Working condition monitoring and closed-loop control method and system for deep fill grouting process

Abstract

This invention relates to the field of geotechnical engineering, specifically to a method and system for monitoring and closed-loop control of grouting conditions in deep backfill processes. Addressing the false rejection phenomenon caused by filter cake buildup at the orifice during grouting, this invention utilizes the thrust-suction pulse characteristics of the grouting pump. Based on flow signal analysis of the time boundary, it obtains the formation elastic rebound ratio by analyzing the differences in dynamic response at different pressure ranges to offset grout viscosity interference. It also obtains the flow resistance fluctuation rate by analyzing flow resistance fluctuation characteristics, and uses the flow resistance fluctuation rate as a confidence weight to analyze energy distribution uniformity to obtain the non-uniformity of suction resistance, thus identifying micro-instability. This allows for the determination of grouting conditions and the generation of pump control commands. It effectively avoids the masking of formation rebound characteristics by non-Newtonian fluid viscosity, accurately distinguishes between effective soil compaction and filter cake blockage, and captures early signs of micro-instability in the orifice flow channel before macroscopic pressure exceeds limits, achieving early warning and closed-loop control of blockage risks, thereby improving the quality of grouting reinforcement.

Description

Technical Field

[0001] This invention relates to the field of geotechnical engineering technology, specifically to a method and system for monitoring and closed-loop control of working conditions during the deep backfill grouting process. Background Technology

[0002] Deep fill engineering is widely used in high-fill roadbeds and deep pit backfilling projects. These strata are characterized by strong heterogeneity and complex pore structures, often requiring layered grouting for reinforcement. During grouting, grout particles filter and accumulate at the soil pore inlets around the grouting holes, gradually forming a dense filter cake, leading to localized blockage of the grouting channels. At this point, the pressure gauge reading at the orifice rises rapidly and reaches the design final pressure, exhibiting high-pressure characteristics. However, this high pressure may originate from two completely different conditions: one is the true high pressure formed after effective soil compaction, and the other is the false high pressure caused by the filter cake blockage at the orifice, known as false rejection.

[0003] Current technologies primarily rely on two macroscopic indicators—orifice pressure and grout volume—for grouting monitoring. However, when pressure increases, existing technologies cannot distinguish between genuine high pressure from effective soil compaction and pseudo-high pressure caused by orifice filter cake blockage. This difficulty in identification can lead to misjudgments of the grouting status: mistaking pseudo-high pressure for effective compaction and stopping grouting will result in incomplete grouting and insufficient reinforcement; continuing grouting may lead to grout leakage, ground heave, and other engineering accidents. Therefore, accurately identifying pseudo-rejection phenomena during the grouting process is a pressing technical problem that needs to be solved in deep fill grouting projects. Summary of the Invention

[0004] To address the technical problem of difficulty in distinguishing between effective stress increase and ineffective resistance increase caused by false rejection, which leads to incomplete or ineffective grouting and affects the quality of engineering reinforcement, the present invention aims to provide a method and system for monitoring and closed-loop control of working conditions during deep fill grouting. The specific technical solution adopted is as follows: On the one hand, the present invention provides a method for monitoring and closed-loop control of working conditions during deep backfill grouting, the method comprising: The pressure and flow signals during the grouting process are collected synchronously. The time boundary of the grouting pump push stroke is determined based on the flow signal to obtain the loading segment and unloading segment of the current pulse. The pressure and flow signals in the loading segment and the unloading segment are filtered to obtain the loading segment pressure sequence, the loading segment flow sequence, and the unloading segment pressure sequence. By comparing the pressure sequence of the loading section and the pressure sequence of the unloading section, and based on the symmetrical influence of fluid viscous resistance on pressure loading and pressure dissipation, the formation elastic rebound ratio is obtained; the instantaneous flow resistance is calculated based on the pressure sequence and the flow rate sequence of the loading section, and the stability of the orifice flow channel is determined based on the dispersion of the instantaneous flow resistance, thus obtaining the flow resistance fluctuation rate; using the flow resistance fluctuation rate as the confidence weight, the energy distribution during the grouting process is analyzed, thus obtaining the non-uniformity of grout suction resistance. The grouting conditions are determined based on the formation elastic rebound ratio, the flow resistance fluctuation rate, and the non-uniformity of the grout suction resistance, and corresponding pump control commands are generated based on the determination results.

[0005] Furthermore, the method for obtaining the loading segment and the unloading segment includes: The moment when the flow signal crosses the preset flow validity threshold from low to high is taken as the start time of the push-out, and the moment when the flow signal crosses the preset flow validity threshold from high to low is taken as the end time of the push-out. The interval between the start time of the push-out and the end time of the push-out is taken as the loading segment, and the interval within the preset observation window after the end time of the push-out is taken as the unloading segment.

[0006] Furthermore, the methods for obtaining the loading section pressure sequence, the loading section flow rate sequence, and the unloading section pressure sequence include: Calculate the grouting volume of the loading section. When the grouting volume is greater than or equal to the preset minimum grouting volume threshold, determine that the loading section is an effective loading section and the unloading section is an effective unloading section. The loading section pressure sequence is obtained based on the pressure signal within the effective loading section, the loading section flow sequence is obtained based on the flow signal within the effective loading section, and the unloading section pressure sequence is obtained based on the pressure signal within the effective unloading section.

[0007] Furthermore, the method for calculating the formation elastic rebound ratio includes: Obtain the pressure peak value in the loading segment pressure sequence, obtain the time when the pressure value in the loading segment pressure sequence first reaches the pressure peak value at a preset first proportion and a preset second proportion, and obtain the pressure loading duration based on the time difference between the two times. The moment when the pressure value in the unloading section pressure sequence first reaches the preset third proportion is obtained. If such a moment exists, the pressure dissipation half-life is obtained based on the time difference between the moment and the start time of the unloading section. If the preset third proportion is not reached after exceeding the preset observation window, the duration of the observation window is taken as the pressure dissipation half-life. The formation elastic rebound ratio is calculated based on the pressure loading duration, the pressure dissipation half-life, and the preset regularization parameters.

[0008] Furthermore, the method for calculating the flow resistance fluctuation rate includes: The instantaneous flow resistance sequence is calculated based on the loading section pressure sequence and the loading section flow rate sequence. The instantaneous flow resistance sequence is detrended to obtain the residual sequence. The standard deviation of the residual sequence and the arithmetic mean of the instantaneous flow resistance sequence are calculated. When the arithmetic mean is greater than or equal to a preset flow resistance threshold, the flow resistance fluctuation rate is obtained based on the standard deviation and the arithmetic mean. When the average value is less than the flow resistance threshold, the flow resistance fluctuation rate is set to zero.

[0009] Furthermore, the method for calculating the non-uniformity of the suction resistance includes: The energy density of the current pulse is calculated based on the loading segment pressure sequence and the loading segment flow sequence, and the confidence weight of the energy density is determined based on the flow resistance volatility and the preset historical volatility benchmark. The energy density distribution of multiple pulses within a preset observation window is compared with an ideal uniform distribution based on the confidence weight to obtain the non-uniformity of the suction resistance.

[0010] Further, the step of comparing the energy density distribution of multiple pulses within a preset observation window with an ideal uniform distribution based on confidence weights to obtain the non-uniformity of the suction resistance includes: The energy densities of multiple pulses within the preset observation window are arranged in ascending order, and the sorted energy densities are multiplied by the corresponding confidence weights to obtain the weighted energy value of each pulse. The weighted energy values ​​of the pulses at each position within the observation window and the previous pulses are accumulated to obtain the weighted energy accumulation value at each position. The weighted energy accumulation value at each position is then normalized with the total weighted energy of all pulses within the observation window to obtain the weighted energy accumulation percentage at each position. Based on the weighted cumulative energy ratio at each position within the observation window, a weighted cumulative energy ratio curve is constructed at each position within the observation window. The non-uniformity of the suction resistance is obtained based on the degree of deviation of the weighted cumulative energy ratio curve from the ideal uniform distribution.

[0011] Furthermore, the method for determining the grouting condition includes: If the non-uniformity of the suction resistance is greater than the preset blockage limit value, it is determined to be a false high-pressure deadlock condition. If the formation elastic rebound ratio is greater than the preset high resistance benchmark, it is determined to be a pseudo-high pressure rigid closure condition. If the formation elastic rebound ratio is less than the preset elasticity judgment benchmark, it is judged as a smooth flow channel or an abnormal slurry leakage condition. If the flow resistance fluctuation rate is greater than or equal to the preset stability warning line, it is determined to be an unstable working condition of the orifice flow channel; Otherwise, it is determined to be an effective compaction condition.

[0012] Furthermore, the method for generating the pump control command includes: For the pseudo-high pressure deadlock condition or pseudo-high pressure rigid closure condition, a filter cake destruction and cleaning command is generated. The filter cake destruction and cleaning command is a high-frequency micro-amplitude flutter command or an intermittent pulse cleaning command. For the aforementioned conditions of unobstructed flow or abnormal slurry flow, a slowdown and waiting-to-cure command is generated; In response to the unstable working condition of the orifice flow channel, a speed reduction coefficient is calculated based on the flow resistance fluctuation rate and the preset stability warning line, and a speed reduction and flow stabilization command is generated. For the effective compaction condition, a steady-state maintenance command is generated to maintain the current grouting parameters.

[0013] On the other hand, the present invention provides a working condition monitoring and closed-loop control system for deep backfill grouting process, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the working condition monitoring and closed-loop control method for deep backfill grouting process.

[0014] The present invention has the following beneficial effects: By determining the time boundary of the grouting pump's thrust stroke to obtain the loading and unloading sections, the continuous grouting process can be discretized into independent analysis units, facilitating the extraction of the dynamic response characteristics of a single pulse. By filtering the pressure and flow signals of the loading and unloading sections, complete single-pulse dynamic data can be obtained, providing accurate input for subsequent feature calculations.

[0015] By utilizing the symmetrical influence of fluid viscous resistance on pressure loading and dissipation to determine the formation elastic rebound ratio, fluid viscosity interference can be counteracted, accurately restoring the true elastic response of the formation skeleton in a high-viscosity slurry environment, and enabling the extraction of engineering features weakly correlated with slurry ratio. By determining the flow resistance fluctuation rate based on the dispersion of instantaneous flow resistance, the microscopic stability characteristics of the orifice channel can be identified, distinguishing between high pressure in a stable channel and high pressure in an unstable channel, and capturing early signs of blockage before macroscopic pressure exceeds limits. By using the flow resistance fluctuation rate as a confidence weight to reduce the influence of false energy consumption, false high energy consumption data caused by unstable accumulation can be eliminated, accurately revealing the true distribution and deadlock characteristics of slurry suction resistance.

[0016] By determining the grouting conditions based on multi-dimensional features and generating pump control commands, it is possible to achieve closed-loop control of comprehensive condition identification and automatic intervention, avoiding misjudgments caused by a single pressure index. Attached Figure Description

[0017] Figure 1 This is a flowchart of a method for monitoring and closed-loop control of working conditions in a deep backfill grouting process, provided as an embodiment of the present invention. Detailed Implementation

[0018] This invention is primarily applied to grouting reinforcement in deep fill engineering projects, particularly in the layered grouting process of high fill subgrades and deep pit backfilling projects, enabling real-time monitoring and closed-loop control of the grout diffusion state during the grouting process. Specifically, this invention can be applied to grouting sites equipped with electro-hydraulic servo grouting pumps or frequency-controlled mechanical plunger pumps, and can also be applied to the retrofitting of conventional mechanical grouting pumps with added frequency converters. Furthermore, this invention can also be applied to grouting reinforcement of heterogeneous strata exhibiting false rejection phenomena, and to grouting quality control requiring precise differentiation between soil compaction and filter cake blockage, among other scenarios necessitating refined monitoring and dynamic control of the grouting process.

[0019] The following description, in conjunction with the accompanying drawings, details a specific scheme for a method and system for monitoring working conditions and implementing closed-loop control during deep backfill grouting, provided by the present invention.

[0020] Example 1: Please see Figure 1 The diagram illustrates a method flowchart for monitoring and closed-loop control of deep backfill grouting processes according to an embodiment of the present invention. The method includes the following steps: Step S101: Synchronously acquire pressure and flow signals during the grouting process, determine the time boundary of the grouting pump push stroke based on the flow signal, and obtain the loading segment and unloading segment of the current pulse; filter the pressure and flow signals in the loading segment and the unloading segment to obtain the loading segment pressure sequence, the loading segment flow sequence, and the unloading segment pressure sequence.

[0021] Deep fill grouting is a reinforcement technique that improves the mechanical properties of soil by injecting grout into the pores of the stratum. The grouting pump uses a reciprocating piston motion to transport the grout, and the working process includes two stages: pushing and suction. In the pushing stage, the piston moves forward to push the grout into the stratum, increasing the pressure in the pipeline. In the suction stage, the piston moves backward to draw grout from the storage tank, decreasing the pressure in the pipeline. This periodic pushing and suction motion causes the stratum to experience periodic pressure fluctuations, and each complete pushing and suction cycle is an independent stratum response test process.

[0022] Pressure signals reflect the fluid resistance state during grouting, including the elastic response of the formation skeleton and the viscous resistance of the grout flow. Flow signals reflect the delivery capacity of the grouting pump and the permeability of the grout in the formation. Monitoring pressure alone cannot distinguish between an increase in effective stress (soil compaction) and an increase in ineffective resistance (filter cake blockage), as both will cause an increase in the orifice pressure gauge reading. By simultaneously acquiring pressure and flow signals, fluid dynamic characteristics can be calculated, and engineering characteristic indicators reflecting the response characteristics of the formation skeleton can be extracted.

[0023] The time boundaries of the grouting pump's push stroke are the start and end times of the pump's grouting action, defining the physical range of a single push-suction cycle. The start of the push stroke corresponds to the moment the grouting pump piston begins to move forward, and the end of the push stroke corresponds to the moment the piston reaches its forward position and begins to move in the opposite direction. Accurately locking the time boundaries is a prerequisite for data discretization. Only by clearly defining the start and end times of each push-suction cycle can data for the corresponding time period be extracted from the continuous monitoring data stream for analysis. By encapsulating continuous monitoring data into single-pulse grouting records, the continuous physical process can be transformed into standardized data objects with clearly defined physical boundaries, facilitating subsequent independent feature extraction and analysis.

[0024] The loading phase is the time period from the start to the end of the grouting stroke, corresponding to the process of the grouting pump pushing the grout into the formation, during which the pressure in the pipeline continuously increases. The unloading phase is a fixed-duration observation window after the end of the grouting stroke, corresponding to the process of the formation pressure naturally dissipating after the pump stops, during which the pressure in the pipeline gradually decreases. Dividing the grouting cycle into loading and unloading phases allows for the analysis of the pressure build-up characteristics during the pressurization process and the pressure dissipation characteristics during the depressurization process. By analyzing the differences in the dynamic responses of these two phases, engineering characteristic indicators reflecting the formation skeleton properties can be extracted.

[0025] In addition, factors such as equipment jogging, pipeline vibration, or cavitation can generate invalid minor disturbances. These disturbances do not possess complete push-suction cycle characteristics, and including them in the analysis will lead to distorted feature extraction results. Validity verification is performed by calculating the grouting volume of the loading section to determine whether the grouting volume reaches the minimum effective grouting volume threshold. Records generated by invalid disturbances are then removed to ensure that subsequent analyses are based on complete and valid push-suction cycle data.

[0026] As a concrete example, a high-frequency pressure sensor and an electromagnetic flow sensor are installed at the high-pressure manifold at the outlet of the grouting pump. Pressure and flow signals are synchronously acquired at a fixed sampling frequency of 100Hz. The output is a time-aligned raw data stream, which is then filtered using a low-pass filter with a cutoff frequency of 200Hz to remove high-frequency electromagnetic interference and high-frequency mechanical vibration noise from the pipeline before analog-to-digital conversion. To further improve the signal-to-noise ratio, a moving average filtering algorithm is used in software to preprocess the raw data, with a sliding window length of 5 sampling points.

[0027] Furthermore, in some implementations of the embodiments of the present invention, the method for obtaining the loading segment and the unloading segment includes: The moment when the flow signal crosses the preset flow validity threshold from low to high is taken as the start time of the push-out, and the moment when the flow signal crosses the preset flow validity threshold from high to low is taken as the end time of the push-out. The interval between the start time of the push-out process and the end time of the push-out process is taken as the loading segment, and the interval within the preset observation window after the end time of the push-out process is taken as the unloading segment.

[0028] The grouting pump employs a reciprocating piston motion. At the start of the pushing stroke, the piston moves forward, pushing the grout into the formation, and the flow rate in the pipeline rises from zero. By monitoring the moment when the flow signal crosses a preset effective flow rate threshold from low to high, the initial instant of the grouting pump's pushing action can be precisely located. At the end of the pushing stroke, the piston reaches its forward position and begins to move in the opposite direction, and the flow rate in the pipeline begins to decrease from its peak. By monitoring the moment when the flow signal crosses a preset effective flow rate threshold from high to low, the instant when the piston reverses direction can be precisely located, i.e., the dividing point between the end of the pushing action and the start of the suction action. The effective flow rate threshold is used to filter out minor disturbances, ensuring that what is detected is the actual pushing action rather than noise interference.

[0029] The loading phase corresponds to the grouting process, during which formation pressure gradually increases, requiring analysis of pressure build-up characteristics. The unloading phase corresponds to the pressure dissipation process after pump shutdown, requiring analysis of formation rebound characteristics. The start and end times of the grouting stroke define the time range of the grouting process. The interval within a pre-set observation window after the end of the grouting stroke is the unloading phase, capturing the natural dissipation of formation pressure. By dividing the continuous monitoring data into loading and unloading phases using these two time boundaries, the differences in dynamic response between the pressurization and depressurization processes can be analyzed separately.

[0030] As a specific example, the preset effective flow rate threshold is set at 5% of the rated flow rate of the grouting pump. The system monitors the flow rate data in real time and acquires the moment when the flow rate signal reaches the effective flow rate threshold. If the flow rate value at the previous moment is greater than the flow rate value at the next moment, then that moment is taken as the start time of the grouting process; if the flow rate value at the previous moment is less than the flow rate value at the next moment, then that moment is taken as the end time of the grouting process. The interval between the start time and the end time of the grouting process is defined as the loading segment, and the interval within the preset observation window after the end time of the grouting process is defined as the unloading segment. The observation window duration is set to 1.0 second, and this value can be adjusted according to the empirical value of the formation permeability.

[0031] In another specific implementation, in order to prevent false judgments caused by noise at a single sampling point, after detecting that the flow signal has crossed the preset flow validity threshold, it is further checked whether the flow remains in the state after crossing in the subsequent three consecutive sampling points. If so, the start or end time of the push-out is confirmed to be valid; otherwise, the flow data is monitored until the confirmation condition is met.

[0032] Furthermore, in some implementations of the embodiments of the present invention, the method for obtaining the loading section pressure sequence, the loading section flow sequence, and the unloading section pressure sequence includes: Calculate the grouting volume of the loading section. When the grouting volume is greater than or equal to the preset minimum grouting volume threshold, determine that the loading section is an effective loading section and the unloading section is an effective unloading section. The loading section pressure sequence is obtained based on the pressure signal within the effective loading section, the loading section flow sequence is obtained based on the flow signal within the effective loading section, and the unloading section pressure sequence is obtained based on the pressure signal within the effective unloading section.

[0033] Grouting volume refers to the volume of grout injected into the formation during a single grouting cycle, reflecting the pumping capacity and grout penetration within the formation. The calculation of the single-pulse grouting volume serves as a basis for determining the physical significance of that grouting cycle. Factors such as equipment jogging, pipeline vibration, or cavitation can generate minute disturbance signals. These signals lack complete grouting cycle characteristics, exhibiting extremely low flow rates or abnormal pressures. Including invalid disturbances in the analysis will distort the feature extraction results, thus affecting the accuracy of operational condition determination. Validity verification involves eliminating invalid disturbances with grouting volumes below a threshold to ensure that subsequent analyses are based on complete and valid grouting cycle data.

[0034] The loading section pressure sequence records the pressure change over time during the grouting process, used to extract pressure establishment features. The loading section flow rate sequence records the flow rate change over time during the grouting process, used in conjunction with the pressure sequence to calculate instantaneous flow resistance. The unloading section pressure sequence records the pressure dissipation process over time after pump shutdown, used to extract pressure dissipation features, corresponding to the pressurization process, flow rate process, and depressurization process, respectively, and is a necessary input for subsequent dynamic response feature extraction.

[0035] As a concrete example, the grouting volume of the loading section is calculated using numerical integration. The formula is the average of the flow rates of adjacent sections in the loading section multiplied by the sum of the sampling intervals. The calculated grouting volume is compared with a preset minimum grouting volume threshold. When the grouting volume is greater than or equal to the preset minimum grouting volume threshold, the loading section is determined to be an effective loading section, and the unloading section is determined to be an effective unloading section. When the grouting volume is less than the preset minimum grouting volume threshold, the current loading section and unloading section are determined to be invalid disturbances and discarded. The sequence of pressure values ​​in the effective loading section, arranged in chronological order, is used as the loading section pressure sequence; the sequence of flow rates in the effective loading section, arranged in chronological order, is used as the loading section flow sequence; and the sequence of pressure values ​​in the effective unloading section, arranged in chronological order, is used as the unloading section pressure sequence. The preset minimum grouting volume threshold is 0.2 liters, which can be set according to the theoretical discharge capacity of the grouting pump, typically 20% of the theoretical discharge capacity.

[0036] It should be noted that the pressure value and flow rate value are the numerical values ​​of a single sampling point in the pressure signal and flow rate signal, respectively.

[0037] Step S102: Compare the pressure sequence of the loading section and the pressure sequence of the unloading section, and obtain the formation elastic rebound ratio based on the symmetric influence of fluid viscous resistance on pressure loading and pressure dissipation; calculate the instantaneous flow resistance based on the pressure sequence and flow rate sequence of the loading section, determine the stability of the orifice flow channel based on the dispersion of the instantaneous flow resistance, and obtain the flow resistance fluctuation rate; use the flow resistance fluctuation rate as the confidence weight to analyze the energy distribution during the grouting process and obtain the non-uniformity of grout suction resistance.

[0038] During the loading phase, the pumped energy input overcomes the fluid viscous resistance and formation permeability resistance, causing the pressure to gradually increase. During the unloading phase, the formation skeleton exhibits elastic rebound, the excess pore water pressure gradually dissipates, and the pressure gradually decreases. The viscous effect of the slurry has a similar order of magnitude of resistance during both pressurization and unloading processes, meaning that the pressure decay rate alone cannot distinguish between the true rebound of the formation skeleton and the interference of fluid viscosity. By analyzing the differences in the dynamic responses during the loading and unloading phases and utilizing their symmetry to cancel out the common viscous components, normalized characteristics reflecting the response characteristics of the formation skeleton can be extracted.

[0039] When the grouting channel becomes blocked, grout particles form a dense filter cake structure at the orifice, and random collisions between particles cause drastic fluctuations in flow resistance. By monitoring the degree of flow resistance fluctuation, it is possible to identify whether the orifice flow channel is in a stable state and to detect early signs of micro-instability before macroscopic pressure exceeds the limit.

[0040] When a grouting channel locks up, most of the energy is consumed in overcoming localized high resistance (filter cake), resulting in extremely uneven energy distribution. When grouting is smooth, the energy distribution is relatively uniform. By analyzing the uniformity of energy distribution and eliminating false high-energy-consumption data caused by unstable accumulation, deadlock characteristics can be identified from a statistical perspective.

[0041] Furthermore, in some implementations of the embodiments of the present invention, the method for calculating the formation elastic rebound ratio includes: Obtain the pressure peak value in the loading segment pressure sequence, obtain the time when the pressure value in the loading segment pressure sequence first reaches the pressure peak value at a preset first proportion and a preset second proportion, and obtain the pressure loading duration based on the time difference between the two times. The moment when the pressure value in the unloading section pressure sequence first reaches the preset third proportion is obtained. If such a moment exists, the pressure dissipation half-life is obtained based on the time difference between the moment and the start time of the unloading section. If the preset third proportion is not reached after exceeding the preset observation window, the duration of the observation window is taken as the pressure dissipation half-life. The formation elastic rebound ratio is calculated based on the pressure loading duration, the pressure dissipation half-life, and the preset regularization parameters.

[0042] The peak pressure is the maximum pressure within the loading section, marking the peak of formation pressure establishment during the grouting pump's operation. The peak pressure serves as a reference point, used to locate different pressure levels on the pressure curve, thereby determining the characteristic time points of the pressure loading and pressure dissipation phases. Normalization based on the peak pressure eliminates the influence of pressure amplitude differences across different grouting cycles, making the characteristic times comparable.

[0043] The pressure build-up process is not an ideal step signal, but rather a gradual process with a rising edge. Directly using the absolute pressure value cannot accurately describe the dynamic characteristics of pressure changes. By determining two pressure levels using a preset first and second proportion of the pressure peak value, and calculating the time difference between the two pressure levels as the pressure loading duration, the speed at which pressure rises from the lower to the higher level can be quantified, reflecting the rate of pressure build-up under the combined effects of pumping energy input and formation resistance. The pressure dissipation process also exhibits a gradual change. By determining the target pressure decay level using a preset third proportion of the pressure peak value, and calculating the time required to decay to that level as the pressure dissipation half-life, the rate of pressure decrease after pump shutdown can be quantified.

[0044] Pressure loading duration is the time required for pressure to rise from a lower level to a higher level, characterizing the speed at which formation pressure builds up under pumping thrust. Pressure dissipation half-life is the time required for pressure to decay to half its peak level after pumping stops, characterizing the speed at which excess pore water pressure dissipates under the combined effects of fluid viscous resistance and formation permeability resistance. Pressure loading duration incorporates information on pumping energy input and viscous resistance, while pressure dissipation half-life incorporates information on formation elastic rebound and viscous resistance.

[0045] The viscosity effect of slurry exerts a similar order of magnitude of resistance during both pressurization and depressurization, a common characteristic of fluid dynamics. Both the pressure loading duration and the pressure dissipation half-life contain viscous resistance components. By calculating their ratio, these common viscous resistance components can be mathematically canceled out, with the remaining difference primarily reflecting the response characteristics of the formation framework. A larger ratio indicates a significantly prolonged unloading process relative to the loading process, suggesting significant elastic rebound or permeability resistance in the formation. A smaller ratio indicates a rapid completion of the unloading process, suggesting rigid closure or unobstructed flow channels in the formation.

[0046] During the initial stage of grouting or no-load testing, pressure builds up extremely rapidly, and the pressure loading time approaches zero. Directly calculating the ratio may lead to division by zero errors or abnormal results. The regularization parameter is a very small positive value used to replace the pressure loading time approaching zero when calculating the ratio, preventing calculation overflow. The value of the regularization parameter is much smaller than the pressure loading time under normal conditions, so it does not affect the calculation results under normal conditions, and only plays a numerical protection role under abnormal conditions.

[0047] As a concrete example, the peak pressure in the loading phase pressure sequence is obtained. The sequence is then searched in the negative time direction to locate the first time point when the pressure value first reaches the peak pressure, representing a preset first proportion. The second time point is located at the second preset second proportion of the time point. If no matching point is found after a reverse search to the beginning of the sequence, the beginning time is used as a replacement point. The time difference between the first and second times is calculated as the pressure loading duration. In the unloading phase pressure sequence, the sequence is searched in the positive time direction to locate the first time point when the pressure value first reaches the peak pressure, representing a preset third proportion. The time difference between the third time point and the beginning of the unloading phase is calculated as the pressure dissipation half-life. The formation elastic rebound ratio is calculated based on the pressure loading duration and the pressure dissipation half-life, using the formula: pressure dissipation half-life divided by the larger of the pressure loading duration and a preset regularization parameter.

[0048] The first preset ratio is set to 10% of the peak pressure, the second preset ratio is set to 90% of the peak pressure, and the third preset ratio is set to 50% of the peak pressure. The preset regularization parameter is set to 0.01 seconds. When the pressure loading duration approaches zero, the preset regularization parameter is used to replace the pressure loading duration in the calculation to prevent calculation overflow. If the pressure has not decayed to 50% of the peak pressure by the end of the unloading observation window, the pressure dissipation half-life is directly set to equal the observation window duration.

[0049] Furthermore, in some implementations of the present invention, the method for calculating the flow resistance fluctuation rate includes: The instantaneous flow resistance sequence is calculated based on the loading section pressure sequence and the loading section flow rate sequence. The instantaneous flow resistance sequence is detrended to obtain the residual sequence. The standard deviation of the residual sequence and the arithmetic mean of the instantaneous flow resistance sequence are calculated. When the arithmetic mean is greater than or equal to a preset flow resistance threshold, the flow resistance fluctuation rate is obtained based on the standard deviation and the arithmetic mean. When the average value is less than the flow resistance threshold, the flow resistance fluctuation rate is set to zero.

[0050] Instantaneous flow resistance is the resistance encountered by fluid flowing through a pipeline and formation per unit time, defined as the ratio of pressure to flow rate. During grouting, instantaneous flow resistance changes over time, reflecting real-time changes in the fluid dynamics state. The accumulation and collapse of grout particles at the pore inlet causes high-frequency oscillations in the instantaneous flow resistance. By calculating the instantaneous flow resistance and extracting its fluctuation characteristics, the microscopic stability state of the orifice flow channel can be identified.

[0051] The resistance variation within a single pulse is influenced by the pumping sinusoidal waveform, exhibiting a nonlinear baseline shape. To extract the pure oscillation component, the baseline trend introduced by the pumping waveform needs to be removed. By detrending, the residual of the instantaneous flow resistance sequence relative to the baseline sequence is calculated. The residual sequence characterizes the high-frequency flow resistance oscillation component after removing the pumping waveform, reflecting the stochastic dynamic process caused by particle accumulation and collapse.

[0052] In the initial stage of grouting or during the low-pressure no-load phase, the reference resistance is extremely small. At this time, the calculated flow resistance fluctuation rate will be incorrectly amplified, leading to distorted stability assessment. A preset pressure threshold is used to determine whether the current state is non-pressurized. When the average flow resistance is lower than the preset pressure threshold, it is determined that the current state is non-pressurized, and the flow resistance fluctuation rate is directly set to zero to avoid the phenomenon of artificially high fluctuation rate caused by extremely small reference resistance.

[0053] The flow resistance fluctuation rate characterizes the degree of dispersion of resistance relative to the mean. By calculating the ratio of the standard deviation of the residual sequence to the mean of the original sequence, i.e., the coefficient of variation, the influence of pressure amplitude can be normalized, focusing on reflecting the stability of the flow channel's microstructure. A larger ratio indicates more severe resistance fluctuations, and the flow channel is in an unstable state; a smaller ratio indicates more stable resistance, and the flow channel is in a stable state.

[0054] As a specific example, instantaneous flow resistance is calculated based on the pressure and flow rates of the loading section, resulting in an instantaneous flow resistance sequence. The calculation formula is the pressure value divided by the corresponding flow rate value. A moving average method is used to detrend the instantaneous flow resistance sequence, and a moving average sequence of instantaneous flow resistance is calculated. The difference between the instantaneous flow resistance and the moving average sequence is used as the residual sequence. The standard deviation of the residual sequence and the arithmetic mean of the instantaneous flow resistance sequence are calculated. The arithmetic mean is compared with a preset flow resistance threshold. When the average is greater than or equal to the preset flow resistance threshold, the ratio of the standard deviation to the arithmetic mean is used as the flow resistance fluctuation rate. When the average is less than the preset flow resistance threshold, the flow resistance fluctuation rate is set to zero. The moving average method uses a sliding window length of 5 sampling points, and the preset flow resistance threshold is set to 0.5 MPa, or 1.5 times the average flow resistance obtained during the start-up clean water circulation or no-load pipeline phase.

[0055] In another specific implementation, in order to avoid numerical calculation singularities caused by the flow rate approaching zero at the moment of piston reversal, the middle 80% (discarding the first 10% and the last 10%) of the stable data segment of the loading segment pressure sequence and the loading segment flow sequence is extracted for the calculation of instantaneous flow resistance.

[0056] Furthermore, in some implementations of the present invention, the method for calculating the non-uniformity of the suction resistance includes: The energy density of the current pulse is calculated based on the loading segment pressure sequence and the loading segment flow sequence, and the confidence weight of the energy density is determined based on the flow resistance volatility and the preset historical volatility benchmark. The energy density distribution of multiple pulses within a preset observation window is compared with an ideal uniform distribution based on the confidence weight to obtain the non-uniformity of the suction resistance.

[0057] Energy density is the energy contained in a unit volume of slurry, reflecting the energy intensity transported by the slurry during a single pulse. The energy density of different pulses may vary due to differences in formation conditions. When the formation permeability is good, the slurry can penetrate smoothly, resulting in a lower energy density; when the formation is blocked, local resistance increases, leading to a higher energy density.

[0058] Flow resistance volatility reflects the stability of the flow channel's microstructure. A lower flow resistance volatility indicates a more stable flow channel and more reliable current data; a higher flow resistance volatility indicates a more unstable flow channel, and the current data may contain false high-energy-consumption information caused by particle accumulation and collapse. A preset historical volatility benchmark serves as an adaptive background value for flow resistance volatility, reflecting the volatility level under normal grouting conditions. By comparing the current flow resistance volatility with the preset historical volatility benchmark, higher weights are assigned to flows close to the benchmark, while lower weights are assigned to flows significantly higher than the benchmark. This ensures that false high-energy-consumption data generated by unstable accumulation is downweighted or eliminated in subsequent analyses.

[0059] When a grouting channel becomes blocked, most of the energy is consumed in overcoming localized high resistance, with only a few pulses carrying high energy, resulting in extremely uneven energy distribution. When grouting is smooth, the energy density of each pulse is relatively close, and the energy distribution is more uniform. By calculating the cumulative distribution of weighted energy density and comparing it with the ideal energy density distribution, the degree of deviation in energy distribution can be quantified. A greater deviation indicates a more uneven energy distribution and a higher degree of blockage in the grouting channel; a smaller deviation indicates a more uniform energy distribution and a smoother grouting channel.

[0060] As a concrete example, we first calculate the energy density of the current pulse, which can be expressed by the following formula: in, This represents the energy density of the k-th pulse. , These represent the pressure and flow rates at the nth sampling point, respectively. This represents the sampling time interval, where m represents the number of sampling points within the current pulse loading segment. This represents the grouting volume of the k-th pulse.

[0061] Then, the confidence weights are calculated, and the formula can be expressed as: in, This represents the confidence weight of the k-th pulse. This represents the flow resistance fluctuation rate of the k-th pulse. This represents the historical volatility benchmark.

[0062] The system maintains the energy density of multiple pulses within a preset observation window. Based on confidence weights, the distribution of energy density is compared with a preset ideal energy density distribution. The non-uniformity of suction resistance is obtained from the comparison results. The preset observation window length is set to 20, used to store the weighted energy density data of the most recent 20 pulses.

[0063] In another specific implementation, during the cold start learning period, the flow resistance volatility of the first 5 pulses is stored in a cache queue, and the median is used as the initial value of the preset historical volatility benchmark. During continuous operation, the preset historical volatility benchmark is updated using an exponentially weighted moving average method, with the forgetting factor set to 0.1. When the flow resistance volatility exceeds the safety limit, which is set to 3 times the historical volatility benchmark, it is considered an abnormal change, and the preset historical volatility benchmark is not updated in that instance to prevent the benchmark from being contaminated by noise.

[0064] Furthermore, in some implementations of this invention, the step of comparing the energy density distribution relationship of multiple pulses within a preset observation window with an ideal uniform distribution based on confidence weights to obtain the non-uniformity of the suction resistance includes: The energy densities of multiple pulses within the preset observation window are arranged in ascending order, and the sorted energy densities are multiplied by the corresponding confidence weights to obtain the weighted energy value of each pulse. The weighted energy values ​​of the pulses at each position within the observation window and the previous pulses are accumulated to obtain the weighted energy accumulation value at each position. The weighted energy accumulation value at each position is then normalized with the total weighted energy of all pulses within the observation window to obtain the weighted energy accumulation percentage at each position. Based on the weighted cumulative energy ratio at each position within the observation window, a weighted cumulative energy ratio curve is constructed at each position within the observation window. The non-uniformity of the suction resistance is obtained based on the degree of deviation of the weighted cumulative energy ratio curve from the ideal uniform distribution.

[0065] The weighted energy accumulation ratio curve characterizes the weighted energy accumulation distribution. Under an ideal uniform distribution, the contribution of each pulse to the total energy is relatively balanced, and the weighted energy accumulation ratio is linearly related to the pulse number accumulation ratio. Therefore, according to the Lorentz curve principle, the diagonal can be used as the baseline of the ideal uniform distribution. The more uneven the actual energy distribution, the further the weighted energy accumulation ratio curve deviates from the ideal diagonal.

[0066] The degree of deviation is used to characterize the non-uniformity of the actual energy distribution relative to the ideal uniform distribution. When the grouting channel is locked, most of the energy is consumed in overcoming local high resistance, resulting in an abnormally high energy density for a few pulses and extremely uneven energy distribution; the degree of deviation is large. When the grouting process is smooth, the energy density of each pulse is relatively close, the energy distribution is relatively uniform, and the degree of deviation is small. By calculating the area difference between the weighted cumulative energy ratio curve and the ideal diagonal, the weighted Gini coefficient is obtained. This weighted Gini coefficient is used as the non-uniformity of grout suction resistance to quantify the non-uniformity of energy distribution and to determine the degree of blockage in the grouting channel.

[0067] As a concrete example, the energy densities of 20 pulses within the observation window are obtained and arranged in ascending order, while the corresponding confidence weights are simultaneously rearranged. For each sorted pulse, the product of its energy density and confidence weight is calculated as the weighted energy value; the weighted energy values ​​of the first j pulses are accumulated to obtain the cumulative weighted energy value, which can be expressed by the formula: in, This represents the weighted cumulative energy value of the first j pulses within the observation window. This represents the energy density of the i-th pulse after sorting. This represents the confidence weight of the i-th pulse after sorting.

[0068] Furthermore, the weighted cumulative energy value is normalized to the weighted total energy of all pulses within the observation window to obtain the weighted cumulative energy ratio, which can be expressed by the formula: in, This represents the weighted cumulative energy percentage of the first j pulses after sorting, and L represents the observation window length. This represents the weighted cumulative energy value of the first j pulses within the observation window. This represents the energy density of the i-th pulse after sorting. This represents the confidence weight of the i-th pulse after sorting. This represents a smoothing factor to prevent the denominator from being zero. It takes the value 1e-6 and its dimensions are consistent with the denominator.

[0069] Based on the weighted energy accumulation ratio, the non-uniformity of suction resistance is calculated using the Lorentz curve principle. The calculation formula can be expressed as follows: in, The non-uniformity of suction resistance for the k-th pulse is represented by , j represents the j-th pulse sorted by energy density from smallest to largest, and L represents the observation window length. , These represent the weighted cumulative energy percentages of the first j and the first j-1 pulses after sorting, respectively. .

[0070] The area under the actual weighted energy cumulative percentage curve is used for discrete approximation calculation. The larger this area, the closer the actual energy distribution is to an ideal uniform distribution. By subtracting the area term from 1, the degree of approximation to uniformity can be converted into a degree of non-uniformity. Therefore, when the non-uniformity of grout suction resistance increases, it indicates a more uneven energy distribution; when the non-uniformity of grout suction resistance decreases, it indicates a more uniform energy distribution. That is, when a few pulses bear most of the weighted energy consumption, the non-uniformity of grout suction resistance increases, indicating that the grouting energy is mainly consumed in local high-resistance areas, rather than being mainly used for effective infiltration and compaction of soil pores. This can be used to identify the risk of false high pressure or deadlock.

[0071] Step S103: Determine the grouting condition based on the formation elastic rebound ratio, the flow resistance fluctuation rate, and the grout suction resistance non-uniformity, and generate corresponding pump control commands according to the determination results.

[0072] Grouting condition determination is based on three dimensions: formation elastic rebound ratio, flow resistance fluctuation rate, and grout suction resistance non-uniformity, to identify the current state of the grouting process. Different grouting states reflect different physical conditions of the formation and grouting channel, requiring differentiated control strategies. By prioritizing multi-dimensional features, conflicts caused by relying on single indicators can be avoided, accurately identifying the current grouting state and providing a basis for generating targeted control commands.

[0073] Pump control commands are generated by the system based on the grouting condition assessment and are used to adjust the operating parameters of the grouting pump. Different grouting states correspond to different control objectives, including maintaining current grouting parameters, reducing grout flow rate, and disrupting blockage structures. By generating pump control commands and implementing closed-loop regulation, a complete closed loop from monitoring to intervention can be achieved, providing early warning and intervention before macroscopic pressure exceeds limits, ensuring the continuity and effectiveness of grouting operations.

[0074] Furthermore, in some implementations of the present invention, the method for determining the grouting condition includes: If the non-uniformity of the suction resistance is greater than the preset blockage limit value, it is determined to be a false high-pressure deadlock condition. If the formation elastic rebound ratio is greater than the preset high resistance benchmark, it is determined to be a pseudo-high pressure rigid closure condition. If the formation elastic rebound ratio is less than the preset elasticity judgment benchmark, it indicates that the pressure dissipates rapidly, and it is judged as a smooth flow channel or an abnormal slurry leakage condition. If the flow resistance fluctuation rate is greater than or equal to the preset stability warning line, it is determined to be an unstable working condition of the orifice flow channel; Otherwise, it is determined to be an effective compaction condition.

[0075] Pseudo-high-pressure deadlock condition refers to a state where the grouting channel is severely blocked. When the non-uniformity of grout suction resistance exceeds the preset blockage limit, it indicates that the grouting energy distribution is extremely uneven, with most of the energy consumed in overcoming local high resistance. This is usually because grout particles filter and accumulate at the soil pore inlets around the grouting hole, forming a dense filter cake structure, leading to local deadlock of the grouting channel. At this point, regardless of elasticity or fluctuation characteristics, it is considered that the grouting channel can no longer deliver grout normally.

[0076] Pseudo-high-pressure rigid closure refers to a state where grout fails to effectively penetrate the soil skeleton. When the elastic rebound ratio of the formation is greater than the preset high-resistance benchmark, it usually means that the grout is blocked at the orifice or hard interlayer interface, and the formation exhibits rigid closure, lacking effective elastic rebound. Although the pressure may increase, it is not due to effective soil compaction, but rather to ineffective high-pressure accumulation caused by limited fluid channels.

[0077] When the elastic rebound ratio of the formation is less than the preset elasticity judgment benchmark, it indicates that the pressure dissipates very quickly and the formation exhibits the characteristics of unobstructed flow channels, which is judged as an unobstructed flow channel or abnormal slurry leakage condition.

[0078] The orifice flow channel instability condition refers to a state where the grouting channel is on the verge of instability but not yet completely blocked. When the flow resistance fluctuation rate is greater than or equal to the preset stability warning line, it indicates that the grout particles have undergone unstable accumulation and collapse oscillation at the pore inlet. The channel is not completely locked, but the flow resistance fluctuation is abnormally high, indicating that the particle accumulation process is in a dynamically unstable state. If no intervention is taken, it will quickly evolve into complete lockup.

[0079] Effective compaction refers to the state in which grout smoothly penetrates the soil pores and produces an effective compaction effect. When the non-uniformity of grout suction resistance does not exceed the preset blockage limit, the formation elastic rebound ratio is not lower than the preset elasticity judgment benchmark, and the flow resistance fluctuation rate does not reach the preset stability warning line, the formation exhibits normal elastic rebound, with small fluctuations in flow resistance and relatively uniform energy distribution. At this time, the pressure increase mainly stems from the improvement of soil strength, which is the effective state that grouting reinforcement aims to achieve.

[0080] As a specific example, the blockage limit is set to 0.4; the high resistance benchmark is set as follows: in adaptive mode, the high resistance benchmark is set to 200% to 300% of the average formation elastic rebound ratio during the first 5 pulses (effective compaction stage) of grouting; in empirical fixed value mode, considering that the pressure dissipation time is often several times the loading time when there is severe blockage, the high resistance benchmark can be directly set to 3.0 to 5.0; the elasticity judgment benchmark is set to 80% of the average formation elastic rebound ratio during the first 5 pulses of grouting; the stability warning line is set to twice the historical volatility benchmark.

[0081] Furthermore, in some implementations of the embodiments of the present invention, the method for generating the pump control command includes: For the pseudo-high pressure deadlock condition or pseudo-high pressure rigid closure condition, a filter cake destruction and cleaning command is generated. The filter cake destruction and cleaning command is a high-frequency micro-amplitude flutter command or an intermittent pulse cleaning command. For conditions where the flow channel is unobstructed or there is abnormal slurry flow, a slowdown and waiting-to-cure command is generated; In response to the unstable working condition of the orifice flow channel, a speed reduction coefficient is calculated based on the flow resistance fluctuation rate and the preset stability warning line, and a speed reduction and flow stabilization command is generated. For the effective compaction condition, a steady-state maintenance command is generated to maintain the current grouting parameters.

[0082] The filter cake destruction and cleaning command is a control command generated for pseudo-high-pressure deadlock or pseudo-high-pressure rigid closure conditions. When the grouting channel is blocked, grout particles form a dense filter cake structure at the orifice, leading to channel deadlock or rigid closure. The filter cake destruction and cleaning command utilizes fluid dynamics to impact and destroy the dense rigid filter cake structure. The filter cake destruction and cleaning command can be a high-frequency micro-amplitude flutter command or an intermittent pulse cleaning command. The high-frequency micro-amplitude flutter command is suitable for adjustable grouting pumps equipped with electronic control units. By controlling the grouting pump to perform high-frequency, short-stroke reciprocating motion, a high-frequency hydraulic water hammer wave is generated between the pipeline and the orifice, and the energy of the water hammer wave is used to break the filter cake structure accumulated at the orifice. The intermittent pulse cleaning command is suitable for conventional mechanical grouting pumps. By controlling the grouting pump to perform a cycle of emergency stop, waiting, and emergency start, the low-frequency water hammer effect generated by the large transient inertia of the fluid is used to destroy the rigid closed structure.

[0083] The slow-down and waiting-to-cure command is a control command for situations where the flow channel is unobstructed or there is abnormal slurry runoff. It is used to reduce the slurry flow rate to delay slurry runoff.

[0084] The speed reduction and flow stabilization command is a control command generated in response to the instability of the orifice flow channel. When the flow channel is in a critical instability state, the slurry particles undergo unstable accumulation and collapse oscillation at the pore inlet. The speed reduction and flow stabilization command reduces the kinetic energy and collision probability between particles by decreasing the slurry flow velocity, thereby slowing down the densification process of the accumulated structure.

[0085] The steady-state maintenance command is a control command generated for effective compaction conditions. When the stratum exhibits normal elastic rebound and the flow resistance fluctuation is small, the grout is smoothly penetrating into the soil pores and producing an effective compaction effect. The steady-state maintenance command keeps the current grouting pump stroke frequency and discharge parameters unchanged, and continues to perform routine grouting operations until the design grouting volume or the actual final pressure standard is reached.

[0086] As a specific example, for pseudo-high-pressure deadlock or pseudo-high-pressure rigid closure conditions, a filter cake destruction and cleaning command is generated. This command can be either a high-frequency micro-amplitude flutter command or an intermittent pulse cleaning command. The high-frequency micro-amplitude flutter command is suitable for adjustable grouting pumps equipped with electronic control units. The stroke is set to 10% to 15% of full scale, the operating frequency is set to 2 Hz to 4 Hz, and the operation continues for 10 to 15 seconds, utilizing high-frequency hydraulic water hammer waves to break up the filter cake structure. The intermittent pulse cleaning command is suitable for conventional mechanical grouting pumps, controlling the pump to perform a cycle of emergency stop, waiting for 2 to 3 seconds, and emergency start for 1 second, repeated 3 to 5 times, utilizing low-frequency water hammer effects to destroy the rigid closure structure.

[0087] For instability issues in the orifice flow channel, the system calculates a speed reduction coefficient based on the flow resistance fluctuation rate and a preset stability warning line, generating a speed reduction and flow stabilization command. This adjusts the grouting pump's stroke frequency to the product of the current frequency and the speed reduction coefficient. Specifically, the speed reduction coefficient is the larger of the stability warning line divided by the flow resistance fluctuation rate and 0.5. The stability warning line is set at twice the historical fluctuation rate benchmark. A negative feedback mechanism dynamically determines the speed reduction magnitude based on the degree of fluctuation exceeding the limit, with a minimum reduction to 50% of the original speed.

[0088] For effective compaction conditions, the system generates a steady-state maintenance command to keep the current grouting pump stroke frequency and discharge parameters unchanged, and continue to perform routine grouting operations until the designed grouting volume or the actual final pressure standard is reached.

[0089] Example 2: The present invention also provides a working condition monitoring and closed-loop control system for deep backfill grouting process, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the above-mentioned working condition monitoring and closed-loop control method for deep backfill grouting process.

Claims

1. A method for monitoring and closed-loop control of working conditions during deep backfill grouting, characterized in that, The method includes: The pressure and flow signals during the grouting process are collected synchronously. The time boundary of the grouting pump push stroke is determined based on the flow signal to obtain the loading segment and unloading segment of the current pulse. The pressure and flow signals in the loading segment and the unloading segment are filtered to obtain the loading segment pressure sequence, the loading segment flow sequence, and the unloading segment pressure sequence. By comparing the pressure sequence of the loading section and the pressure sequence of the unloading section, and based on the symmetrical influence of fluid viscous resistance on pressure loading and pressure dissipation, the formation elastic rebound ratio is obtained; the instantaneous flow resistance is calculated based on the pressure sequence and the flow rate sequence of the loading section, and the stability of the orifice flow channel is determined based on the dispersion of the instantaneous flow resistance, thus obtaining the flow resistance fluctuation rate; using the flow resistance fluctuation rate as the confidence weight, the energy distribution during the grouting process is analyzed, thus obtaining the non-uniformity of grout suction resistance. The grouting conditions are determined based on the formation elastic rebound ratio, the flow resistance fluctuation rate, and the non-uniformity of the grout suction resistance, and corresponding pump control commands are generated based on the determination results.

2. The method for monitoring and closed-loop control of working conditions in the deep backfill grouting process according to claim 1, characterized in that, The methods for obtaining the loading and unloading sections include: The moment when the flow signal crosses the preset flow validity threshold from low to high is taken as the start time of the push-out, and the moment when the flow signal crosses the preset flow validity threshold from high to low is taken as the end time of the push-out. The interval between the start time of the push-out and the end time of the push-out is taken as the loading segment, and the interval within the preset observation window after the end time of the push-out is taken as the unloading segment.

3. The method for monitoring and closed-loop control of working conditions in the deep backfill grouting process according to claim 1, characterized in that, The methods for obtaining the loading section pressure sequence, loading section flow sequence, and unloading section pressure sequence include: Calculate the grouting volume of the loading section. When the grouting volume is greater than or equal to the preset minimum grouting volume threshold, determine that the loading section is an effective loading section and the unloading section is an effective unloading section. The loading section pressure sequence is obtained based on the pressure signal within the effective loading section, the loading section flow sequence is obtained based on the flow signal within the effective loading section, and the unloading section pressure sequence is obtained based on the pressure signal within the effective unloading section.

4. The method for monitoring and closed-loop control of working conditions in the deep backfill grouting process according to claim 1, characterized in that, The method for calculating the formation elastic rebound ratio includes: Obtain the pressure peak value in the loading segment pressure sequence, obtain the time when the pressure value in the loading segment pressure sequence first reaches the pressure peak value at a preset first proportion and a preset second proportion, and obtain the pressure loading duration based on the time difference between the two times. The moment when the pressure value in the unloading section pressure sequence first reaches the preset third proportion is obtained. If such a moment exists, the pressure dissipation half-life is obtained based on the time difference between the moment and the start time of the unloading section. If the preset third proportion is not reached after exceeding the preset observation window, the duration of the observation window is taken as the pressure dissipation half-life. The formation elastic rebound ratio is calculated based on the pressure loading duration, the pressure dissipation half-life, and the preset regularization parameters.

5. The method for monitoring and closed-loop control of working conditions in the deep backfill grouting process according to claim 1, characterized in that, The method for calculating the flow resistance fluctuation rate includes: The instantaneous flow resistance sequence is calculated based on the loading section pressure sequence and the loading section flow rate sequence. The instantaneous flow resistance sequence is detrended to obtain the residual sequence. The standard deviation of the residual sequence and the arithmetic mean of the instantaneous flow resistance sequence are calculated. When the arithmetic mean is greater than or equal to a preset flow resistance threshold, the flow resistance fluctuation rate is obtained based on the standard deviation and the arithmetic mean. When the average value is less than the flow resistance threshold, the flow resistance fluctuation rate is set to zero.

6. The method for monitoring and closed-loop control of working conditions in the deep backfill grouting process according to claim 1, characterized in that, The method for calculating the non-uniformity of suction resistance includes: The energy density of the current pulse is calculated based on the loading segment pressure sequence and the loading segment flow sequence, and the confidence weight of the energy density is determined based on the flow resistance volatility and the preset historical volatility benchmark. The energy density distribution of multiple pulses within a preset observation window is compared with an ideal uniform distribution based on the confidence weight to obtain the non-uniformity of the suction resistance.

7. The method for monitoring and closed-loop control of working conditions in the deep backfill grouting process according to claim 6, characterized in that, The step of comparing the energy density distribution of multiple pulses within a preset observation window with an ideal uniform distribution based on confidence weights to obtain the non-uniformity of the suction resistance includes: The energy densities of multiple pulses within the preset observation window are arranged in ascending order, and the sorted energy densities are multiplied by the corresponding confidence weights to obtain the weighted energy value of each pulse. The weighted energy values ​​of the pulses at each position within the observation window and the previous pulses are accumulated to obtain the weighted energy accumulation value at each position. The weighted energy accumulation value at each position is then normalized with the total weighted energy of all pulses within the observation window to obtain the weighted energy accumulation percentage at each position. Based on the weighted cumulative energy ratio at each position within the observation window, a weighted cumulative energy ratio curve is constructed at each position within the observation window. The non-uniformity of the suction resistance is obtained based on the degree of deviation of the weighted cumulative energy ratio curve from the ideal uniform distribution.

8. The method for monitoring and closed-loop control of working conditions in the deep backfill grouting process according to claim 1, characterized in that, The methods for determining the grouting condition include: If the non-uniformity of the suction resistance is greater than the preset blockage limit value, it is determined to be a false high-pressure deadlock condition. If the formation elastic rebound ratio is greater than the preset high resistance benchmark, it is determined to be a pseudo-high pressure rigid closure condition. If the formation elastic rebound ratio is less than the preset elasticity judgment benchmark, it is judged as a smooth flow channel or an abnormal slurry leakage condition. If the flow resistance fluctuation rate is greater than or equal to the preset stability warning line, it is determined to be an unstable working condition of the orifice flow channel; Otherwise, it is determined to be an effective compaction condition.

9. The method for monitoring and closed-loop control of working conditions in the deep fill grouting process according to claim 8, characterized in that, The method for generating the pump control command includes: For the pseudo-high pressure deadlock condition or pseudo-high pressure rigid closure condition, a filter cake destruction and cleaning command is generated. The filter cake destruction and cleaning command is a high-frequency micro-amplitude flutter command or an intermittent pulse cleaning command. For the aforementioned conditions of unobstructed flow or abnormal slurry flow, a slowdown and waiting-to-cure command is generated; In response to the unstable working condition of the orifice flow channel, a speed reduction coefficient is calculated based on the flow resistance fluctuation rate and the preset stability warning line, and a speed reduction and flow stabilization command is generated. For the effective compaction condition, a steady-state maintenance command is generated to maintain the current grouting parameters.

10. A condition monitoring and closed-loop control system for deep backfill grouting process, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the working condition monitoring and closed-loop control method for deep backfill grouting process as described in any one of claims 1 to 9.

Images