnitude and frequency of the change.

[0080] Denoising the segmented pressure signal involves filtering out or smoothing any random noise, measurement errors, or transient interference that may be present in the original pressure signal. The aim is to improve the purity and reliability of the signal, ensuring the accuracy of subsequent analysis and preventing transient noise and outliers from misleading the judgment of changes in fluid component properties. Specifically, denoising can be achieved using various signal processing techniques, such as moving average filtering, median filtering, wavelet denoising, or Kalman filtering. The appropriate denoising algorithm should be selected based on the specific application scenario and noise characteristics.

[0081] For the denoised pressure signal, its amplitude and frequency of change within a continuous time window are calculated. The amplitude of change can be understood as the difference between the maximum and minimum values ​​of the pressure signal within a preset time window, or the standard deviation of the signal over a period of time, reflecting the range of pressure fluctuations. The frequency of change can refer to the number of times the pressure signal crosses a baseline (e.g., the average value) within the same time window, or the main frequency components extracted through frequency domain analysis methods such as Fourier transform, reflecting the speed or periodicity of pressure fluctuations. The calculation of these parameters typically involves sliding window analysis, which involves performing local statistical analysis on the signal over a continuous time period to capture its dynamic characteristics.

[0082] The severity of changes in fluid component properties can be determined by comprehensively analyzing the amplitude and frequency of pressure signal changes. For example, a large amplitude and high frequency of pressure signal changes usually indicate drastic changes in fluid component properties; conversely, a small amplitude and low frequency of changes indicate relatively stable fluid component properties. In practical applications, different thresholds can be set or multi-dimensional judgment models can be established to quantify this severity, thereby providing a basis for subsequent adjustments to control strategies.

[0083] This application first denoises the segmented pressure signal, effectively eliminating the interference of instantaneous noise and outliers, ensuring the accuracy and reliability of subsequent analysis. Based on this, by calculating the amplitude and frequency of the denoised pressure signal within a continuous time window, the dynamic characteristics of fluid pressure can be quantitatively captured. The amplitude directly reflects the range of pressure fluctuations caused by changes in fluid composition, while the frequency reveals the rate and periodicity of this change. By comprehensively considering these two key parameters, the severity of changes in fluid component properties can be assessed more comprehensively and accurately, thus providing a reliable basis for the selection of subsequent adaptive calibration modes.

[0084] The above technical solution effectively avoids interference from instantaneous noise and outliers in judging the severity of changes in fluid component properties, improving the accuracy and robustness of the assessment. Simultaneously, by quantitatively analyzing the amplitude and frequency of pressure signal changes, the judgment of the severity of fluid component property changes becomes more refined and objective. This allows for more accurate triggering of rapid calibration modes or maintenance of conventional calibration modes, further enhancing the response speed and adaptability of adaptive calibration of characteristic parameters, and ensuring the effectiveness and stability of the differential pressure balance control strategy under different fluid conditions.

[0085] like Figure 2 As shown in the embodiments, this application also discloses a differential pressure balance control system for a plugging and draining device, comprising: Pressure signal acquisition module 1 is used to acquire the pressure signal of the sealing and guiding device and to process the pressure signal in segments. Feature parameter extraction module 2 is used to extract feature parameters reflecting the fluid state from the segmented pressure signal; The flow mode determination module 3 is used to determine the flow mode of the fluid based on characteristic parameters. Control strategy adjustment module 4 is used to adjust the differential pressure balance control strategy according to the flow pattern, wherein: When the flow pattern is in a steady state, the conventional differential pressure feedback regulation method is adopted; When the flow mode is unstable, the pressure signal is trend estimated to obtain the overall differential pressure change trend. Based on the overall differential pressure change trend, the opening of the control valve is adjusted, while the adjustment speed of the control valve is limited and the minimum opening of the control valve is maintained.

[0086] The differential pressure balance control system for the sealing and drainage device disclosed in this application achieves intelligent balance control of the differential pressure in the sealing and drainage device through modular design. In complex underground engineering environments, especially when the water flow contains dissolved gases forming a liquid-gas two-phase flow, this system monitors the pressure in real time through a pressure signal acquisition module, deeply analyzes the fluid state through a feature parameter extraction module, and identifies the current fluid flow pattern through a flow mode judgment module. Based on this judgment, the control strategy adjustment module can adaptively select an appropriate control strategy, thereby effectively avoiding misjudgment and over-adjustment under unstable flow conditions, ensuring the stable operation and safety of the sealing and drainage device.

[0087] The above are merely embodiments of this application and are not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for differential pressure balance control of a sealing and drainage device, characterized in that, include: Acquire the pressure signal of the sealing and drainage device, and process the pressure signal in segments; Extract characteristic parameters reflecting the fluid state from the segmented pressure signal; Determine the flow mode of the fluid based on the characteristic parameters; Adjust the differential pressure balance control strategy according to the flow pattern, where: When the flow pattern is in a steady state, the conventional differential pressure feedback regulation method is adopted; When the flow mode is unstable, the pressure signal is trend estimated to obtain the overall differential pressure change trend. Based on the overall differential pressure change trend, the opening of the control valve is adjusted, while the adjustment speed of the control valve is limited and the minimum opening of the control valve is maintained.

2. The differential pressure balance control method for a sealing and drainage device according to claim 1, characterized in that, The step of determining the flow mode of the fluid based on characteristic parameters includes: Acquire the pressure signal of the sealing and drainage device, and perform preliminary processing on the pressure signal; The pre-processed pressure signal is decomposed into multiple scale components. Extracting component characteristics from multiple scale components; Determine the flow pattern of the fluid based on its component characteristics.

3. The differential pressure balance control method for a sealing and drainage device according to claim 1, characterized in that, The step of extracting characteristic parameters reflecting the fluid state from the segmented pressure signal includes: For the segmented pressure signal, extract the fluctuation degree, main frequency components, skewness and kurtosis of the pressure signal; Cross-correlation analysis was performed on the fluctuation degree, main frequency components, skewness, and kurtosis of the extracted pressure signal; Based on the results of cross-correlation analysis, fluid component identification rules are constructed; Based on fluid component identification rules, the presence and contribution ratio of different components in the fluid are determined as characteristic parameters reflecting the fluid state.

4. The differential pressure balance control method for a sealing and drainage device according to claim 1, characterized in that, The step of extracting characteristic parameters reflecting the fluid state from the segmented pressure signal includes: Multi-scale energy normalization is performed on the segmented pressure signal to eliminate the differences in pressure fluctuation amplitude caused by different components. Energy intensity, fluctuation asymmetry, and local fluctuation frequency are extracted from the normalized pressure signal at different frequency scales. By comparing energy intensity at different frequency scales, the relative contribution of each component in the fluid is determined. Combined with wave asymmetry and local wave frequency, the existence and state of all components in the fluid are identified as characteristic parameters reflecting the fluid state.

5. The differential pressure balance control method for a sealing and drainage device according to claim 1, characterized in that, The step of extracting characteristic parameters reflecting the fluid state from the segmented pressure signal includes: Nonlinear characteristic transformation is performed on the segmented pressure signal to enhance the pressure fluctuation characteristics caused by nonlinear interactions between fluid components; Energy intensity, fluctuation asymmetry, and local fluctuation frequency are extracted from the pressure signal after nonlinear characteristic transformation at multiple frequency scales. By analyzing the interaction between energy intensity, wave asymmetry, and local wave frequencies at different frequency scales, the cooperative interaction modes of various components in the fluid and their contribution to the overall fluid state are identified, serving as characteristic parameters reflecting the fluid state.

6. The differential pressure balance control method for a sealing and drainage device according to claim 1, characterized in that, The step of extracting characteristic parameters reflecting the fluid state from the segmented pressure signal includes: Real-time adaptive calibration of feature parameters is performed on the segmented pressure signal to dynamically adjust the parameter thresholds or recognition rules for feature extraction. For the adaptively calibrated pressure signal, extract the fluctuation degree, main frequency components, skewness and kurtosis of the pressure signal; Based on the fluctuation degree, main frequency components, skewness, and kurtosis of the extracted pressure signal, the presence and contribution ratio of different components in the fluid are determined, which serve as characteristic parameters reflecting the fluid state.

7. The differential pressure balance control method for a sealing and drainage device according to claim 6, characterized in that, The step of performing real-time adaptive calibration of feature parameters on the segmented pressure signal to dynamically adjust the parameter thresholds or recognition rules for feature extraction includes: The pressure signal after segmented processing is evaluated in real time to determine the degree of change in the properties of the fluid components. When the changes in the properties of fluid components exceed a preset threshold, a rapid calibration mode is triggered, which increases the update frequency of feature parameters and uses short-term historical data and real-time data difference analysis to adjust the parameter thresholds or recognition rules for feature extraction. When the drastic change in the properties of the fluid components does not exceed the preset threshold, the normal calibration mode is maintained. In this mode, a calibration algorithm based on long-term historical data trend analysis is used to adjust the parameter thresholds or recognition rules for feature extraction.

8. The differential pressure balance control method for a sealing and drainage device according to claim 7, characterized in that, The step of evaluating the real-time rate of change of the segmented pressure signal to determine the severity of the change in fluid component properties includes: The segmented pressure signal is decomposed into components at multiple time scales. For components at multiple time scales, calculate their energy intensity within a preset time window; By comparing energy intensities at different time scales, the changing trends and drastic degrees of fluid component properties at different time scales can be identified, thus serving as a measure of the drastic degree of change in fluid component properties.

9. The differential pressure balance control method for a sealing and drainage device according to claim 7, characterized in that, The step of evaluating the real-time rate of change of the segmented pressure signal to determine the severity of the change in fluid component properties includes: The segmented pressure signal is denoised to eliminate the effects of instantaneous noise and outliers. For the denoised pressure signal, calculate its variation amplitude and frequency within a continuous time window; The degree of drastic change in the properties of fluid components can be determined based on the magnitude and frequency of the change.

10. A differential pressure balance control system for a sealing and drainage device, using the differential pressure balance control method for a sealing and drainage device as described in any one of claims 1-9, characterized in that, Includes the following modules: The pressure signal acquisition module is used to acquire the pressure signal of the sealing and guiding device and to process the pressure signal in segments. The feature parameter extraction module is used to extract feature parameters reflecting the fluid state from the segmented pressure signal; The flow mode determination module is used to determine the flow mode of the fluid based on characteristic parameters. The control strategy adjustment module is used to adjust the differential pressure balance control strategy according to the flow pattern, wherein: When the flow pattern is in a steady state, the conventional differential pressure feedback regulation method is adopted; When the flow mode is unstable, the pressure signal is trend estimated to obtain the overall differential pressure change trend. Based on the overall differential pressure change trend, the opening of the control valve is adjusted, while the adjustment speed of the control valve is limited and the minimum opening of the control valve is maintained.

Images