Real-time monitoring method, system and storage medium for jet grouting pile construction quality

Abstract

The present invention provides a real-time monitoring method, system and storage medium for the construction quality of jet grouting piles, and relates to the technical field of jet grouting pile construction quality. The present invention deploys pressure, flow and microseismic sensors at the jet grouting pile pipeline outlet and drill bit, collects multi-dimensional time series data with a high signal-to-noise ratio, and adaptively filters and drift corrects it. Based on Takens phase space reconstruction, dynamic trajectories are generated, rheological entropy is calculated, and a four-dimensional space-time diffusion field is constructed. After projecting the field onto the construction plane, the underground cave network model is continuously corrected using Bayesian updating, multi-level warnings are triggered according to quantitative classification of entropy values ​​and filling rates, and a filling quality heat map is generated, thereby achieving high-precision, dynamic and visual monitoring and guidance of the slurry diffusion process.

Description

Technical Field

[0001] The invention relates to the technical field of jet grouting pile construction quality, and in particular to a real-time monitoring method, system and storage medium for jet grouting pile construction quality. Background Art

[0002] Jet-grouting pile construction is a key technology widely used for reinforcing weak foundations and filling karst areas. Its construction quality directly determines the foundation's bearing capacity and long-term stability. Traditional construction acceptance often relies on coring and static load testing. This process is not only cumbersome and time-consuming, but also fails to accurately reflect the distribution of grout in complex underground structures, such as caves and fissures.

[0003] Monitoring the quality of jet grouting pile construction has traditionally relied on single physical methods such as coring, static load testing, or transient electromagnetic and acoustic penetration. These methods not only have long sampling cycles and cumbersome processes, but because they only measure the response of a single path, they struggle to capture the coupled changes in pressure, flow, and multidimensional formation microseismic information during the grouting process. They also fail to promptly reflect the dynamic characteristics of nonlinear slurry diffusion within complex underground cave and fracture networks. Even some systems that have added pressure or flow threshold alarms can only provide simple over-limit alerts for a single signal, lacking in-depth exploration of high-frequency disturbances and non-stationary behavior, making it difficult to accurately assess cave connectivity, slurry replacement rates, and their spatiotemporal distribution.

[0004] Existing technologies lack the four-dimensional field modeling method that combines multi-source time series data with spatial coordinates and construction timestamps. The detection results are mostly discrete points or static curves, which cannot generate high-resolution construction plane distribution maps, nor can they iterate and update monitoring data and geological models in real time. There is no graded early warning and visual closed loop, making it difficult to quickly convert monitoring findings into grouting parameter adjustments and on-site construction scheduling, which leads to a disconnect between construction and acceptance, and makes it impossible to implement continuous, dynamic, and guided quality assessment and risk warning of the slurry diffusion process.

[0005] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not form the prior art that is already known to a person of ordinary skill in the art. Summary of the Invention

[0006] The object of the present invention is to provide a method, system and storage medium for real-time monitoring of the construction quality of jet grouting piles to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A method for real-time monitoring of the construction quality of jet grouting piles, comprising the following steps:

[0009] Step 1: Collect the grouting pressure, grouting flow rate and microseismic signals of each drilling point during the grouting process to form a multi-dimensional time series signal, and pre-process the multi-dimensional time series signal, which includes adaptive filtering and drift correction;

[0010] Step 2: Based on the Takens phase space reconstruction principle, the preprocessed multidimensional time series signal is subjected to delayed embedding to generate a phase space vector sequence, revealing the dynamic trajectory of the slurry flow process. Based on this trajectory, the distance distribution of adjacent vectors in the trajectory is calculated, and the rheological entropy value representing the degree of disordered diffusion of the slurry is output. Combined with the borehole spatial coordinates and the construction timestamp, the spatiotemporal field of the slurry diffusion behavior is constructed;

[0011] Step 3: Project the spatiotemporal field of slurry diffusion behavior onto the construction plane to generate an entropy distribution map. Based on the entropy distribution map, the underground cave network model is iteratively modified through the Bayesian update mechanism. Multi-level entropy thresholds are set to trigger a three-level warning. Based on the entropy distribution map and the model fitting results, the cave filling quality grade of each drilling area is output;

[0012] Step 4: Issue a warning notification of the corresponding level based on the cave filling quality level, overlay the quality level information on the construction plane, and generate an interactive filling quality heat map.

[0013] Furthermore, the logic of adaptive filtering and drift correction for multi-dimensional time series signals is as follows:

[0014] Using three times the length of the grouting pump's working cycle as a sliding window, the multidimensional time series signal is divided into data segments of equal step length according to the sliding window length. For each data segment, the following operations are performed: 1) Intrinsic mode functions are extracted sequentially through empirical mode decomposition, and the first three modes containing mechanical vibrations of the drilling rig and pipeline are removed to obtain the residual signal after vibration removal;

[0015] 2) For the residual grouting pressure signal, a polynomial curve of no higher than fifth order is used to fit its baseline drift. This baseline curve is subtracted from the original pressure signal to obtain a drift-corrected pressure signal.

[0016] 3) For the residual grouting flow signal, a two-way zero-phase digital low-pass filter with a cutoff frequency of 10 Hz is used to process it to obtain the drift-corrected grouting flow signal to retain the high-frequency components reflecting the grouting mutation;

[0017] 4) For the residual microseismic signal, retain the original vibration mode, design a 1-200 Hz bandpass filter to suppress power frequency interference and high-frequency noise, and output the filtered microseismic signal.

[0018] Furthermore, the logic of takingns phase space reconstruction of the drift-corrected pressure signal is as follows:

[0019] The drift-corrected grouting pressure signal is used as the main channel. Its autocorrelation function is calculated and the moment corresponding to the first zero crossing point is taken as the time delay of phase space reconstruction. Based on this time delay, a pseudo nearest neighbor discrimination algorithm is used to incrementally test different phase space reconstruction dimensions. The proportion of false nearest neighbors is monitored in each trial. When this proportion drops below 5%, the final phase space reconstruction dimension is determined.

[0020] Based on the determined time delay and reconstruction dimension, the delay embedding method is used to generate phase space vector sequences for the drift-corrected grouting pressure signal, grouting flow signal, and filtered microseismic signal. The Euclidean distance between two adjacent points in the phase space vector sequence of each signal is calculated in turn to obtain the distance sequence of the pressure signal, grouting flow, and microseismic signal in the phase space.

[0021] For each phase space, the distance interval width is set , and combined with the maximum value of the Euclidean distance in the phase space to divide K distance intervals of equal width, the Euclidean distance values ​​between each adjacent vector pair are divided into corresponding distance intervals, and the number of distance occurrences in each distance interval is counted, thereby obtaining a distance distribution histogram.

[0022] Furthermore, the specific logic for calculating the rheological entropy value based on the distance distribution obtained by phase space reconstruction is as follows:

[0023] Grouting pressure , grouting flow and microseismic signals The phase space distance distribution is performed separately:

[0024] Set the length of the phase space vector sequence of the xth type to , then the total number of adjacent vector pairs is , count the logarithm of the Euclidean distance of the xth signal falling into the kth distance interval , calculate the probability corresponding to the distance interval , , and satisfy the normalization condition ;

[0025] Calculate the rheological entropy value that characterizes the degree of disorder in distance distribution for:

[0026] ;

[0027] in, Is the index of the signal type, 1, 2, 3 correspond to , is the index of the distance interval, is the total number of distance intervals, which is determined by the maximum Euclidean distance of the x-th signal The width of the distance interval OK, that is ;

[0028] Generate a comprehensive rheological entropy value representing the global diffusion disorder through weighted fusion : , are weight coefficients, and ,when When , it is determined that the diffusion behavior of the slurry in the phase space is abnormal. .

[0029] Furthermore, the logic of constructing the space-time field of slurry diffusion behavior includes:

[0030] For each drilling point i, its spatial coordinates are known and its comprehensive rheological entropy , and form an observation set of the spatiotemporal field of slurry diffusion behavior ;

[0031] Any point to be interpolated in the construction area is recorded as ,from Select n observation points with the closest time and space distance to form the observation point set , j is the index of the observation point in the observation point set, n is the number of observation points in the observation point set, where ;

[0032] Define the weighted space-time distance between the jth observation point and the point to be interpolated: is the space-time distance between the point to be interpolated and the j-th observation point, ;

[0033] Based on time and space distance Calculate the interpolation weights: , is the spatial correlation coefficient;

[0034] The comprehensive rheological entropy value of the interpolation point at the time-space coordinate point is obtained by weighted average interpolation:

[0035] ;

[0036] in, is the comprehensive rheological entropy value of the interpolation point, is the comprehensive rheological entropy value at the jth observation point.

[0037] Furthermore, the drilling points and interpolation points at the current moment are selected in the construction plane, and their comprehensive rheological entropy values ​​are calculated. Projecting onto a two-dimensional plane grid and generating an entropy distribution map, the grid resolution is aligned with the cave model;

[0038] The Bayesian updating mechanism is used to modify the underground cave network model, including:

[0039] The state parameter vector of the underground cave network is defined as ,in is the grouting filling rate, is the cave connectivity index, is the equivalent diameter of the cave;

[0040] Divide the entropy distribution map into M equal-area grids, and the index of the grid , the comprehensive rheological entropy of the oth grid center is , the maximum possible entropy value is recorded as ;

[0041] Assume that the observation noise follows a Gaussian distribution with mean 0 and variance Gaussian distribution, then calculate the likelihood function:

[0042] ;

[0043] in, Based on the current Theoretical entropy value prediction at the oth grid under the parameter;

[0044] Defining the prior distribution , then iteratively update the posterior:

[0045] ;

[0046] in, For Integrate over the domain of ;

[0047] Iterate several times until the posterior distribution converges, and then take the optimal model parameters Update cave network model parameters.

[0048] Furthermore, by combining the entropy distribution map with the optimal model parameters obtained through Bayesian updating, the cave filling quality of each drilling area is graded and an early warning is triggered. At the same time, a heat map is generated on the construction plane. The specific logic is as follows:

[0049] For each grid o of each borehole, its comprehensive rheological entropy value and optimal fill rate , implement four-level assessment:

[0050] Excellent: and , good level: and ,intermediate: and , difference: and ;

[0051] Multi-level warning trigger: Level 1 warning: any single grid within 5 minutes ; Second level warning: define adjacent grids to form a connected area ,and , average rheological entropy ; Level 3 warning: Fill rate after Bayesian update And the corresponding posterior probability ;

[0052] On the construction plane, the grid is used as a unit, and the quality level of each grid is superimposed with different color levels, and rendered with excellent, good, medium and poor colors. When the early warning is triggered, the corresponding grid is automatically highlighted and a notification is pushed.

[0053] The present invention further provides a real-time monitoring system for the construction quality of jet grouting piles, which is used to implement the above-mentioned real-time monitoring method for the construction quality of jet grouting piles, comprising:

[0054] The data acquisition module is used to collect the grouting pressure, grouting flow rate and microseismic signals of each drilling point during the grouting process, form a multi-dimensional time series signal, and pre-process the multi-dimensional time series signal, including adaptive filtering and drift correction;

[0055] The entropy calculation module is used to generate a phase space vector sequence through delayed embedding based on the Takens phase space reconstruction principle for the preprocessed multi-dimensional time series signal, revealing the dynamic trajectory of the slurry flow process. Based on this trajectory, the distance distribution of adjacent vectors in the trajectory is calculated, and the rheological entropy value representing the degree of disordered diffusion of the slurry is output. Combined with the borehole spatial coordinates and the construction timestamp, the spatiotemporal field of the slurry diffusion behavior is constructed;

[0056] The model evaluation module is used to project the spatiotemporal field of slurry diffusion behavior onto the construction plane to generate an entropy distribution map. Based on the entropy distribution map, the underground cave network model is iteratively modified through a Bayesian update mechanism. Multi-level entropy thresholds are set to trigger a three-level warning. Based on the entropy distribution map and the model fitting results, the cave filling quality grade of each drilling area is output;

[0057] The early warning judgment module is used to issue early warning notifications of corresponding levels based on the cave filling quality level, superimpose quality level information on the construction plane, and generate an interactive filling quality heat map.

[0058] The present invention also provides a non-volatile computer-readable storage medium containing computer-executable instructions. When the computer-executable instructions are executed by one or more processors, the processors execute the method for real-time monitoring of the construction quality of rotary jet piles.

[0059] Compared with the prior art, the present invention has the following beneficial effects:

[0060] The present invention synchronously collects three signals of pressure, flow and microseismicity at the pipeline outlet and drill bit of the jet grouting pile construction equipment, and effectively eliminates mechanical vibration and sensor drift noise through preprocessing methods such as empirical mode decomposition, adaptive filtering and drift correction, thereby obtaining multi-dimensional time series data with a high signal-to-noise ratio; based on these data, Takens phase space reconstruction is used to extract the dynamic trajectory, and the rheological entropy is calculated to quantitatively characterize the degree of disorder in the slurry diffusion process, which can deeply explore the high-frequency disturbance and nonlinear diffusion characteristics, and realize high-resolution imaging of the dynamic evolution of slurry in complex cave-crack networks.

[0061] The present invention combines the rheological entropy value at each borehole with the spatial coordinates and timestamps to construct a four-dimensional space-time field, and continuously corrects the underground cave network model through a Bayesian update mechanism, so that the model parameters (filling rate, connectivity, equivalent diameter) are closely matched with real-time monitoring data, forming a high-precision, dynamic spatial mapping of the grouting effect.

[0062] The present invention combines the entropy distribution map projected onto the construction plane with the updated model, which can not only give a quantitative classification of the filling quality of each area, but also intuitively present it in the form of a heat map on the plane, providing a scientific and closed-loop decision-making basis for subsequent grouting parameter optimization and construction resource scheduling. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 Schematic diagram of the overall method flow of the present invention;

[0064] Figure 2 This is a line graph of pressure-flow-microseismic rheological entropy values ​​of the present invention;

[0065] Figure 3 This is a fitting curve diagram of the pressure rheological entropy value-comprehensive rheological entropy value of the present invention;

[0066] Figure 4 It is a line graph of pressure rheological entropy value-comprehensive rheological entropy value of the present invention;

[0067] Figure 5 This is a graph showing the microseismic rheological entropy value-comprehensive rheological entropy value diagram of the present invention;

[0068] Figure 6 This is a flow chart of the overall system module of the present invention. DETAILED DESCRIPTION

[0069] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to specific embodiments.

[0070] It should be noted that, unless otherwise defined, the technical or scientific terms used in the present invention should have the usual meanings understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative position relationships. When the absolute position of the object being described changes, the relative position relationship may also change accordingly.

[0071] Example:

[0072] See also Figure 1-Figure 5 , the present invention provides a technical solution:

[0073] A method for real-time monitoring of the construction quality of jet grouting piles, comprising the following steps:

[0074] Step 1: Collect the grouting pressure, grouting flow rate and microseismic signals of each drilling point during the grouting process to form a multi-dimensional time series signal, and pre-process the multi-dimensional time series signal, which includes adaptive filtering and drift correction;

[0075] Sensor deployment: Pressure and flow sensors are deployed at the outlet of the grouting pipeline to monitor the state of the slurry output by the pumping system, and microseismic sensors are deployed at the drill bit to capture the geotechnical response during the drilling process;

[0076] Pressure sensors capture changes in resistance when slurry breaks through the formation, such as a sudden pressure increase during the initial stage of cave filling and a sudden pressure drop when the cave is penetrated. Flow sensors monitor the stability of slurry delivery. A sudden increase in flow may indicate a breakthrough of the cave, while a sudden drop suggests slurry leakage. Microseismic sensors receive signals of rock fracture. For example, vibrations in the 1-200 Hz frequency range can indicate shearing of the cave boundary or filling shock.

[0077] The logic of adaptive filtering and drift correction for multi-dimensional time series signals is:

[0078] Using three times the length of the grouting pump's working cycle as a sliding window, the multidimensional time series signal is divided into data segments of equal step length according to the sliding window length. For each data segment, the following operations are performed: 1) Intrinsic mode functions are extracted sequentially through empirical mode decomposition, and the first three modes containing mechanical vibrations of the drilling rig and pipeline are removed to obtain the residual signal after vibration removal;

[0079] 2) For the residual grouting pressure signal, a polynomial curve of no higher than fifth order is used to fit its baseline drift. This baseline curve is subtracted from the original pressure signal to obtain a drift-corrected pressure signal.

[0080] 3) For the residual grouting flow signal, a two-way zero-phase digital low-pass filter with a cutoff frequency of 10 Hz is used to process it to obtain the drift-corrected grouting flow signal to retain the high-frequency components reflecting the grouting mutation;

[0081] 4) For the residual microseismic signal, retain the original vibration mode, design a 1-200 Hz bandpass filter to suppress power frequency interference and high-frequency noise, and output the filtered microseismic signal;

[0082] Grouting pumps typically reciprocate their pistons at a specific cycle. This vibration cycle encompasses both the mechanical motion itself and the pressure fluctuations within the pipeline. Using this cycle as a benchmark, segmenting the signal allows each segment to contain several complete pump strokes, facilitating the following modal decomposition:

[0083] The window length is set at three times the grouting pump's operating cycle. For example, a 2-second cycle pump uses a 6-second window to ensure coverage of the complete pumping pulsation and transition process. The original signal is processed in segments to avoid the aliasing of mechanical interference and valid events in long-term signals. The reason for the triple length is that shorter than one cycle may not be able to completely separate the pump vibration mode, and too long a segment will introduce more field changes and disrupt the stability of the signal within the segment. Triple the period to cover sufficient modal cycles while taking into account local stability.

[0084] The EMD (Empirical Mode Decomposition) de-oscillation method decomposes complex signals into a series of IMFs (Intrinsic Mode Functions), each of which corresponds to a different frequency band and amplitude envelope of the signal, without the need to pre-set basis functions such as Fourier or wavelet.

[0085] The first to third order IMFs mostly correspond to high-frequency, large-amplitude vibrations, such as mechanical vibrations from drilling motors, pipeline resonance, and pump shock. After removing these order IMFs, the residual signal can better highlight the subtle responses of sudden seepage changes to pressure and flow during grouting.

[0086] For each signal segment, the first three IMFs (intrinsic mode functions) are stripped: IMF1, which is high-frequency circuit noise greater than 200Hz, such as sensor power supply interference; IMF2, which is 20-50Hz drilling rig gear vibration, such as hydraulic motor meshing frequency; and IMF3, which is 5-15Hz pipeline hydraulic oscillation, that is, pressure waves caused by the opening and closing of pump valves. The retained residual signal focuses on the interaction events between the slurry and the formation. The residual signal is equal to the original signal minus IMF1 + IMF2 + IMF3.

[0087] Mechanical vibration energy is concentrated in low-frequency modes, namely IMF1 to IMF3, while grout dynamic events, such as pressure surges and rock fractures, are common in IMF4 to IMF6 (80-200Hz). Microseismic signals cannot be processed using EMD because effective microseismic events, such as 35Hz rock fractures, overlap with the mechanical vibration frequency band, and EMD may mistakenly eliminate effective components.

[0088] The pressure signal is fitted with a polynomial of order no higher than five to account for the slowly varying baseline caused by sensor temperature drift and viscosity gradient. After deducting the baseline, the true fluctuations, such as changes in cave filling resistance, are highlighted. The flow signal is filtered with a 10Hz low-pass filter to eliminate pumping pulsations, while retaining sudden fluctuations greater than 10Hz.

[0089] Polynomials with an order no higher than fifth are simple to fit and have fewer parameters, making them suitable for aligning slowly changing baseline drifts. Polynomials with orders higher than fifth are prone to overfitting and may introduce false trends. Polynomials with orders less than fifth are sufficient to fit most temperature drift and zero drift curves.

[0090] Sudden changes in flow rate can indicate blockage or funneling, but equipment vibrations can also be superimposed in the mid- and high-frequency bands. Using bidirectional zero-phase low-pass filtering and a 10Hz cutoff digital filter, forward-reverse filtering can eliminate phase distortion and preserve the time location of sudden events.

[0091] The core value of bandpass filtering of microseismic signals to retain the 1-200Hz frequency band is to filter out low-frequency interference, i.e., drilling rig displacement vibration less than 1Hz, suppress high-frequency noise, i.e., electrical interference greater than 200Hz, and lock in effective events, i.e., rock shear waves (cave boundary destruction) of 30-120Hz and slurry cavity shock waves of 50-80Hz.

[0092] Step 2: Based on the Takens phase space reconstruction principle, the preprocessed multidimensional time series signal is subjected to delayed embedding to generate a phase space vector sequence, revealing the dynamic trajectory of the slurry flow process. Based on this trajectory, the distance distribution of adjacent vectors in the trajectory is calculated, and the rheological entropy value representing the degree of disordered diffusion of the slurry is output. Combined with the borehole spatial coordinates and the construction timestamp, the spatiotemporal field of the slurry diffusion behavior is constructed;

[0093] Phase space reconstruction is the process of converting one-dimensional time series signals into high-dimensional dynamic trajectories, aiming to reveal the hidden laws of slurry diffusion. Specifically, pressure trajectories reflect the energy fluctuations of slurry breaking through the formation, flow trajectories describe the continuity of slurry transport, and microseismic trajectories characterize the response of geotechnical structures. Placing them separately in phase space is like drawing flowing water into three-dimensional streamlines, which allows for more intuitive observation of turning points, mutations, or blockages in the diffusion process.

[0094] The logic of takingns phase space reconstruction of the drift-corrected pressure signal is:

[0095] The drift-corrected grouting pressure signal is used as the main channel. Its autocorrelation function is calculated and the moment corresponding to the first zero crossing point is taken as the time delay of phase space reconstruction. Based on this time delay, a pseudo nearest neighbor discrimination algorithm is used to incrementally test different phase space reconstruction dimensions. The proportion of false nearest neighbors is monitored in each trial. When this proportion drops below 5%, the final phase space reconstruction dimension is determined.

[0096] The first step in embedding a time series signal into a multidimensional space is to determine the delay, or the time interval between adjacent coordinate components on the time axis. The autocorrelation function is calculated using the main channel, the drift-corrected pressure signal, and the time corresponding to the first zero-crossing point is found. This point is exactly where the signal is least correlated with itself, preserving dynamic information while avoiding excessive redundancy between adjacent components.

[0097] If the time delay is too small, the reconstructed point clusters will overlap, preventing the complete trajectory from being visible. If the time delay is too large, the continuity of the process will be lost. The first zero-crossing point provides the golden time to balance the two. For example, when photographing the construction process of jet grouting piles, if the time delay is too large, the interval between photos will be too long, and the key frame of cave penetration will be missed. If the delay is too small, the photos taken in succession will all overlap, and only the drill bit but not the drilling process will be visible. The first zero-crossing point is equal to the optimal shooting interval, which allows each action to be clearly seen while ensuring continuity.

[0098] Based on the determined time delay and reconstruction dimension, the delay embedding method is used to generate phase space vector sequences for the drift-corrected grouting pressure signal, grouting flow signal, and filtered microseismic signal. The Euclidean distance between two adjacent points in the phase space vector sequence of each signal is calculated in turn to obtain the distance sequence of the pressure signal, grouting flow, and microseismic signal in the phase space.

[0099] In the pseudo nearest neighbor algorithm, in low dimensions, many points that should be far apart in high-dimensional space are projected to their nearest neighbors. As the embedding dimension increases, these false neighbors gradually separate. When the proportion of pseudo nearest neighbors drops below 5%, it means that the spatial relationships of most points have been correctly restored and the dimensionality increase can be stopped. This method ensures that the embedding space is large enough to accommodate the complexity of the system without wasting computing resources due to multiple dimensions.

[0100] After determining the time delay and embedding dimension, not only can the pressure signal be delayed and embedded, but flow and microseismic signals can also be incorporated into the same framework to generate multiple sets of phase space vectors separately or jointly. This has two advantages: different channels will simultaneously display the joint dynamic characteristics of seepage, vibration, and pressure at the same time; if a single signal is affected by noise, details will be lost. Multi-channel joint embedding can complement each other and enhance the robustness of reconstruction.

[0101] For each phase space, the distance interval width is set , and combined with the maximum value of the Euclidean distance in the phase space to divide K distance intervals of equal width, the Euclidean distance values ​​between each adjacent vector pair are divided into corresponding distance intervals, and the number of distance occurrences in each distance interval is counted, thereby obtaining a distance distribution histogram.

[0102] Considering the trajectory in phase space as a continuously advancing curve, calculating the Euclidean distance between two adjacent points is quantifying the size of the jump in the system state. If the slurry diffuses smoothly, the distance distribution is relatively concentrated; if turbulence, channel blockage or rapid leakage occurs, a larger jump will occur and the distance will become more dispersed. Dividing all adjacent distances into the same interval and counting the number of occurrences in each interval will produce a distance distribution histogram, and the concept of "entropy" is then used to measure its disorder. The more uniform the distribution (the more disordered), the higher the rheological entropy value. This value is called "rheological entropy", which intuitively depicts the degree of diffusion disorder of the slurry in phase space.

[0103] Each drilling sensor has fixed coordinates in space and generates a rheological entropy value at each moment. Combining these data constitutes a four-dimensional diffusion behavior observation field.

[0104] The specific logic for calculating the rheological entropy value based on the distance distribution obtained by phase space reconstruction is as follows:

[0105] Grouting pressure , grouting flow and microseismic signals The phase space distance distribution is performed separately:

[0106] Assume that the length of the phase space vector sequence of the x-th signal is , then the total number of adjacent vector pairs is , count the logarithm of the Euclidean distance of the xth signal falling into the kth distance interval , calculate the probability corresponding to the distance interval , , and satisfy the normalization condition , which ensures that all intervals together cover all possible distances;

[0107] Reflects the duration of the signal dynamics behavior, The larger it is, the longer the signal observation time is, and the more complete the trajectory evolution is. The bigger, The bigger it is, the longer the jet grouting pile drilling time will be. The larger it is, the higher the statistical reliability;

[0108] Describe the statistical characteristics of the phase space trajectory distance distribution, Concentration, stable state transition (ordered diffusion), Dispersion, state transition chaos (disordered diffusion); The bigger, The bigger;

[0109] Calculate the rheological entropy value that characterizes the degree of disorder in distance distribution for:

[0110] ;

[0111] in, Is the index of the signal type, 1, 2, 3 correspond to , is the index of the distance interval, is the total number of distance intervals, which is determined by the maximum Euclidean distance of the x-th signal The width of the distance interval OK, that is ;

[0112] Low entropy When it is close to 0, it is highly ordered (uniform formation and stable diffusion), and most The trajectory jump amplitude is relatively simple and predictable, which indicates that the system state changes are relatively stable and high entropy near When , there is high disorder (caves / cracks chaotic diffusion), The more uniform distribution indicates that the state jump often spans multiple amplitude levels and the process is more chaotic, indicating that complex phenomena such as turbulence, blockage reorganization or channel instability may occur in the slurry diffusion;

[0113] If the maximum distance of a certain signal increases, that is, the Leading to more boxes or wider full range, without adding high concentration points in the box, may make the distribution more dispersed, thereby pushing up If the concentration of boxes increases, some becomes larger, the others become smaller, then will decline;

[0114] Describes the discretization accuracy of the phase space trajectory, Small, coarse-grained analysis (suitable for simple strata), Large, fine-grained analysis (suitable for complex caves); maximum distance Increase, distance interval width The bigger, The bigger;

[0115] Generate a comprehensive rheological entropy value representing the global diffusion disorder through weighted fusion : , are weight coefficients, and ,when When , it is determined that the diffusion behavior of the slurry in the phase space is abnormal. ;

[0116] when If the grouting pressure is large, it means that the grouting pressure is dominant. Attention should be paid to abnormal grouting resistance. Large, that is, microseismic signals dominate, and attention should be paid to the risk of rock mass fracture. Increase means that the diffusion behavior of the entire grouting-formation system is more disordered and unstable. A decrease indicates that the system runs more smoothly and is controllable; when Sudden increase, If it increases significantly, the specific signal abnormality source is located;

[0117] The quantitative threshold for describing the uncontrolled diffusion of slurry is 75% of the theoretical maximum disorder, which is considered abnormal. When the value exceeds this value, it indicates that the distance distribution is close to uniform and the jumps of the three signals are highly chaotic, suggesting that serious leakage, blockage switching or sudden formation rupture may have occurred.

[0118] The specific data of some sample numbers and comprehensive rheological entropy values ​​are shown in Table 1.

[0119] Table 1

[0120]

[0121] By analyzing the first 15 rows of data, we can observe a certain correlation between different rheological entropy values. For example, the data shows that there is a certain positive correlation between the pressure rheological entropy value and the comprehensive rheological entropy value. As the sample number increases, the pressure rheological entropy value fluctuates between 0.14 and 0.22, and the comprehensive rheological entropy value also varies between 0.19 and 0.25. In particular, for sample number 14, the pressure rheological entropy value reaches 0.17, while the comprehensive rheological entropy value is 0.25. This indicates that under higher pressure rheological entropy conditions, the comprehensive rheological entropy value also increases, which may mean that the fluidity and overall performance of the fluid are improved under higher pressure environments.

[0122] When analyzing the relationship between flow rate rheological entropy and microseismic rheological entropy, we found that as flow rate rheological entropy increased, so did microseismic rheological entropy. For example, sample number 5 had a flow rate rheological entropy of 0.30 and a corresponding microseismic rheological entropy of 0.18, while sample number 1 had a flow rate rheological entropy of 0.20 and a microseismic rheological entropy of 0.25. This suggests that an increase in flow rate may affect the intensity of microseismic signals, reflecting the vibration characteristics generated by the fluid during movement.

[0123] The relationship between the comprehensive rheological entropy and the flow rheological entropy is also noteworthy. As the sample number increases, the comprehensive rheological entropy fluctuates between 0.20 and 0.26, while the flow rheological entropy also increases from 0.20 to 0.35. This trend suggests that changes in flow have a certain impact on the overall rheological properties.

[0124] In summary, the relationships and trends between these rheological entropy values ​​can provide certain guidance for optimizing fluid flow characteristics and are worthy of further exploration and verification in practical applications.

[0125] The logic of constructing the space-time field of slurry diffusion behavior includes:

[0126] For each drilling point i, its spatial coordinates are known and its comprehensive rheological entropy , and form an observation set of the spatiotemporal field of slurry diffusion behavior ;

[0127] Represents the overall diffusion disorder of the i-th borehole at this moment. The larger the degree, the more chaotic and unstable the slurry behavior at this location is. The larger the value, the greater the risk of leakage, turbulence or damage to the ground structure in the area where the drilling point is located. The smaller it is, the smoother and more expected the grouting behavior is;

[0128] Any point to be interpolated in the construction area is recorded as ,from Select n observation points with the closest time and space distance to form the observation point set , j is the index of the observation point in the observation point set, n is the number of observation points in the observation point set, where ; Number of samples The larger the value, the wider the interpolation range but the smoother the result may be. The smaller the value, the more localized but the greater the risk of being affected by a single point anomaly;

[0129] Define the weighted space-time distance between the jth observation point and the point to be interpolated: is the space-time distance between the point to be interpolated and the j-th observation point, ;

[0130] 、 Both represent horizontal plane distances. Indicates the control vertical distance influence, The smaller it is, the closer this point is to the point to be interpolated in time and space, and a higher interpolation weight should be given. The larger it is, the lower the interpolation contribution;

[0131] Based on time and space distance Calculate the interpolation weights: , is the spatial correlation coefficient;

[0132] Used to measure the contribution of the j-th observation value to the result of the estimated point, exponential decay function With the distance in time and space Increase and decrease rapidly, The larger it is, the more sensitive the parameter is to distance. The weight of small distance samples is extremely large, and the weight of long distance samples is close to zero. The smaller the value, the more evenly the weight distribution is, and more distant points also contribute; the sum of the denominators ensures , so that the cumulative contribution of all sample weights is 100%;

[0133] The comprehensive rheological entropy value of the interpolation point at the time-space coordinate point is obtained by weighted average interpolation:

[0134] ;

[0135] in, is the comprehensive rheological entropy value of the interpolation point, is the comprehensive rheological entropy value at the jth observation point;

[0136] Represents the comprehensive rheological entropy value of the interpolation point, that is, the comprehensive disorder evaluation value obtained by time-space weighted average based on the surrounding n observation points. The higher, The greater the contribution to the result, if the point to be estimated is close to a high entropy observation point, then will be pulled up, on the contrary, if they are close to low entropy points, then Lower; The larger it is, the more chaotic and unstable the slurry diffusion behavior is at that location and at that time; The smaller the size, the smoother and more predictable the behavior; by traversing the entire construction plane or three-dimensional space and calculating , you can draw a continuous entropy field map, identify high-risk areas and time windows, and support visual monitoring and early warning triggering.

[0137] Step 3: Project the spatiotemporal field of slurry diffusion behavior onto the construction plane to generate an entropy distribution map. Based on the entropy distribution map, the underground cave network model is iteratively modified through the Bayesian update mechanism. Multi-level entropy thresholds are set to trigger a three-level warning. Based on the entropy distribution map and the model fitting results, the cave filling quality grade of each drilling area is output;

[0138] Select the drilling point and interpolation point at the current moment in the construction plane, and use their comprehensive rheological entropy values Projecting onto a two-dimensional plane grid and generating an entropy distribution map, the grid resolution is aligned with the cave model;

[0139] Comprehensive rheological entropy value at the center of the grid The larger it is, the higher the possibility of abnormal slurry diffusion, turbulence or channel instability in the region is; the smaller it is, the diffusion behavior is stable and orderly;

[0140] The Bayesian updating mechanism is used to modify the underground cave network model, including:

[0141] The state parameter vector of the underground cave network is defined as ,in is the grouting filling rate, is the cave connectivity index, is the equivalent diameter of the cave;

[0142] Grouting filling rate The closer it is to 1, the cave is almost filled with slurry, and the closer it is to 0, the less filled it is; The larger the value, the denser and more highly connected the cave network is. The larger the value, the larger the size of the single pore or connected body;

[0143] Describes the current structural state of the underground cave network after grouting, As the size increases, more space is occupied by the slurry and the compaction is more complete. As the pores increase, the interconnection between the cave and the surrounding pores becomes higher, and there are more lubrication and seepage channels. Increase, the volume of a single cavity or a continuous area increases;

[0144] Divide the entropy distribution map into M equal-area grids, and the index of the grid , the comprehensive rheological entropy of the oth grid center is , the maximum possible entropy value is recorded as ;

[0145] Assume that the observation noise follows a Gaussian distribution with mean 0 and variance Gaussian distribution, then calculate the likelihood function:

[0146] ;

[0147] in, Based on the current Theoretical entropy value prediction at the oth grid under the parameter;

[0148] is determined by the current model parameters Through phase space diffusion simulation and projection calculation, Given a model state , the probability of observing this set of entropy distribution; for a certain grid o, Measuring model predictions and The degree of agreement between , then the exponential term is close to , the likelihood is large, the model state The explanatory power of this grid is strong; when Much greater than , the exponential term approaches 0, the likelihood is small, indicating that the model is difficult to explain the anomaly;

[0149] Represents the joint probability of all grid observations. If the likelihood value is large, the current parameter If the overall plane data can be well explained and the likelihood value is small, the parameters need to be adjusted to be close to the actual measurement;

[0150] If the parameters are adjusted so that most Closer to the corresponding , then each exponential term increases, and the joint likelihood Increases, conversely, if the deviation increases, the likelihood decreases;

[0151] like As the noise increases, the noise is assumed to be larger and the distribution is wider, making the same deviation The penalty on likelihood is reduced if The smaller the value, the more severe the penalty, and the smaller the model fitting error needs to be to maintain a high likelihood;

[0152] Defining the prior distribution , then iteratively update the posterior:

[0153] ;

[0154] in, For Integrate over the domain of ;

[0155] Can be set based on historical grouting projects, geological drilling data or expert experience, such as grouting filling rate Obeying the Beta distribution, it tends to have a medium to high fill rate. Obey a uniform or biased distribution in a certain interval, It may be limited based on the actual cave scale experience; the prior density is large Value, which means it is more common and reasonable before there is no data, and the prior density is small Value, not very recognized by geology;

[0156] Posterior density , which means that after observing the entropy distribution, the parameter The updated belief of this formula is: likelihood × prior, which takes into account both data and experience. The denominator is normalized by a constant to ensure that the overall integral of the posterior density is 1.

[0157] Parameters with large posterior : It is consistent with the prior experience and can fit the current observation well. The parameters with small posterior either violate the prior experience or fit the data poorly. If The likelihood is improved and highly consistent with the prior, then the posterior is significantly increased. If there is a deviation from either the data or the prior, the posterior decreases;

[0158] Iterate several times until the posterior distribution converges, and then take the optimal model parameters Update the parameters of the cave network model; the iterative process continuously adjusts the parameters with observation data. The distribution of is estimated until the posterior distribution is stable;

[0159] Through this Bayesian update system, the entropy distribution can be organically combined with the geological model, parameters can be continuously corrected, and simulation accuracy can be improved, ultimately obtaining a more realistic cave network structure to guide subsequent construction warnings and parameter optimization.

[0160] Step 4: Issue a warning notification of the corresponding level based on the cave filling quality level, overlay the quality level information on the construction plane, and generate an interactive filling quality heat map;

[0161] Combining the entropy distribution map with the optimal model parameters obtained through Bayesian updating, the cave filling quality of each drilling area is graded and an early warning is triggered. At the same time, a heat map is generated on the construction plane. The specific logic is as follows:

[0162] For each grid o of each borehole, its comprehensive rheological entropy value and optimal fill rate , implement four-level assessment:

[0163] Excellent: and , good level: and ,intermediate: and , difference: and ;

[0164] The grades of excellent, good, medium and poor reflect the degree of completeness of grouting filling of the current caves in the area and the stability of grouting behavior; is the comprehensive rheological entropy value of the oth grid, which quantifies the disorder of the slurry diffusion at this location, The larger it is, the more chaotic and unstable the diffusion process will be. The smaller it is, the smoother and more orderly the diffusion; is the grouting filling rate prediction obtained after Bayesian update of the oth grid, and its value is , The larger the volume, the more the cavities or pores are filled with slurry. The smaller it is, the less filling there is, and there are cavities or channels;

[0165] Excellent: and Indicates very low disorder (smooth diffusion) and very high filling rate, and the construction quality is the best; good grade: and The indication is slightly to moderately disordered, with possible slight surging, but the fill rate is good and the overall quality is good;

[0166] intermediate: and Indicates high disorder (turbulence or channel instability), medium filling rate, need attention but not extreme; Poor: and The indication is that the diffusion is very chaotic and the filling rate is low, which may indicate serious leakage or voids and requires urgent treatment;

[0167] Multi-level warning trigger: Level 1 warning: any single grid within 5 minutes ; Indicates that the strong disorder at a single point persists, indicating that the location may be experiencing ongoing blockage, circulation breakout, or local leakage;

[0168] Second level warning: define adjacent grids to form a connected area ,and , average rheological entropy ; This indicates that the average rheological entropy value of the contiguous area exceeds the threshold value. is the number of grids, requiring at least 4 adjacent points to have common anomalies, and the regional average is greater than This indicates a large-scale disorder of moderate or higher levels. The significance lies in the fact that these are not isolated points, but rather contiguous areas of medium to high disorder, indicating that the diffusion anomaly is spatially diffuse and requires intervention on a larger scale.

[0169] Level 3 warning: Fill rate after Bayesian update And the corresponding posterior probability , that is, judging with high confidence that the fill rate is too low; It represents the confidence level that the filling rate is lower than 0.4 under the observed entropy value. This indicates that the predicted filling effect is extremely poor, and the posterior confidence level is greater than 90%, indicating a high degree of confidence in this judgment. This is the highest level of warning, indicating that there is a high probability of serious voids or construction failure in this area.

[0170] On the construction plane, the quality level of each grid is superimposed with different color levels, and rendered in excellent, good, medium and poor colors. When the warning is triggered, the corresponding grid is automatically highlighted and a notification is pushed. That is, when the warning conditions are met, the relevant grid flashes or a pop-up window is enlarged to remind the operator to respond immediately. Click a single grid to view the corresponding 、 , continuous monitoring curves and posterior confidence, dragging or zooming can focus on different areas, and support rapid positioning of risk points.

[0171] See also Figure 6 The present invention further provides a real-time monitoring system for the construction quality of jet grouting piles, wherein the system is used to implement the above-mentioned real-time monitoring method for the construction quality of jet grouting piles, comprising:

[0172] The data acquisition module is used to collect the grouting pressure, grouting flow rate and microseismic signals of each drilling point during the grouting process, form a multi-dimensional time series signal, and pre-process the multi-dimensional time series signal, including adaptive filtering and drift correction;

[0173] The entropy calculation module is used to generate a phase space vector sequence through delayed embedding based on the Takens phase space reconstruction principle for the preprocessed multi-dimensional time series signal, revealing the dynamic trajectory of the slurry flow process. Based on this trajectory, the distance distribution of adjacent vectors in the trajectory is calculated, and the rheological entropy value representing the degree of disordered diffusion of the slurry is output. Combined with the borehole spatial coordinates and the construction timestamp, the spatiotemporal field of the slurry diffusion behavior is constructed;

[0174] The model evaluation module is used to project the spatiotemporal field of slurry diffusion behavior onto the construction plane to generate an entropy distribution map. Based on the entropy distribution map, the underground cave network model is iteratively modified through a Bayesian update mechanism. Multi-level entropy thresholds are set to trigger a three-level warning. Based on the entropy distribution map and the model fitting results, the cave filling quality grade of each drilling area is output;

[0175] The early warning judgment module is used to issue early warning notifications of corresponding levels based on the cave filling quality level, superimpose quality level information on the construction plane, and generate an interactive filling quality heat map.

[0176] The present invention also provides a non-volatile computer-readable storage medium containing computer-executable instructions. When the computer-executable instructions are executed by one or more processors, the processors execute the method for real-time monitoring of the construction quality of rotary jet piles.

[0177] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0178] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software depends on the specific application and design constraints of the technical solution.

[0179] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, and may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment as needed.

[0180] The above is only a specific implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be covered by the scope of protection of the present application.

Claims

1. A method for real-time monitoring of the construction quality of jet grouting piles, characterized in that: The specific steps include: Step 1: Collect the grouting pressure, grouting flow rate and microseismic signals of each drilling point during the grouting process to form a multi-dimensional time series signal, and pre-process the multi-dimensional time series signal, which includes adaptive filtering and drift correction; Step 2: Based on the Takens phase space reconstruction principle, the preprocessed multidimensional time series signal is subjected to delayed embedding to generate a phase space vector sequence, revealing the dynamic trajectory of the slurry flow process. Based on this trajectory, the distance distribution of adjacent vectors in the trajectory is calculated, and the rheological entropy value representing the degree of disordered diffusion of the slurry is output. Combined with the borehole spatial coordinates and the construction timestamp, the spatiotemporal field of the slurry diffusion behavior is constructed; Step 3: Project the spatiotemporal field of slurry diffusion behavior onto the construction plane to generate an entropy distribution map. Based on the entropy distribution map, the underground cave network model is iteratively modified through the Bayesian update mechanism. Multi-level entropy thresholds are set to trigger a three-level warning. Based on the entropy distribution map and the model fitting results, the cave filling quality grade of each drilling area is output; Step 4: Issue a warning notification of the corresponding level based on the cave filling quality level, overlay the quality level information on the construction plane, and generate an interactive filling quality heat map.

2. A method for real-time monitoring of the construction quality of jet grouting piles according to claim 1, characterized in that: The logic of adaptive filtering and drift correction for multi-dimensional time series signals is: Using three times the length of the grouting pump's working cycle as a sliding window, the multidimensional time series signal is divided into data segments of equal step length according to the sliding window length. For each data segment, the following operations are performed: 1) Intrinsic mode functions are extracted sequentially through empirical mode decomposition, and the first three modes containing mechanical vibrations of the drilling rig and pipeline are removed to obtain the residual signal after vibration removal; 2) For the residual grouting pressure signal, a polynomial curve of no higher than fifth order is used to fit its baseline drift. This baseline curve is subtracted from the original pressure signal to obtain a drift-corrected pressure signal. 3) For the residual grouting flow signal, a two-way zero-phase digital low-pass filter with a cutoff frequency of 10 Hz is used to process it to obtain the drift-corrected grouting flow signal to retain the high-frequency components reflecting the grouting mutation; 4) For the residual microseismic signal, retain the original vibration mode, design a 1-200 Hz bandpass filter to suppress power frequency interference and high-frequency noise, and output the filtered microseismic signal.

3. A method for real-time monitoring of the construction quality of jet grouting piles according to claim 2, characterized in that: The logic of takingns phase space reconstruction of the drift-corrected pressure signal is: The drift-corrected grouting pressure signal is used as the main channel. Its autocorrelation function is calculated and the moment corresponding to the first zero crossing point is taken as the time delay of phase space reconstruction. Based on this time delay, a pseudo nearest neighbor discrimination algorithm is used to incrementally test different phase space reconstruction dimensions. The proportion of false nearest neighbors is monitored in each trial. When this proportion drops below 5%, the final phase space reconstruction dimension is determined. Based on the determined time delay and reconstruction dimension, the delay embedding method is used to generate phase space vector sequences for the drift-corrected grouting pressure signal, grouting flow signal, and filtered microseismic signal. The Euclidean distance between two adjacent points in the phase space vector sequence of each signal is calculated in turn to obtain the distance sequence of the pressure signal, grouting flow, and microseismic signal in the phase space. For each phase space, the distance interval width is set , and combined with the maximum value of the Euclidean distance in the phase space to divide K distance intervals of equal width, the Euclidean distance values ​​between each adjacent vector pair are divided into corresponding distance intervals, and the number of distance occurrences in each distance interval is counted, thereby obtaining a distance distribution histogram.

4. A method for real-time monitoring of the construction quality of jet grouting piles according to claim 3, characterized in that: The specific logic for calculating the rheological entropy value based on the distance distribution obtained by phase space reconstruction is as follows: Grouting pressure , grouting flow and microseismic signals The phase space distance distribution of is performed separately: Set the length of the phase space vector sequence of the xth type to , then the total number of adjacent vector pairs is , count the logarithm of the Euclidean distance of the xth signal falling into the kth distance interval , calculate the probability corresponding to the distance interval , , and satisfy the normalization condition ; Calculate the rheological entropy value that characterizes the degree of disorder in distance distribution for: ； in, Is the index of the signal type, 1, 2, 3 correspond to , is the index of the distance interval, is the total number of distance intervals, which is determined by the maximum Euclidean distance of the x-th signal The width of the distance interval OK, that is ; Generate a comprehensive rheological entropy value representing the global diffusion disorder through weighted fusion : , are weight coefficients, and ,when When , it is determined that the diffusion behavior of the slurry in the phase space is abnormal. .

5. A method for real-time monitoring of the construction quality of jet grouting piles according to claim 4, characterized in that: The logic of constructing the space-time field of slurry diffusion behavior includes: For each drilling point i, its spatial coordinates are known and its comprehensive rheological entropy , and form an observation set of the spatiotemporal field of slurry diffusion behavior ; Any point to be interpolated in the construction area is recorded as ,from Select n observation points with the closest time and space distance to form the observation point set , j is the index of the observation point in the observation point set, n is the number of observation points in the observation point set, where ; Define the weighted space-time distance between the jth observation point and the point to be interpolated: is the space-time distance between the point to be interpolated and the j-th observation point, ; Based on time and space distance Calculate the interpolation weights: , is the spatial correlation coefficient; The comprehensive rheological entropy value of the interpolation point at the time-space coordinate point is obtained by weighted average interpolation: ； in, is the comprehensive rheological entropy value of the interpolation point, is the comprehensive rheological entropy value at the jth observation point.

6. A method for real-time monitoring of the construction quality of jet grouting piles according to claim 5, characterized in that: Select the drilling point and interpolation point at the current moment in the construction plane, and use their comprehensive rheological entropy values Projecting onto a two-dimensional plane grid and generating an entropy distribution map, the grid resolution is aligned with the cave model; The Bayesian updating mechanism is used to modify the underground cave network model, including: The state parameter vector of the underground cave network is defined as ,in is the grouting filling rate, is the cave connectivity index, is the equivalent diameter of the cave; Divide the entropy distribution map into M equal-area grids, and the index of the grid , the comprehensive rheological entropy of the oth grid center is , the maximum possible entropy value is recorded as ; Assume that the observation noise follows a Gaussian distribution with mean 0 and variance Gaussian distribution, then calculate the likelihood function: ； in, Based on the current Theoretical entropy value prediction at the oth grid under the parameter; Defining the prior distribution , then iteratively update the posterior: ； in, For Integrate over the domain of ; Iterate several times until the posterior distribution converges, and then take the optimal model parameters Update cave network model parameters.

7. A method for real-time monitoring of the construction quality of jet grouting piles according to claim 6, characterized in that: Combining the entropy distribution map with the optimal model parameters obtained through Bayesian updating, the cave filling quality of each drilling area is graded and an early warning is triggered. At the same time, a heat map is generated on the construction plane. The specific logic is as follows: For each grid o of each borehole, the comprehensive rheological entropy value and optimal fill rate , implement four-level assessment: Excellent: and , good level: and ,intermediate: and , difference: and ; Multi-level warning trigger: Level 1 warning: any single grid within 5 minutes ; Second level warning: define adjacent grids to form a connected area ,and , average rheological entropy ; Level 3 Warning: Fill Rate After Bayesian Update And the corresponding posterior probability ; On the construction plane, the grid is used as a unit, and the quality level of each grid is superimposed with different color levels, and rendered with excellent, good, medium and poor colors. When the early warning is triggered, the corresponding grid is automatically highlighted and a notification is pushed.

8. A real-time monitoring system for the construction quality of jet grouting piles, characterized by: The system is used to implement the method for real-time monitoring of the construction quality of jet grouting piles according to any one of claims 1 to 7, comprising: The data acquisition module is used to collect the grouting pressure, grouting flow rate and microseismic signals of each drilling point during the grouting process, form a multi-dimensional time series signal, and pre-process the multi-dimensional time series signal, including adaptive filtering and drift correction; The entropy calculation module is used to generate a phase space vector sequence through delayed embedding based on the Takens phase space reconstruction principle for the preprocessed multi-dimensional time series signal, revealing the dynamic trajectory of the slurry flow process. Based on this trajectory, the distance distribution of adjacent vectors in the trajectory is calculated, and the rheological entropy value representing the degree of disordered diffusion of the slurry is output. Combined with the borehole spatial coordinates and the construction timestamp, the spatiotemporal field of the slurry diffusion behavior is constructed; The model evaluation module is used to project the spatiotemporal field of slurry diffusion behavior onto the construction plane to generate an entropy distribution map. Based on the entropy distribution map, the underground cave network model is iteratively modified through a Bayesian update mechanism. Multi-level entropy thresholds are set to trigger a three-level warning. Based on the entropy distribution map and the model fitting results, the cave filling quality grade of each drilling area is output; The early warning judgment module is used to issue early warning notifications of corresponding levels based on the cave filling quality level, superimpose quality level information on the construction plane, and generate an interactive filling quality heat map.

9. A non-volatile computer-readable storage medium containing computer-executable instructions, characterized in that: When the computer-executable instructions are executed by one or more processors, the processors are enabled to execute the method for real-time monitoring of the construction quality of jet grouting piles according to any one of claims 1 to 7.

Images