Pressure fusion-based thin-walled cavity site adaptive support monitoring method and system

Abstract

The application provides a thin-wall cavity site adaptive support monitoring method and system based on pressure fusion. The method comprises: obtaining the spatial form and wall state of the thin-wall cavity site, and dividing the site wall into multiple monitoring units; performing limited trial pressure regulation on each support chamber to determine the support action relationship between the support chamber and the monitoring unit; performing space-time correction and reliability evaluation on multi-source pressure data; generating a wall pressure state field, determining the two-way differential pressure safety boundary of each monitoring unit according to the wall state, comparing the wall pressure state field with the two-way differential pressure safety boundary, and determining the risk monitoring unit; generating a target support pressure field, converting the target support pressure field into a target pressure of each support chamber; adjusting each support chamber according to the target pressure, and updating the wall pressure state field according to the adjusted multi-source pressure data until the backfilling operation is completed. The application can improve the stress sensing accuracy of the wall, improve the partition support adaptability, and enhance the safety of the backfilling process.

Description

Technical Field

[0001] This application relates to the field of cultural relic protection and structural safety monitoring technology, and in particular to an adaptive support monitoring method and system for thin-walled cavity sites based on pressure fusion. Background Technology

[0002] Thin-walled hollow sites are widely found in immovable cultural relics such as ancient pottery kilns, casting kilns, and underground flues. These sites are typically constructed of adobe, brick, or other fragile materials, characterized by thin walls, low strength, irregular shapes, and narrow internal spaces. After excavation, protective backfilling is usually required to mitigate the impact of environmental changes and external disturbances on the site. During backfilling, the lateral pressure and localized construction loads generated by the backfill soil continuously act on the site walls. Without effective support, this can easily lead to wall deformation, crack expansion, or even cavity collapse.

[0003] Existing technologies mainly employ grouting reinforcement, rigid formwork support, or flexible material padding. Grouting reinforcement increases the load-bearing capacity by injecting solidifying materials into the wall or surrounding soil, but it is irreversible and may lead to material penetration and chemical erosion. Rigid formwork or supports can provide high support strength, but they are difficult to adapt to cavity cross-sections and wall undulations, easily leading to localized contact and stress concentration. Furthermore, the installation, fixing, and dismantling processes may damage the site itself. Flexible padding causes less disturbance to the wall, but its support capacity is limited and it is difficult to resist dynamically changing backfill pressure. In addition, existing support schemes usually only control the pressure inside the support device or a preset threshold, failing to accurately obtain the actual contact pressure of the wall and the external backfill pressure. They also cannot implement zonal adjustments based on the load-bearing differences of different wall sections, resulting in both insufficient and excessive support, making it difficult to ensure continuous safety during the backfilling process. Summary of the Invention

[0004] In view of this, the present application provides an adaptive support monitoring method and system for thin-walled cavity archaeological sites based on pressure fusion, in order to solve the problems of the existing technology, such as the difficulty in accurately sensing the stress state of the wall, the difficulty in reasonably determining the zoned support pressure, and the lack of dynamic adjustment in the backfilling process.

[0005] A first aspect of this application provides an adaptive support monitoring method for thin-walled cavity archaeological sites based on pressure fusion, comprising: acquiring the spatial morphology and wall state of the thin-walled cavity archaeological site; dividing the archaeological wall into multiple monitoring units; and establishing the association between the support chambers, pressure sensing units, and monitoring units of a partitioned flexible support system; performing restricted trial pressure regulation on each support chamber; determining the support relationship between the support chamber and the monitoring unit based on chamber pressure changes and contact pressure responses; acquiring chamber pressure, wall contact pressure, and external backfill pressure during the backfilling process; and performing spatiotemporal correction and reliability assessment on the multi-source pressure data; Multi-source pressure data is integrated based on correlation, support action, and load transfer relationships between monitoring units to generate a wall pressure state field. The bidirectional differential pressure safety boundary of each monitoring unit is determined according to the wall state. The wall pressure state field is compared with the bidirectional differential pressure safety boundary to identify risk monitoring units. A target support pressure field is generated based on the current pressure state of the risk monitoring units and the predicted backfill load. The target support pressure field is then converted into the target pressure for each support chamber using the support action relationship. Each support chamber is adjusted according to the target pressure, and the wall pressure state field is updated based on the adjusted multi-source pressure data until the backfilling operation is completed.

[0006] A second aspect of this application provides an adaptive support monitoring system for thin-walled cavity archaeological sites based on pressure fusion, comprising: an acquisition module for acquiring the spatial morphology and wall state of the thin-walled cavity archaeological site, dividing the archaeological wall into multiple monitoring units, and establishing the association between the support chambers of the partitioned flexible support body, pressure sensing units, and monitoring units; a pressure regulation module for performing limited trial pressure regulation on each support chamber, and determining the support relationship between the support chamber and the monitoring unit based on chamber pressure changes and contact pressure responses; a correction module for acquiring chamber pressure, wall contact pressure, and external backfill pressure during the backfilling process, and performing spatiotemporal correction and reliability assessment on the multi-source pressure data; and a fusion module. The system is divided into four modules: a 1) a 2) a 3) a 4) a 5) a 6) a 7) 4) a 5) 4) 5)"" 4) 5)"" 5)"" 5) 6)"" 5)"" 5)""" 5"——1."—55"5——————" "— — "——"—"—"—"—"—"—"—"—"—"—"—"—"—"—d"—"—d"—"—"—,——,——— -—— "—— -———"—"—"—"—"—"—"—"—"—"—"—d"—"—d"—"—d"—"—d"—"—"—d"—"—,——,—— -——"—"—"—"—d"—"—d"—"—"—d"—"—d"—"—d"—"—d"—"—d"—"—d"—"—d"—"—d"—"—d"—"—d"—"—,——,—— -—— "—— "————— "——— "——— "——— "——— "——— "—— "— "51"—"—" "— "——"—"—"—"—"—"—"—"—"—"—"—"—"—d ...

[0007] The above-described technical solutions adopted in the embodiments of this application can achieve the following beneficial effects: By acquiring the spatial morphology and wall condition of the thin-walled cavity site, the site wall was divided into multiple monitoring units, and the correlation between the support chambers, pressure sensing units, and monitoring units of the zoned flexible support system was established. Constrained exploratory pressure regulation was performed on each support chamber, and the support relationship between the support chamber and the monitoring unit was determined based on chamber pressure changes and contact pressure responses. Chamber pressure, wall contact pressure, and external backfill pressure during the backfilling process were acquired, and spatiotemporal correction and reliability assessment were performed on the multi-source pressure data. Based on the correlation, support relationship, and the relationships between monitoring units... The load transfer relationship is integrated with multi-source pressure data to generate a wall pressure state field. Based on the wall state, a bidirectional differential pressure safety boundary is determined for each monitoring unit. The wall pressure state field is compared with the bidirectional differential pressure safety boundary to identify risk monitoring units. A target support pressure field is generated based on the current pressure state of the risk monitoring units and the predicted backfill load. The target support pressure field is then converted into target pressures for each support chamber using the support action relationship. Each support chamber is adjusted according to the target pressures, and the wall pressure state field is updated based on the adjusted multi-source pressure data until backfilling is completed. This application improves the accuracy of wall stress perception, enhances the adaptability of zoned support, and strengthens the safety of the backfilling process. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 This is a flowchart illustrating the adaptive support monitoring method for thin-walled cavity archaeological sites based on pressure fusion provided in this application embodiment; Figure 2 This is a schematic diagram of the structure of the adaptive support monitoring system for thin-walled cavity archaeological sites based on pressure fusion provided in the embodiments of this application; Figure 3 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0010] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0011] When implementing protective backfilling for existing thin-walled cavity archaeological sites, methods such as grouting reinforcement, rigid formwork support, rigid bracket support, or flexible material padding are commonly used to withstand the lateral pressure generated by the backfill soil. Grouting reinforcement can improve the overall bearing capacity of the archaeological wall or surrounding soil, but it is irreversible and may cause the reinforcement material to seep in or undergo chemical reactions. Rigid formwork and rigid brackets have high support strength, but they are difficult to fully conform to irregular walls, easily leading to localized stress concentrations. While flexible padding can reduce disturbance to the archaeological site during installation, its support and adjustment capabilities are limited. Some existing solutions also assess and control the support status by monitoring the internal pressure of the support device.

[0012] However, thin-walled cavity sites exhibit significant spatial variations in wall thickness, curvature, crack distribution, and material degradation. Backfill loads also dynamically change with backfill height, compaction, and construction location. Relying solely on the internal pressure of the support device is insufficient to reflect the actual contact pressure between the support structure and the wall, let alone accurately characterize the net differential pressure formed by the combined effects of external backfill pressure and internal support pressure. Therefore, existing technologies suffer from difficulties in accurately perceiving the wall's stress state, determining appropriate support pressure for different wall sections, and dynamically adjusting support pressure according to changes in backfill load, easily leading to insufficient or excessive local support.

[0013] To address the aforementioned issues, this application divides the thin-walled cavity site into multiple monitoring units and establishes spatial relationships between support chambers, pressure sensing units, and monitoring units. By performing constrained exploratory pressure regulation on each support chamber, the support relationship is determined based on chamber pressure changes and wall contact pressure responses. Chamber pressure, wall contact pressure, and external backfill pressure are collected during the backfilling process. The multi-source pressure data undergoes spatiotemporal correction, reliability assessment, and fusion processing to generate a wall pressure state field reflecting the spatial distribution of internal support pressure, external backfill pressure, and net differential pressure. Furthermore, a bidirectional differential pressure safety boundary is determined for each monitoring unit based on the wall state, risk monitoring units are identified, and a target support pressure field is generated by combining the current pressure state and subsequent backfill load prediction results. Finally, the target pressure of each support chamber is inverted using the support relationship, and closed-loop zonal pressure regulation is implemented.

[0014] Through the above technical solutions, this application can improve the accuracy of sensing the stress state of the thin-walled cavity site wall, making the support pressure compatible with the bearing state of different monitoring units and the changes in backfill load; at the same time, it can promptly identify the risks of insufficient and excessive support based on the bidirectional differential pressure safety boundary, and continuously correct the pressure of the support chamber through multi-source pressure feedback, thereby improving the adaptability and dynamic adjustment capability of zoned support, reducing the risk of local stress concentration, crack propagation and cavity collapse in the wall, and enhancing the safety of the backfilling process.

[0015] The technical solution of this application will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0016] Figure 1 This is a schematic flowchart of the adaptive support monitoring method for thin-walled cavity archaeological sites based on pressure fusion provided in this application embodiment. Figure 1 As shown, the method may specifically include: S101, obtain the spatial morphology and wall state of the thin-walled cavity site, divide the site wall into multiple monitoring units, and establish the correlation between the support chambers, pressure sensing units and monitoring units of the partitioned flexible support body. S102, Perform restricted exploratory pressure adjustment on each support chamber, and determine the support relationship between the support chamber and the monitoring unit based on the chamber pressure change and contact pressure response; S103, acquire chamber pressure, wall contact pressure and external backfill pressure during the backfilling process, and perform spatiotemporal correction and reliability assessment on the multi-source pressure data; S104, based on the correlation, support relationship and load transfer relationship between monitoring units, multi-source pressure data are fused to generate the wall pressure state field; S105, determine the bidirectional differential pressure safety boundary of each monitoring unit based on the wall condition, compare the wall pressure state field with the bidirectional differential pressure safety boundary, and determine the risk monitoring unit; S106, Generate the target support pressure field based on the current pressure status of the risk monitoring unit and the prediction results of the subsequent backfill load, and convert the target support pressure field into the target pressure of each support chamber using the support action relationship; S107, adjust each support chamber according to the target pressure, and update the wall pressure state field based on the adjusted multi-source pressure data until the backfilling operation is completed.

[0017] In some embodiments, the spatial morphology and wall state of the thin-walled cavity site are acquired, the site wall is divided into multiple monitoring units, and the relationship between the support chambers, pressure sensing units, and monitoring units of the zoned flexible support system is established, including: Spatial measurement data and wall detection data of the inner and outer surfaces of the thin-walled cavity site were collected to construct a spatial model of the site that characterizes the geometric shape and structural state of the wall surface. Based on the morphological continuity and state differences in the site's spatial model, the zoning boundaries are determined, and the site walls are divided into multiple monitoring units with relatively consistent stress characteristics. The partitioned flexible support structure is placed in the spatial reference corresponding to the site spatial model. Based on the effective range of each support chamber and the sensing range of each pressure sensing unit, its projection area on the site wall is determined. The association weights are established based on the degree of spatial overlap between the projection area and the monitoring unit, and the association relationship between the support chamber, the pressure sensing unit and the monitoring unit is generated.

[0018] Specifically, spatial measurement data is used to uniformly describe the three-dimensional positional relationships of the inner and outer surfaces and cavity boundaries of the thin-walled cavity site, while wall inspection data is used to reflect the material integrity and load-bearing differences of different wall sections. During data acquisition, point cloud data of the inner and outer surfaces of the site can be obtained using non-contact 3D scanning equipment, and multi-station splicing is used to supplement measurements in obscured areas; wall thickness, crack location, crack width, surface erosion degree, water content, and material strength characterization are acquired simultaneously. Coordinate registration, noise removal, and hole repair are performed on the multi-station point cloud data. A unified spatial coordinate system is established with the site reference point, principal axis direction, and vertical direction, and then the wall inspection data is mapped to the corresponding surface grid according to the inspection location. The wall thickness distribution is determined by the normal distance between the inner and outer surfaces, the curvature and curvature variation are determined by the local surface fitting results, and the structural state attributes are formed by the crack and deterioration detection results, thereby constructing a spatial model of the site that includes geometric grid, wall thickness attributes, curvature attributes, and damage attributes.

[0019] The monitoring units are not divided using a fixed-size grid, but rather consider both the continuity of the wall morphology and the differences in structural state. First, locations of morphological abrupt changes are identified based on the rate of change of curvature, the rate of change of wall thickness, and differences in the normal direction. Then, locations of state abrupt changes are identified based on crack boundaries, locations of abrupt changes in material strength, and obvious erosion areas. These two types of abrupt changes are used as candidate partition boundaries. For local areas enclosed by candidate boundaries, the dispersion of wall thickness, curvature, and material state within the region is calculated. When the dispersion is below a set threshold, it is retained as a monitoring unit; when the dispersion is high, it is further subdivided. For local areas that are too small and have similar states to adjacent areas, they can be merged into adjacent monitoring units to avoid excessive fragmentation of monitoring units. Each monitoring unit records its spatial extent, surface normal, average wall thickness, average curvature, damage level, and adjacency relationships, ensuring that the walls within the same monitoring unit have relatively consistent stress characteristics.

[0020] Further, after completing the division of monitoring units, the design model or post-installation measurement model of the zoned flexible support structure is registered to the site spatial model. Each support chamber has an independent chamber identifier, geometric boundary, and adjustable pressure range. The effective range is determined based on the outer contour of the chamber under different pressure states, the extension direction of the flexible membrane layer, and the constraints of adjacent chambers. For pressure sensing units installed on the surface of the flexible support structure, the outer side of the site wall, or in the backfill soil, the sensing range is determined based on the installation location, sensing surface direction, and effective sensing distance. Along the normal direction of the site wall or the shortest contact path, the effective range of the support chamber and the sensing range of the pressure sensing unit are projected onto the site wall to obtain the corresponding projection area.

[0021] The correlation weight can be determined based on the overlap ratio between the projected area and the monitoring unit, and corrected by considering the projection distance, normal angle, and occlusion status. A higher correlation weight is obtained when the overlap area is larger, the projection distance is smaller, and the direction of action is more consistent with the wall normal. The correlation weight is reduced when there are structural protrusions, chamber diaphragms, or installation components obstructing the view. The weights of monitoring units associated with the same support chamber are normalized to form the action correlation from chamber to monitoring unit. The weights of monitoring units covered by the same pressure sensing unit are also normalized to form the observation correlation from sensing unit to monitoring unit. The correlation results can be stored in a correlation matrix, where the matrix rows correspond to support chambers or pressure sensing units, the matrix columns correspond to monitoring units, and the matrix elements are the correlation weights.

[0022] To avoid correlation distortion caused by installation deviations of the flexible support structure, the surface profile of the support structure can be acquired under the initial low-pressure fit condition. The measured profile is then re-registered with the design model, and the projected area is updated based on the displacement of the chamber boundary. If a pressure sensing unit simultaneously covers multiple monitoring units with significantly different states, the observation contribution is split according to the effective sensing area. If a monitoring unit is not effectively covered by any sensing range, the position of the sensing unit is adjusted or the monitoring unit is marked as a low observation integrity area. Through the above processing, a calculable correspondence can be established between the geometric state of the site, the support action area, and the pressure observation area under the same spatial reference, reducing the impact of irregular walls, local damage, and installation deviations on subsequent pressure analysis.

[0023] For example, when processing a near-elliptical thin-walled kiln site, 3D scanning revealed a cavity length of approximately 3.2m, a maximum width of approximately 1.8m, and a wall thickness distribution ranging from 0.12m to 0.28m. The model identified a high-curvature thin-walled area at the top, a cracked area in the side walls, a thick-walled area at the bottom, and a transition zone, dividing the area into 18 monitoring units. The flexible support structure consisted of 6 independent support chambers and 24 contact pressure sensing units. The overlap ratios of the projections of the second support chamber onto the top 4 monitoring units were 0.52, 0.31, 0.12, and 0.05, respectively. After correction for projection distance and normal angle, the corresponding correlation weights were determined to be 0.58, 0.29, 0.09, and 0.04. This clearly defines the main wall area affected by each support chamber and the main monitoring area reflected by each pressure sensing unit, providing a unified spatial basis for subsequent trial pressure adjustment, pressure fusion, and zonal support control.

[0024] In some embodiments, restricted exploratory pressure regulation is performed on each support chamber to determine the support relationship between the support chamber and the monitoring unit based on chamber pressure changes and contact pressure responses, including: Within the pressure regulation range below the corresponding monitoring unit's safe pressure, graded pressure disturbances are applied sequentially to each support chamber, and the chamber pressure and contact pressure corresponding to each pressure level are collected. Based on the increase and spatial distribution of contact pressure before and after the pressure disturbance, determine the degree of pressure response of each support chamber to each monitoring unit; Identify the empty travel between the change in chamber pressure and the establishment of contact pressure, and determine the contact starting point and pressure transmission efficiency based on the relationship between contact pressure and chamber pressure. Based on the degree of pressure response, the starting point of contact, and the pressure transmission efficiency, a support effect relationship is generated that characterizes the weight of each support chamber on each monitoring unit.

[0025] Specifically, the limited trial pressure adjustment is used to identify the actual action path of each support chamber on different monitoring units before formal backfilling, avoiding the estimation of the action range based solely on the design location of the support structure. Before execution, the allowable trial pressure is determined based on the wall thickness, curvature, crack condition, material strength, and initial contact condition of each monitoring unit, and the smaller positive safety margin within the bidirectional differential pressure safety boundary of each monitoring unit is taken as the upper limit of the pressure adjustment.

[0026] To reduce the coupling effect between adjacent chambers, the non-test chambers can be kept at the reference pressure, while only the current test chamber is subjected to gradually increasing low-amplitude pressure disturbances. Each pressure level is maintained until the chamber pressure and contact pressure changes enter a stable range before moving to the next pressure level. The pressure levels can be set according to constant or adaptive amplitude. If the contact pressure increases too rapidly, the local pressure gradient exceeds the limit, or the wall condition becomes abnormal, the current test is immediately terminated and the pressure returns to the previous safe pressure.

[0027] During data acquisition, chamber pressure, output of each contact pressure sensing unit, and status of the corresponding monitoring unit are recorded using a unified timestamp. For each level of pressure disturbance, the increment between the baseline pressure before the disturbance and the contact pressure after stabilization is calculated, and the contact pressure increment is mapped to each monitoring unit according to the association weight between the pressure sensing unit and the monitoring unit. For the same support chamber, if the contact pressure increment of a certain monitoring unit continuously increases with the increase of chamber pressure, and the response timing is consistent with the pressure disturbance, then the monitoring unit is determined to be an effective response area; if only instantaneous fluctuations or inconsistent change directions occur, the corresponding response effectiveness is reduced. Furthermore, pressure response levels are generated based on contact pressure increment, response duration, spatial coverage ratio, and the degree of response attenuation of adjacent monitoring units to distinguish between primary action areas, secondary action areas, and weakly coupled areas.

[0028] The empty travel is used to characterize the ineffective pressure change stage of the support chamber from the start of pressurization to before the flexible membrane layer forms effective contact with the site wall. The chamber pressure sequence and the contact pressure sequence can be time-aligned to determine the pressure range where the chamber pressure has been continuously rising while the contact pressure remains within the noise band. When the contact pressure exceeds the baseline fluctuation limit for multiple consecutive sampling periods and changes in the same direction as the chamber pressure, the corresponding chamber pressure is determined as the contact start point. To avoid misjudgment due to single-point noise, the number of pressure sensing units achieving effective response within the same monitoring unit can be required to meet coverage conditions, or adjacent sensing units can form a continuous response area. A higher contact start point indicates a more significant gap, wrinkle, or localized suspension between the chamber and the wall.

[0029] Furthermore, after the initial bonding point, the pressure transmission efficiency is determined based on the relationship between the increase in contact pressure and the increase in chamber pressure. The increment ratio can be calculated within each effective pressure level and weighted averaged in conjunction with response stability. When there are clear segmentations in the relationship, the transmission efficiency is determined separately for the initial bonding stage and the stable bearing stage. If the chamber pressure continues to increase but the increase in contact pressure slows down, it can be determined that the flexible membrane layer has undergone lateral expansion or a shift in its effective range. Based on this, the stable transmission efficiency of the current monitoring unit is reduced, and the response weights of adjacent monitoring units are updated. Finally, the pressure response level, the bonding point correction coefficient, and the pressure transmission efficiency are normalized and combined to form the action weights from the support chamber to the monitoring unit. These weights are then normalized according to the total action weights of each support chamber to generate a support action relationship matrix.

[0030] To improve the repeatability of the identification results, at least two rounds of pressurization and pressure return tests can be performed on each support chamber to compare the contact pressure response at the same pressure level. When the deviation between the two rounds of response is less than the consistency threshold, the weighted average result is used; when the deviation is large, chamber leakage, membrane slippage, and sensing unit drift are checked, and the test is repeated after eliminating the anomalies. If the bonding starting point decreases significantly with each round, it can be determined that the flexible membrane has flattened after the first round of pressurization, and the support action relationship is updated with the results of stable rounds.

[0031] For example, when conducting test pressure regulation on a thin-walled kiln site with six support chambers, the upper limit of the allowable test pressure for the second support chamber was determined to be 18 kPa, with four pressure levels set at 6 kPa, 9 kPa, 12 kPa, and 15 kPa. During the test, the contact pressure of the top monitoring units A, B, and C increased from 0.4 kPa to 4.8 kPa, 2.9 kPa, and 1.1 kPa, respectively, while the side wall monitoring unit D only increased by 0.2 kPa. Analysis showed that A, B, and C were the effective response regions, while D was a weak coupling region. The contact pressure continuously exceeded the baseline fluctuation limit after the chamber pressure reached 7.2 kPa, therefore 7.2 kPa was determined as the contact starting point; the pressure transmission efficiencies of A, B, and C during the stable bearing stage were 0.46, 0.27, and 0.09, respectively. After corrections based on response level and contact start point, the influence weights of the second support chamber on A, B, C, and D were determined to be 0.58, 0.28, 0.11, and 0.03, respectively. This yields a support action relationship consistent with the actual contact state, providing a reliable basis for subsequent target support pressure field inversion and zoned closed-loop pressure regulation.

[0032] In some embodiments, the chamber pressure, wall contact pressure, and external backfill pressure during the backfilling process are acquired, and spatiotemporal correction and reliability assessment are performed on the multi-source pressure data, including: Pressure data for each support chamber and its corresponding monitoring unit were collected according to a unified sampling sequence, and spatial mapping was completed based on the correlation. The pressure data is corrected for measurement deviations based on the baseline state of the pressure sensing unit and environmental influences, and timing alignment is performed according to the pressure response delay. Based on the spatial consistency, continuous variation characteristics and support relationship of adjacent pressure data, the observation residuals of each pressure data are determined. Data reliability is generated based on the observation residuals and the operational stability of the pressure sensing unit, and abnormal pressure data is deweighted and labeled to obtain corrected multi-source pressure data.

[0033] Specifically, the multi-source pressure data includes the internal pressure of each support chamber, the contact pressure between the flexible support and the site wall, and the external backfill pressure located on the outside of the wall or in the backfill soil. To ensure that different pressure data can be compared at the same time, a unified acquisition controller sends synchronous sampling trigger signals to each pressure sensing unit, and a common clock is used to record the sampling time. Chamber pressure can be directly associated with the corresponding chamber according to the support chamber identifier, while contact pressure and external backfill pressure are mapped to the monitoring unit according to the installation coordinates, sensing direction, and effective sensing range. When one pressure sensing unit covers multiple monitoring units, the observation contribution is allocated according to the spatial overlap ratio; when one monitoring unit corresponds to multiple pressure sensing units, each observation value and its initial weight are retained to avoid directly averaging before correction and thus masking local anomalies.

[0034] After installation, the pressure sensing unit can collect static output under conditions of no backfill load and with the support chamber at reference pressure, forming zero-point reference, noise amplitude, and short-term drift parameters. During backfilling operations, temperature, humidity, power supply fluctuations, and the pre-tightening state of the flexible membrane layer may cause measurement deviations. Therefore, environmental conditions can be collected simultaneously, and zero-point drift and sensitivity changes can be corrected using pre-calibrated compensation relationships. For contact pressure sensing units, force direction deviations can be corrected based on changes in installation curvature and contact area; for external backfill pressure sensing units, local measurement deviations can be corrected based on burial depth, attitude offset, and soil compaction state.

[0035] For communication delays or sample loss, data can be reordered according to timestamps, and interpolation using adjacent valid samples can be used within the allowable missing duration; if the allowable duration is exceeded, observations are not recreated, but the corresponding time period is marked as a data gap. To avoid transient pulses caused by pump and valve actions being mistaken for sudden changes in wall load, a short-time filtering window can be set according to the duration of the pressure regulation command, but true pressure transitions exceeding the safe rate of change are retained.

[0036] Furthermore, after amplitude correction, the time difference for different pressure sequences to reach an effective response is calculated, using the pressure regulation action, backfilling propulsion event, or pressure change moment as a reference. Data with slower responses are then aligned to a unified state moment. Timing alignment can use a fixed response delay or dynamically update the delay parameters based on recent pressure changes.

[0037] The observation residuals are used to determine whether the pressure data are consistent with the surrounding conditions and the support action patterns. For each monitoring unit, a spatial reference pressure can be generated based on the distance to adjacent monitoring units, wall continuity, and load transfer direction; a temporal reference pressure can be generated based on the pressure value of the previous sampling period, the current rate of change, and the backfilling operation status. For contact pressure, the contact pressure change of the corresponding monitoring unit can be predicted based on the chamber pressure change using the support action relationship. The actual observed values ​​are compared with the spatial reference pressure, temporal reference pressure, and action relationship prediction values ​​to form spatial residuals, continuous change residuals, and action response residuals. If an observed value differs significantly from adjacent data, but the corresponding location is at a crack boundary, curvature abrupt change zone, or backfill interface, the true local gradient is retained and not directly judged as an anomaly; the observation residuals are only increased when the difference cannot be explained by the wall condition and backfilling process.

[0038] Data reliability can be determined by observation residuals, signal noise, data missing rate, zero-point drift, and continuous operational stability. Each indicator is first oriented and normalized within its intervals, then weighted according to the pressure sensing unit type to generate a reliability score between 0 and 1. Data with higher reliability is used as the primary observation for subsequent pressure fusion; data with reliability in the middle range has its fusion weight reduced and is marked for review; data with reliability below the anomaly threshold retains its original record but is not directly used for updating the wall pressure state field. If the reliability of the same pressure sensing unit is low for multiple consecutive periods, a sensor status check is triggered, and alternative estimates are generated using adjacent sensing units and support relationships.

[0039] For example, the top monitoring unit of a thin-walled kiln site was covered by two contact pressure sensing units and one external backfill pressure sensing unit, with a uniform sampling period of 1 second. During backfilling, the chamber pressure increased from 10 kPa to 12 kPa, and the contact pressure should normally increase after about 2 seconds. The acquisition results showed that the first contact pressure sensing unit increased by 0.9 kPa, and the second contact pressure sensing unit increased by 0.1 kPa, with the output of the second sensing unit accompanied by high-frequency fluctuations. After temperature compensation and time alignment, the residual between the second sensing unit and the spatial reference pressure was still 0.72, and the residual with the predicted value of the interaction relationship was 0.68, with its confidence level dropping to 0.24. Therefore, it was marked as an anomaly and downweighted; the confidence level of the first sensing unit was 0.91, and it continued to be the primary observation. Through the above processing, the interference of drift, delay, and fault data on pressure fusion can be suppressed while preserving the true local pressure changes, improving the stability and accuracy of the reconstructed stress state of the wall.

[0040] In some embodiments, multi-source pressure data are fused based on correlation relationships, support action relationships, and load transfer relationships between monitoring units to generate a wall pressure state field, including: Based on the correlation, the corrected multi-source pressure data are mapped to the corresponding monitoring units to construct pressure state nodes that characterize external loads, internal support, and observation uncertainty. Based on the relationship of support action, the chamber pressure and wall contact pressure are constrained, and the estimated value of the internal support pressure of each monitoring unit is determined in combination with the data reliability. Spatial propagation constraints are established based on the load transfer relationship between monitoring units to collaboratively estimate external backfill pressure and retain the pressure gradient at the location of structural state abrupt change. The net differential pressure and its variation characteristics are determined based on the estimated internal support pressure and the estimated external backfill pressure, and the pressure state field of the wall is generated.

[0041] Specifically, multi-source pressure fusion uses monitoring units as a unified calculation object, converting pressure observations from different sources, with varying coverage and reliability into wall pressure states with spatial location and physical meaning. First, based on the correlation between support chambers, pressure sensing units, and monitoring units, the corrected chamber pressure, contact pressure, and external backfill pressure are allocated to the corresponding monitoring units. For cases where a single sensing unit covers multiple monitoring units, the observation contribution is split according to spatial overlap weights; for cases where multiple sensing units cover the same monitoring unit, each observation value, data reliability, and sensing direction are retained, without direct equal-weighted averaging. Thus, a pressure state node is established for each monitoring unit. This node records candidate values ​​for external loads, candidate values ​​for internal support, observation time, data reliability, missing data markers, and estimation uncertainties. Connections between pressure state nodes are established according to the adjacency relationships between monitoring units.

[0042] Internal support pressure is not directly equivalent to the pressure inside the support chamber. Instead, it is the actual pressure exerted on the wall after the chamber pressure has passed through the flexible membrane layer, pressure transmission, and spatial diffusion. Based on the action weights, contact points, and pressure transmission efficiency in the support action relationship, the effective pressure exceeding the contact point in each support chamber can be converted into the predicted support pressure for the corresponding monitoring unit. For monitoring units simultaneously affected by multiple support chambers, the predicted pressures of each chamber are superimposed according to their action weights, and unreasonable superposition is limited by a coupling correction term between adjacent chambers. Subsequently, the predicted support pressure is compared with the observed contact pressure, and their contributions are dynamically adjusted based on the reliability of the contact pressure data. When the reliability of the contact pressure is high, the estimated result is closer to the measured contact pressure; when the contact pressure is missing or has low reliability, estimation is mainly based on the chamber pressure and support action relationship, while simultaneously expanding the uncertainty range. If the predicted value continuously deviates from the measured value, the action residual of the corresponding node is updated to identify membrane slippage, local suspension, or chamber leakage.

[0043] The collaborative estimation of external backfill pressure is based on the load transfer relationship between monitoring units. This relationship can be determined by factors such as wall adjacency distance, curvature continuity, wall thickness difference, crack isolation state, and backfill contact range. For adjacent monitoring units with continuous wall thickness and curvature and no through cracks, a strong propagation weight is set to ensure smooth propagation of existing external pressure observations to areas without sensing units. For crack boundaries, abrupt wall thickness changes, corner areas, or backfill interfaces, the propagation weight is reduced to prevent over-smoothing of the actual pressure gradient. During collaborative estimation, the measured external backfill pressure value is used as the observation constraint, and the continuity of pressure changes at adjacent nodes is used as the propagation constraint. Prior load values ​​are formed by combining backfill height, soil compaction state, and construction location. For nodes lacking direct observation, the load can be determined by weighting the propagation results from neighboring nodes with the prior load values. For nodes with highly reliable measured values, the deviation of the propagation results from the measured values ​​is limited.

[0044] Furthermore, after estimating the internal support pressure and external backfill pressure, the net differential pressure of each monitoring unit is calculated according to a unified pressure direction. The effect of external backfill pressure pointing into the cavity can be defined as positive, and the effect of internal support pressure pointing outward of the wall can be defined as negative; the difference between the two represents the net load borne by the wall. The rate of change, acceleration, and duration of change are calculated based on the net differential pressure within a continuous sampling period, and the spatial pressure gradient is calculated based on the difference in net differential pressure between adjacent monitoring units. The pressure state field can be represented using a graph structure with time attributes. Nodes store external backfill pressure, internal support pressure, net differential pressure, and uncertainty; connections store the load propagation direction and pressure gradient, enabling subsequent risk identification to simultaneously consider local pressure magnitude, change trend, and spatial diffusion state.

[0045] For example, monitoring unit A at the top of a thin-walled kiln is covered by a second support chamber and two contact pressure sensing units. The current pressure in the second support chamber is 14 kPa, the contact point is 7.2 kPa, the transmission efficiency is 0.46, and the action weight is 0.58, resulting in a predicted support pressure of approximately 1.81 kPa. The two contact pressure observations are 1.9 kPa and 1.3 kPa, with corresponding confidence levels of 0.92 and 0.28. After confidence-weighted fusion, the estimated internal support pressure is determined to be 1.86 kPa. The backfill pressure measured by the pressure sensing unit outside unit A is 4.7 kPa, with a confidence level of 0.88. The external pressures of adjacent units B and C are 4.4 kPa and 5.0 kPa, respectively. After load propagation constraint correction, the estimated external backfill pressure of unit A is 4.72 kPa, with a net differential pressure of 2.86 kPa. If the external pressure of element D on the other side of the crack boundary is 3.5 kPa, the propagation weight is reduced due to the crack obstruction, preserving the pressure gradient between A and D without forced smoothing. Through the above fusion, a continuous, traceable, and uncertain wall pressure state field can be formed, improving the accuracy of local stress identification and subsequent zonal pressure regulation decisions.

[0046] In some embodiments, a bidirectional differential pressure safety boundary is determined for each monitoring unit based on the wall condition. The wall pressure state field is compared with the bidirectional differential pressure safety boundary to determine the risk monitoring unit, including: Based on the wall condition and boundary constraints of each monitoring unit, the directional bearing capacity against external loads and internal support effects is determined respectively. The directional bearing capacity is safely corrected based on the uncertainty of the wall condition, and a bidirectional differential pressure safety boundary is generated that limits the bidirectional variation range of net differential pressure. Extract the net differential pressure, trend of change, and spatial pressure gradient of each monitoring unit from the wall pressure state field to determine the corresponding differential pressure safety margin; Risk characterization is generated based on differential pressure safety margin and change status, and monitoring units that exceed the bidirectional differential pressure safety boundary or meet trend risk conditions are identified as risk monitoring units.

[0047] Specifically, the bidirectional differential pressure safety boundary is used to limit the two stress states: external backfill pressure dominance and internal support pressure dominance. Since the thin-walled archaeological site walls typically resist bending inwards and pushing outwards from the cavity differently, a single absolute differential pressure threshold cannot be used. The wall thickness, curvature, material strength, crack density, erosion degree, and edge support conditions of each monitoring unit can be read from the site's spatial model, and the stress direction can be determined by combining this with the location of the monitoring unit at the kiln top, sidewall, corner, or bottom.

[0048] The bearing capacity in the external load direction represents the wall's ability to resist the compression of backfill soil into the cavity, and can be determined based on local compressive strength, bending span, external restraint, and the degree of crack penetration. The bearing capacity in the internal support direction represents the wall's ability to resist the outward jacking of the flexible support, and can be determined based on the integrity of the inner surface, local curvature, contact area, and edge restraint. For thin-walled, high-curvature, cracked, or suspended boundary areas, the bearing capacity in both directions is correspondingly reduced.

[0049] Directional bearing capacity can be determined by combining material testing results with simplified mechanical calculations. First, the foundation bearing capacity value is generated based on material strength characteristics, wall thickness, and effective span. Then, it is corrected using curvature correction coefficients, crack weakening coefficients, and edge restraint coefficients. For external load directions, the area of ​​backfill soil and the load diffusion range can also be considered; for internal support directions, the flexible membrane layer bonding area and localized concentrated loads must also be considered. If the monitoring unit consists of multiple material zones, the directional bearing capacity is determined according to the principle of controlling the most unfavorable area, avoiding the average value from masking local weak points. Directional bearing capacity is expressed as net differential pressure; when the external backfill pressure is greater than the internal support pressure, it is recorded as positive; when the internal support pressure is greater than the external backfill pressure, it is recorded as negative.

[0050] The uncertainty of the wall condition is used to avoid overly broad boundaries caused by detection and model errors. The state uncertainty coefficient can be generated based on the completeness of spatial measurements, the dispersion of material detection, the confidence level of crack identification, the stability of historical states, and the uncertainty of the pressure state field. When the detection data is complete, the results from different measuring points are consistent, and the state is stable over a long period, a smaller safety reduction is used; when there are obstructed areas, large fluctuations in material strength, or unclear crack boundaries, the safety reduction is increased. The forward and reverse bearing capacities are reduced separately to obtain the upper limit of the forward differential pressure and the lower limit of the reverse differential pressure, forming an asymmetric bidirectional differential pressure safety boundary. For monitoring units with crack propagation or continuously decreasing wall thickness, the boundary can also be dynamically contracted according to recent state changes.

[0051] The differential pressure safety margin is determined by the distance between the current net differential pressure and the corresponding directional boundary. When the net differential pressure is positive, the upper limit of the positive differential pressure is calculated minus the current net differential pressure; when the net differential pressure is negative, the distance between the current net differential pressure and the lower limit of the reverse differential pressure is calculated. In addition to the current margin, the rate of change, acceleration, and direction of change of the net differential pressure for multiple consecutive cycles are extracted from the wall pressure state field, and the pressure gradient between the monitoring unit and adjacent units is calculated. If the current net differential pressure has not yet exceeded the limit, but the safety margin continues to decrease, the rate of increase of the net differential pressure exceeds the trend threshold, or the local pressure gradient increases rapidly at the crack tip and corner region, then the trend risk condition is determined to be met. For short-term construction disturbances, the duration and backfilling events can be combined for exclusion to avoid directly classifying instantaneous impacts as continuous risks.

[0052] Risk characterization can be generated by weighting the degree of boundary exceedance, differential pressure safety margin, margin decay rate, spatial pressure gradient, and estimation uncertainty. When the net differential pressure exceeds the positive upper limit, it indicates that the external backfill load is insufficient relative to the internal support; when the net differential pressure is below the reverse lower limit, it indicates that the internal support pressure is too high. When the boundary has not been exceeded but the trend risk condition is met, it is marked as an early warning risk monitoring unit; when the boundary has been exceeded or the high gradient area continues to expand, it is marked as a control risk monitoring unit, and the corresponding direction and risk urgency are provided to the subsequent target support pressure field generation process.

[0053] For example, the foundation bearing capacity in the external load direction of top monitoring unit A is 6.2 kPa, and the foundation bearing capacity in the internal support direction is 4.5 kPa. Considering the small wall thickness, the presence of minor cracks, and a state uncertainty coefficient of 0.18, the upper limit of the positive differential pressure is corrected to 5.1 kPa, and the lower limit of the reverse differential pressure is corrected to -3.7 kPa. Currently, the external backfill pressure is 6.4 kPa, the internal support pressure is 2.2 kPa, the net differential pressure is 4.2 kPa, and the positive safety margin is 0.9 kPa, which has not yet exceeded the limit. However, the net differential pressure increased from 3.1 kPa to 4.2 kPa over five consecutive sampling periods, the safety margin continued to decrease, and the pressure gradient with adjacent units reached 1.4 kPa / m. Therefore, unit A is identified as a warning risk monitoring unit. The net differential pressure of side wall unit B is -4.0 kPa, which is lower than the lower limit of the reverse differential pressure of -3.5 kPa, so it is directly identified as a control risk monitoring unit for excessive internal support. By establishing a safety boundary that is asymmetrical in direction and corrected according to changes in condition, both insufficient and excessive support can be identified simultaneously, and the pertinence of risk assessment for weak wall sections can be improved.

[0054] In some embodiments, a target support pressure field is generated based on the current pressure state of the risk monitoring unit and the predicted results of subsequent backfill loads. The target support pressure field is then converted into target pressures for each support chamber using the support action relationship, including: Based on the backfilling operation status and the temporal changes in external backfilling pressure, predict the backfilling load of each risk monitoring unit within the target control period; Based on the backfill load, current pressure status, and bidirectional differential pressure safety boundary, determine the target support pressure that will maintain the predicted net differential pressure within the preset safety margin; Based on the pressure continuity between adjacent monitoring units and the allowable stress variation range of the wall, the target support pressure is collaboratively corrected to generate the target support pressure field. Based on the inversion of the support action relationship, the pressure contribution of each support chamber to the target support pressure field is determined. Under the condition of satisfying the chamber pressure regulation boundary and the pressure difference constraint between adjacent chambers, the target pressure of each support chamber is determined.

[0055] Specifically, the target control cycle can be determined based on the backfilling speed, pressure change rate, and pressure regulating device response time, and is used to predict the external backfilling load that the risk monitoring unit may bear in the next stage. First, the current backfilling height, filler placement location, layer thickness, compaction status, construction cycle, and equipment operating area are obtained, and the external backfilling pressure sequence of each risk monitoring unit is extracted from the wall pressure state field.

[0056] The pressure sequence is decomposed to distinguish between continuous loads resulting from increased backfill height, stage loads caused by localized compaction, and short-term pulses caused by construction disturbances. For continuous loads, extrapolation can be performed based on the pressure increment per unit backfill height and the current advance speed. For stage loads, the increment range can be determined by combining historical pressure responses at the same work location. For short-term pulses, they are only included in the prediction results when the duration reaches a set condition. Furthermore, based on the load propagation direction and time delay between adjacent monitoring units, the expected new load is propagated to neighboring areas to obtain the predicted backfill load values ​​and prediction uncertainties for each risk monitoring unit within the target control period.

[0057] The target support pressure is used to keep the predicted net differential pressure within the two-way differential pressure safety boundary, while maintaining a safety margin to absorb prediction errors and construction disturbances. In the calculation, the current external backfill pressure is added to the predicted additional load to obtain the predicted external pressure, which is then combined with the current internal support pressure to determine the predicted net differential pressure under unadjusted conditions. If the predicted net differential pressure is close to the upper limit of the positive differential pressure, the internal support pressure is increased; if the predicted net differential pressure is close to the lower limit of the negative differential pressure, the internal support pressure is decreased.

[0058] The preset safety margin can be dynamically determined based on the uncertainty of the wall condition, the uncertainty of load prediction, and the urgency of the risk. A larger margin is used for monitoring units with thin walls, dense cracks, or low observation integrity. The target support pressure is equal to the wall contact pressure required to bring the predicted net differential pressure back to the corresponding position inside the safe range, and is limited by the allowable pressure change range of the monitoring unit to prevent a rapid reversal of the wall force direction caused by a single pressure adjustment.

[0059] The target support pressure calculated independently by each monitoring unit may exhibit local abrupt changes, thus requiring spatial collaborative correction. Pressure continuity constraints can be established based on the adjacency relationship of monitoring units, wall curvature continuity, wall thickness differences, and crack isolation status. For adjacent units with continuous structures and acting jointly by the same support chamber, the target support pressure difference and pressure gradient are limited to ensure a smooth transition in the adjusted stress distribution. For crack boundaries, areas of abrupt wall thickness changes, and corner areas, the continuity constraint strength is reduced while retaining necessary differentiated support. Collaborative correction prioritizes achieving a safety margin for risk monitoring units while minimizing pressure variations in non-risk monitoring units. It also verifies whether the predicted net differential pressure of each unit remains within the bidirectional differential pressure safety boundary after correction, ultimately forming a target support pressure field covering all monitoring units.

[0060] The target support pressure field needs to be converted into a target pressure that each support chamber can handle. Based on the action weights, contact points, and pressure transmission efficiency in the support action relationships, an action matrix can be established to represent the change in effective pressure within a chamber and its impact on the support pressure in the monitoring unit. The difference between the target support pressure field and the current internal support pressure field is used as the pressure to be compensated, and the pressure contribution that each support chamber should provide is then calculated.

[0061] During the solution process, the optimization direction is to minimize the target support pressure deviation and the chamber pressure adjustment amount, and the minimum pressure, maximum pressure, single-cycle change, pressure change rate, and pressure difference between adjacent chambers are limited for each chamber. For monitoring units with multiple chambers acting together, pressure contribution is allocated first according to the efficiency of action. For cases where a single chamber simultaneously affects both risk and non-risk areas, the adjustment range of that chamber is reduced and compensation from adjacent chambers is invoked to avoid over-support in non-risk areas. If the target support pressure field exceeds the achievable range of all chambers, high-risk units are prioritized according to the urgency of risk, and a backfill pause or manual reinforcement prompt is generated.

[0062] For example, the current external backfill pressure of top monitoring unit A is 4.7 kPa, the internal support pressure is 1.9 kPa, and the upper limit of the positive differential pressure is 5.1 kPa. Based on the thickness and compaction location of the next layer of backfill, the external pressure is predicted to increase by 1.4 kPa within the target control period, reaching a predicted external pressure of 6.1 kPa. To maintain a safety margin of 1.0 kPa, the target net differential pressure is determined to be 4.1 kPa, corresponding to a target support pressure of 2.0 kPa. Due to the thinner wall of adjacent unit B, the initial target support pressure is 2.6 kPa. After continuous pressure constraint correction, A and B are adjusted to 2.15 kPa and 2.45 kPa, respectively. The action matrix shows that the action weights of chamber 2 on A and B are 0.58 and 0.28, respectively, and the corresponding weights for chamber 3 are 0.17 and 0.46. The solution shows that the target pressure in chamber 2 increases from 14 kPa to 14.8 kPa, and in chamber 3 from 12.5 kPa to 13.4 kPa, while the pressure difference between adjacent chambers remains within the allowable range. Through this process, the predicted load and the wall safety state can be converted into executable zoned pressure regulation parameters, allowing the support pressure to dynamically match the backfilling process.

[0063] In some embodiments, the pressure contribution of each support chamber to the target support pressure field is inverted based on the support action relationship. Under the condition of satisfying the chamber pressure regulation boundary and the pressure difference constraint between adjacent chambers, the target pressure of each support chamber is determined, including: Based on the weight of action and pressure transmission efficiency in the support action relationship, an action matrix is ​​constructed to show the effect of pressure change in the support chamber on pressure change in the monitoring unit. By combining the uncertainty of the action matrix and the relationship between support actions, candidate action scenarios representing different pressure transmission states are generated; A pressure inversion model is constructed using the deviation between the support pressure and the target support pressure field under each candidate action scenario, the change in chamber pressure, and the pressure difference between adjacent chambers as joint constraint objectives. The pressure inversion model is solved within the feasible region defined by the bidirectional differential pressure safety boundary and the chamber pressure regulation boundary, and the corresponding chamber pressure is determined as the target pressure.

[0064] Specifically, pressure inversion is used to convert the target support pressure at the monitoring unit level into a pressure setpoint that the support chamber can execute. First, an action matrix is ​​constructed with the support chamber as columns and the monitoring unit as rows. The matrix elements are determined by the action weight of the corresponding chamber to the monitoring unit, the pressure transmission efficiency after bonding, and the current effective contact state. For chamber pressures below the bonding start point, the effective pressure change is not included; after exceeding the bonding start point, the chamber pressure increment is multiplied by the pressure transmission efficiency and the action weight to obtain the support pressure contribution to the corresponding monitoring unit.

[0065] In some examples, if the flexible membrane flattens, slips, or changes in contact area, the matrix elements can be corrected based on the chamber pressure and contact pressure responses within the most recent control cycle. When a monitoring unit is affected by multiple chambers, the contributions of each chamber are summed; when a chamber affects multiple monitoring units, the differences in each action path are preserved to avoid proportionally distributing the chamber pressure.

[0066] The relationship between support actions is affected by the state of flexible materials, wall morphology, and installation deviations, and the action matrix is ​​not fixed. Multiple candidate action scenarios can be generated based on the action weight reset confidence interval, pressure transmission efficiency fluctuation range, and fit-in point deviation obtained from trial pressure regulation. The baseline scenario uses the current estimate; the weak transmission scenario reduces the transmission efficiency in areas with folds, gaps, or low-confidence observations; the offset scenario adjusts the action weights between adjacent monitoring units; and the coupling enhancement scenario increases the cross-contribution of the shared action area between adjacent chambers.

[0067] The number of candidate scenarios can be determined based on the degree of uncertainty in the relationship. When the uncertainty is low, only the baseline scenario and the upper and lower boundary scenarios are retained. When the uncertainty is high, scenarios with offset of the effective range and local failure scenarios are added. Each scenario forms a corresponding action matrix and is configured with scenario occurrence weights to verify the feasibility and safety of the target pressure under different transmission states.

[0068] The pressure inversion model uses the target pressure of each support chamber as the variable to be determined. For each candidate action scenario, the target pressure is subtracted from the current chamber pressure to obtain the chamber pressure change, and then the predicted support pressure for each monitoring unit is calculated using the corresponding action matrix. The model's primary objective is to minimize the deviation between the predicted support pressure and the target support pressure field, while simultaneously limiting the total chamber pressure change and the pressure difference between adjacent chambers. For monitoring units with high risk urgency and small differential pressure safety margin, the penalty weight for pressure matching deviation is increased; for non-risk monitoring units, the focus is on limiting the deviation of support pressure from the current state. The chamber pressure change constraint is used to reduce unnecessary large-scale pressurization or depressurization, and the adjacent chamber pressure difference constraint is used to prevent the formation of excessive local gradients in the flexible support. Multi-scenario worst-case deviation constraints can also be added to ensure that the same set of target pressures does not cause significant under- or over-support in any candidate scenario.

[0069] The feasible region is jointly defined by the bidirectional differential pressure safety boundary and the chamber pressure regulation boundary. For each monitoring unit, the predicted net differential pressure is calculated based on the predicted external backfill pressure and the predicted support pressure under the candidate scenario, and the predicted net differential pressure is required to be between the upper limit of the positive differential pressure and the lower limit of the reverse differential pressure, while retaining a preset safety margin. For each support chamber, the minimum holding pressure, the maximum allowable pressure, the maximum change in a single cycle, and the pressure change rate are defined; for adjacent chambers, the target pressure difference and the conflict of change direction are defined.

[0070] If a candidate solution meets the requirements in the baseline scenario but causes a positive out-of-bounds error in the weak transmission scenario, the solution is deemed not robust and feasible. A constrained iterative optimization method can be used to solve the problem, gradually adjusting variables starting from the current chamber pressure, calculating pressure deviations and boundary constraint violations in multiple scenarios each time, until the objective function converges or reaches the iteration upper limit. If no solution satisfies all constraints, the pressure matching requirements for non-critical monitoring units are relaxed sequentially according to risk level, but the bidirectional differential pressure safety boundary is not relaxed, and a pause backfilling or supplementary support command is output.

[0071] For example, the top monitoring units A, B, and C require additional support pressures of 0.35 kPa, 0.50 kPa, and 0.20 kPa, respectively. The effective action coefficients of the second chamber for A, B, and C are 0.27, 0.13, and 0.05, respectively, and the corresponding coefficients for the third chamber are 0.08, 0.24, and 0.16, respectively. Considering local folds in the second chamber, its transmission efficiency may decrease by 15%, thus generating a baseline scenario and a weak transmission scenario. Currently, the pressures in the second and third chambers are 14.0 kPa and 12.5 kPa, respectively, with a maximum single-cycle increment of 1.5 kPa for both, and an allowable pressure difference of 2.5 kPa between adjacent chambers. The model solves for a target pressure of 14.8 kPa for the second chamber and 13.4 kPa for the third chamber. After verification in two scenarios, the predicted support pressure deviations for A, B, and C were all below 0.12 kPa, the predicted net differential pressure remained within the bidirectional boundaries, and the pressure difference between adjacent chambers was 1.4 kPa. This allows for obtaining robust target pressures that meet safety constraints even when there are fluctuations in the interaction relationship, reducing the impact of single matrix estimation bias on the zonal pressure regulation results.

[0072] In some embodiments, each support chamber is adjusted according to the target pressure, and the wall pressure state field is updated based on the adjusted multi-source pressure data until the backfilling operation is completed, including: Based on the current pressure and target pressure of each support chamber, a zone pressure regulation command is generated, and pressurization or depressurization is performed under pressure change constraints. Collect multi-source pressure data after pressure adjustment, determine the actual support response based on the support action relationship, and update the internal support pressure, external backfill pressure and net differential pressure of each monitoring unit; The updated net differential pressure is compared with the bidirectional differential pressure safety boundary. If the risk monitoring unit fails to return to a safe state, the target pressure of the support chamber is corrected and the pressure is readjusted. When the risk range expands or the pressure regulation results do not meet the safety conditions, the backfilling of the area is suspended and the area enters an abnormal support state; when each monitoring unit continues to meet the safety conditions and the backfilling progress reaches the completion condition, the closed-loop pressure regulation ends.

[0073] Specifically, closed-loop pressure regulation uses the current pressure, target pressure, allowable rate of change, and risk urgency of the corresponding risk monitoring unit as inputs for each support chamber to generate zoned pressure regulation commands. First, the pressure deviation between the target pressure and the current pressure is calculated. A positive deviation generates a pressurization command, a negative deviation generates a depressurization command, and the current pressure is maintained when the deviation is within the control dead zone. To avoid abrupt changes in the overall morphology of the flexible support structure caused by simultaneous changes in multiple chambers, the pressure regulation order can be determined according to risk urgency, pressure deviation magnitude, and chamber coupling degree, prioritizing the regulation of chambers directly acting on high-risk monitoring units and with weak coupling to other chambers. For multiple chambers acting on the same monitoring unit, a step-by-step alternating adjustment method can be used, maintaining stability and collecting the response after each pressure increment before determining the next increment.

[0074] The controller adjusts the opening and duration of the corresponding pump valves for pressurization or depressurization. The single adjustment amount, pressure change per unit time, and pressure difference between adjacent chambers must not exceed the set boundaries. When approaching the target pressure, the pump valve opening is gradually reduced to avoid over-adjustment due to pressure inertia. Depressurization is performed in stages to prevent a sudden drop in internal support pressure from causing the external backfill load to momentarily dominate. If the chamber pressure change rate is abnormal, there is no effective pressure response after valve action, or unexpected linkage occurs between adjacent chambers, the current command is stopped, and the corresponding chamber is marked as pending inspection. During pressure regulation, the number of chambers simultaneously pressurizing and depressurizing within the same wall area can be limited, reducing the possibility of rapid reversal of the local pressure gradient.

[0075] After each pressure regulation phase, chamber pressure, wall contact pressure, and external backfill pressure are acquired according to a unified sampling sequence, and deviation correction, timing alignment, and reliability assessment are performed. Based on the support action relationship, the actual pressure change in the chamber is converted into the predicted support response of each monitoring unit, and then compared with the measured increment of contact pressure. When the predicted response matches the measured response, the internal support pressure is updated according to the measured value; when the measured response is lower but the data reliability is higher, it is determined that the actual pressure transmission efficiency has decreased, and the corresponding action weight is adjusted; when the measured data reliability is lower, the predicted response is retained and the estimation uncertainty is increased. The external backfill pressure is updated collaboratively based on direct observation, load propagation of adjacent monitoring units, and the current backfill status, and then the net differential pressure, rate of change, and spatial pressure gradient are recalculated to form an updated wall pressure state field.

[0076] The updated net differential pressure is compared with the corresponding bidirectional differential pressure safety boundary. When the net differential pressure of the risk monitoring unit returns to the safe range and the safety margin meets the set requirements, the monitoring unit is switched to continuous observation mode, and the current chamber pressure is maintained until the next control cycle. When the net differential pressure still exceeds the boundary, or although it does not exceed the boundary but the safety margin continues to decrease, the target pressure is adjusted according to the deviation between the actual support response and the expected support response. If the actual transmission efficiency is lower than expected, the target pressure can be appropriately increased within the chamber pressure adjustment boundary, or adjacent chambers that have a secondary effect on the risk monitoring unit can be called in for joint compensation; if the internal support pressure is excessive, the pressure of the primary functioning chamber is reduced, and the synchronous pressurization of adjacent chambers is restricted. After each correction, the feasible region check is re-performed to ensure that other monitoring units do not generate new reverse boundary overflows due to compensation adjustments.

[0077] Abnormal support status is used to handle situations where conventional closed-loop regulation cannot restore safety. When the number of risk monitoring units continues to increase, the risk expands outward along adjacent relationships, the net differential pressure continuously exceeds the limit, or the target support pressure cannot be provided even after the chamber reaches the pressure regulation limit, the packing delivery, compaction, or equipment movement in the corresponding area is suspended, and non-abnormal chambers are kept at a safe pressure. The system performs differentiated processing based on the source of the abnormality: when a chamber leaks, the corresponding gas path is isolated and adjacent chambers are called for compensation; when contact breaks, the pressure regulation amplitude is reduced and the contact point is re-identified; when the pressure sensing unit is abnormal, the nearest observation and action relationship estimation are used as substitutes; when the wall condition is abnormal, manual inspection and temporary reinforcement information are output. Backfilling of the area can only be resumed after the abnormality is eliminated and the safety conditions are met for several consecutive sampling cycles.

[0078] For example, due to increased external backfill pressure, a 0.7 kPa support gap was formed in the top monitoring unit A, and the target pressure in the second chamber was adjusted from 14.0 kPa to 14.8 kPa. The controller divided the adjustment process into four 0.2 kPa increments, each held for 20 seconds. After completing the first two stages, the contact pressure increased by only 0.09 kPa, lower than the expected value of 0.16 kPa. The system corrected the pressure transmission efficiency from 0.46 to 0.39 and increased the target pressure in the third chamber by 0.3 kPa to compensate. After readjustment, the net differential pressure in unit A decreased from 5.3 kPa to 4.0 kPa, the positive safety margin was restored to 1.1 kPa, and the pressure gradient between adjacent units was within the allowable range. During subsequent backfilling, each monitoring unit remained safe for 10 consecutive control cycles. When the backfill height reached the design endpoint and the pressure change stabilized, the system ended active closed-loop pressure regulation and switched to pressure maintenance monitoring. Through the above process, a closed-loop control system can be formed that connects pressure setting, execution feedback, state reconstruction, risk review, and target correction, thereby improving the stability of zoned support and the ability to handle anomalies during the backfilling process.

[0079] The following are system embodiments of this application, which can be used to execute the method embodiments of this application. For details not disclosed in the system embodiments of this application, please refer to the method embodiments of this application.

[0080] Figure 2 This is a schematic diagram of the adaptive support monitoring system for thin-walled cavity archaeological sites based on pressure fusion, provided in an embodiment of this application. Figure 2 As shown, the system includes: The acquisition module 201 is used to acquire the spatial morphology and wall state of the thin-walled cavity site, divide the site wall into multiple monitoring units, and establish the relationship between the support chambers, pressure sensing units and monitoring units of the partitioned flexible support body. The pressure regulating module 202 is used to perform limited exploratory pressure regulation on each support chamber and determine the support relationship between the support chamber and the monitoring unit based on the chamber pressure change and contact pressure response. The correction module 203 is used to acquire the chamber pressure, wall contact pressure and external backfill pressure during the backfilling process, and to perform spatiotemporal correction and reliability assessment on the multi-source pressure data. The fusion module 204 is used to fuse multi-source pressure data based on correlation, support relationship and load transfer relationship between monitoring units to generate a wall pressure state field. Comparison module 205 is used to determine the bidirectional differential pressure safety boundary of each monitoring unit based on the wall condition, compare the wall pressure state field with the bidirectional differential pressure safety boundary, and determine the risk monitoring unit; The generation module 206 is used to generate a target support pressure field based on the current pressure status of the risk monitoring unit and the prediction results of subsequent backfill load, and to convert the target support pressure field into the target pressure of each support chamber using the support action relationship. The update module 207 is used to adjust each support chamber according to the target pressure and update the wall pressure state field according to the adjusted multi-source pressure data until the backfilling operation is completed.

[0081] In some embodiments, Figure 2 The acquisition module 201 collects spatial measurement data and wall detection data of the inner and outer surfaces of the thin-walled cavity site, and constructs a site spatial model representing the geometric shape and structural state of the wall surface; it determines the partition boundaries based on the morphological continuity and state differences in the site spatial model, and divides the site wall into multiple monitoring units with relatively consistent stress characteristics; it places the partitioned flexible support body in the spatial reference corresponding to the site spatial model, and determines its projection area on the site wall based on the effective range of each support chamber and the sensing range of each pressure sensing unit; it establishes correlation weights according to the degree of spatial overlap between the projection area and the monitoring unit, and generates the correlation relationship between the support chamber, the pressure sensing unit and the monitoring unit.

[0082] In some embodiments, Figure 2 Within a pressure regulation range below the safe pressure of the corresponding monitoring unit, the pressure regulating module 202 sequentially applies graded pressure disturbances to each support chamber, collecting the chamber pressure and contact pressure corresponding to each pressure level; based on the contact pressure increment and spatial distribution before and after the pressure disturbance, it determines the pressure response degree of each support chamber to each monitoring unit; it identifies the empty stroke between the chamber pressure change and the establishment of the contact pressure, and determines the contact start point and pressure transmission efficiency based on the relationship between the contact pressure and the chamber pressure; based on the pressure response degree, contact start point, and pressure transmission efficiency, it generates a support action relationship characterizing the weight of each support chamber's action on each monitoring unit.

[0083] In some embodiments, Figure 2 The correction module 203 collects pressure data from each support chamber and corresponding monitoring unit according to a unified sampling sequence, and completes spatial mapping based on the correlation relationship; it corrects the measurement deviation of the pressure data according to the baseline state of the pressure sensing unit and the environmental influence, and performs timing alignment according to the pressure response delay; it determines the observation residual of each pressure data based on the spatial consistency, continuous change characteristics and support relationship of adjacent pressure data; it generates data credibility based on the observation residual and the operational stability of the pressure sensing unit, and deweights abnormal pressure data to obtain corrected multi-source pressure data.

[0084] In some embodiments, Figure 2 The fusion module 204 maps the corrected multi-source pressure data to the corresponding monitoring units according to the correlation relationship, and constructs pressure state nodes that characterize external loads, internal supports, and observation uncertainties. It constrains the chamber pressure and wall contact pressure according to the support action relationship, and determines the estimated internal support pressure of each monitoring unit based on the data reliability. It establishes spatial propagation constraints based on the load transfer relationship between monitoring units, performs collaborative estimation of external backfill pressure, and retains the pressure gradient at the location of structural state change. It determines the net differential pressure and its variation characteristics based on the estimated internal support pressure and the estimated external backfill pressure, and generates the wall pressure state field.

[0085] In some embodiments, Figure 2 The comparison module 205 determines the directional bearing capacity against external loads and internal support effects based on the wall state and boundary constraints of each monitoring unit; it performs a safety correction on the directional bearing capacity based on the uncertainty of the wall state, and generates a bidirectional differential pressure safety boundary that limits the bidirectional variation range of net differential pressure; it extracts the net differential pressure, variation trend and spatial pressure gradient of each monitoring unit from the wall pressure state field, and determines the corresponding differential pressure safety margin; it generates a risk characterization based on the differential pressure safety margin and variation state, and identifies monitoring units that exceed the bidirectional differential pressure safety boundary or meet the trend risk conditions as risk monitoring units.

[0086] In some embodiments, Figure 2 The generation module 206 predicts the backfill load of each risk monitoring unit within the target control period based on the backfill operation status and the temporal changes of the external backfill pressure; combined with the backfill load, the current pressure status, and the bidirectional differential pressure safety boundary, it determines the target support pressure that maintains the predicted net differential pressure within the preset safety margin; based on the pressure continuity between adjacent monitoring units and the allowable stress variation range of the wall, it collaboratively corrects the target support pressure to generate the target support pressure field; based on the support action relationship, it inverts the pressure contribution of each support chamber to the target support pressure field, and determines the target pressure of each support chamber under the condition of satisfying the chamber pressure regulation boundary and the pressure difference constraint between adjacent chambers.

[0087] In some embodiments, Figure 2 The generation module 206 constructs an action matrix from the change in pressure in the support chamber to the change in pressure in the monitoring unit based on the action weight and pressure transmission efficiency in the support action relationship. Combining the action matrix and the uncertainty of the support action relationship, it generates candidate action scenarios that characterize different pressure transmission states. Using the deviation between the support pressure and the target support pressure field, the change in chamber pressure, and the pressure difference between adjacent chambers under each candidate action scenario as joint constraint targets, it constructs a pressure inversion model. The pressure inversion model is solved within the feasible region defined by the bidirectional differential pressure safety boundary and the chamber pressure regulation boundary, and the corresponding chamber pressure is determined as the target pressure.

[0088] In some embodiments, Figure 2 The update module 207 generates zoned pressure regulation commands based on the current pressure and target pressure of each support chamber, and performs pressurization or depressurization under pressure change constraints; it collects multi-source pressure data after pressure regulation, determines the actual support response based on the support action relationship, and updates the internal support pressure, external backfill pressure, and net differential pressure of each monitoring unit; it compares the updated net differential pressure with the bidirectional differential pressure safety boundary, and when the risk monitoring unit has not returned to a safe state, it corrects the target pressure of the support chamber and regulates the pressure again; when the risk range expands or the pressure regulation result does not meet the safety conditions, it suspends regional backfilling and enters an abnormal support state; when each monitoring unit continuously meets the safety conditions and the backfilling progress reaches the completion condition, it ends the closed-loop pressure regulation.

[0089] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0090] Figure 3 This is a schematic diagram of the electronic device 3 provided in an embodiment of this application. Figure 3As shown, the electronic device 3 of this embodiment includes: a processor 301, a memory 302, and a computer program 303 stored in the memory 302 and executable on the processor 301. When the processor 301 executes the computer program 303, it implements the steps in the various method embodiments described above. Alternatively, when the processor 301 executes the computer program 303, it implements the functions of each module / unit in the various system embodiments described above.

[0091] Electronic device 3 can be a desktop computer, laptop, handheld computer, cloud server, or other electronic device. Electronic device 3 may include, but is not limited to, processor 301 and memory 302. Those skilled in the art will understand that... Figure 3 This is merely an example of electronic device 3 and does not constitute a limitation on electronic device 3. It may include more or fewer components than shown, or different components.

[0092] The processor 301 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0093] The memory 302 can be an internal storage unit of the electronic device 3, such as a hard disk or memory of the electronic device 3. The memory 302 can also be an external storage device of the electronic device 3, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the electronic device 3. The memory 302 can also include both internal and external storage units of the electronic device 3. The memory 302 is used to store computer programs and other programs and data required by the electronic device.

[0094] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0095] If integrated modules / units are implemented as software functional units and sold or used as independent products, they can be stored in a readable storage medium (e.g., a computer-readable storage medium). Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program may include computer program code, which may be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable storage medium may include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0096] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for adaptive support monitoring of thin-walled cavity archaeological sites based on pressure fusion, characterized in that, include: The spatial morphology and wall condition of the thin-walled cavity site were obtained, the site wall was divided into multiple monitoring units, and the relationship between the support chambers, pressure sensing units and monitoring units of the zoned flexible support system was established. The pressure adjustment of each support chamber is limited and exploratory. Based on the chamber pressure change and contact pressure response, the support relationship between the support chamber and the monitoring unit is determined. Acquire chamber pressure, wall contact pressure and external backfill pressure during the backfilling process, and perform spatiotemporal correction and reliability assessment on the multi-source pressure data; Based on the aforementioned correlation, support relationship, and load transfer relationship between monitoring units, multi-source pressure data are fused to generate a wall pressure state field. The bidirectional differential pressure safety boundary of each monitoring unit is determined based on the wall condition. The wall pressure state field is compared with the bidirectional differential pressure safety boundary to determine the risk monitoring unit. The target support pressure field is generated based on the current pressure status of the risk monitoring unit and the prediction results of the subsequent backfill load. The target support pressure field is then converted into the target pressure of each support chamber using the support action relationship. Adjust each support chamber according to the target pressure, and update the wall pressure state field based on the adjusted multi-source pressure data until the backfilling operation is completed.

2. The method according to claim 1, characterized in that, The process involves acquiring the spatial morphology and wall condition of the thin-walled cavity site, dividing the site wall into multiple monitoring units, and establishing the correlation between the support chambers of the zoned flexible support structure, pressure sensing units, and monitoring units, including: Spatial measurement data and wall detection data of the inner and outer surfaces of the thin-walled cavity site were collected to construct a spatial model of the site that characterizes the geometric morphology and structural state of the wall surface. Based on the morphological continuity and state differences in the aforementioned site spatial model, the partition boundaries are determined, and the site wall is divided into multiple monitoring units with relatively consistent stress characteristics. The partitioned flexible support structure is placed in the spatial reference corresponding to the site space model. Based on the effective range of each support chamber and the sensing range of each pressure sensing unit, its projection area on the site wall is determined. The association weights are established based on the degree of spatial overlap between the projection area and the monitoring unit, and the association relationship between the support chamber, the pressure sensing unit and the monitoring unit is generated.

3. The method according to claim 1, characterized in that, The process of performing restricted exploratory pressure regulation on each support chamber, and determining the support relationship between the support chamber and the monitoring unit based on chamber pressure changes and contact pressure responses, includes: Within the pressure regulation range below the corresponding monitoring unit's safe pressure, graded pressure disturbances are applied sequentially to each support chamber, and the chamber pressure and contact pressure corresponding to each pressure level are collected. Based on the increase and spatial distribution of contact pressure before and after the pressure disturbance, determine the degree of pressure response of each support chamber to each monitoring unit; Identify the empty travel between the change in chamber pressure and the establishment of contact pressure, and determine the contact starting point and pressure transmission efficiency based on the relationship between contact pressure and chamber pressure. Based on the pressure response level, contact point, and pressure transmission efficiency, a support effect relationship is generated that characterizes the weight of each support chamber on each monitoring unit.

4. The method according to claim 1, characterized in that, The acquisition of chamber pressure, wall contact pressure, and external backfill pressure during the backfilling process, and the spatiotemporal correction and reliability assessment of the multi-source pressure data, include: Pressure data of each support chamber and corresponding monitoring unit are collected according to a unified sampling sequence, and spatial mapping is completed based on the aforementioned correlation. The pressure data is corrected for measurement deviation based on the baseline state of the pressure sensing unit and environmental influences, and timing alignment is performed according to the pressure response delay. Based on the spatial consistency and continuous variation characteristics of adjacent pressure data and the relationship of the support action, the observation residuals of each pressure data are determined. Data reliability is generated based on the observation residuals and the operational stability of the pressure sensing unit, and abnormal pressure data is downweighted and labeled to obtain corrected multi-source pressure data.

5. The method according to claim 1, characterized in that, The process of fusing multi-source pressure data based on the aforementioned correlations, support relationships, and load transfer relationships between monitoring units to generate a wall pressure state field includes: According to the aforementioned correlation, the corrected multi-source pressure data is mapped to the corresponding monitoring unit to construct pressure state nodes that characterize external loads, internal support, and observation uncertainty. Based on the aforementioned support action relationship, the chamber pressure and wall contact pressure are constrained, and the estimated value of the internal support pressure of each monitoring unit is determined in combination with the data reliability. Spatial propagation constraints are established based on the load transfer relationship between monitoring units, external backfill pressure is estimated collaboratively, and pressure gradients at locations of abrupt structural changes are preserved. The net differential pressure and its variation characteristics are determined based on the estimated internal support pressure and the estimated external backfill pressure, and a wall pressure state field is generated.

6. The method according to claim 1, characterized in that, The process of determining the bidirectional differential pressure safety boundary for each monitoring unit based on the wall condition, comparing the wall pressure state field with the bidirectional differential pressure safety boundary, and determining the risk monitoring unit includes: Based on the wall condition and boundary constraints of each monitoring unit, the directional bearing capacity against external loads and internal support effects is determined respectively. The directional bearing capacity is safety-corrected based on the uncertainty of the wall condition, thereby generating a bidirectional differential pressure safety boundary that limits the bidirectional variation range of net differential pressure. Extract the net differential pressure, trend of change, and spatial pressure gradient of each monitoring unit from the pressure state field of the wall to determine the corresponding differential pressure safety margin; Based on the differential pressure safety margin and its changing state, a risk characterization is generated, and monitoring units that exceed the bidirectional differential pressure safety boundary or meet the trend risk conditions are identified as risk monitoring units.

7. The method according to claim 1, characterized in that, The process of generating a target support pressure field based on the current pressure status of the risk monitoring unit and the prediction results of subsequent backfill loads, and converting the target support pressure field into the target pressure of each support chamber using the support action relationship, includes: Based on the backfilling operation status and the temporal changes in external backfilling pressure, predict the backfilling load of each risk monitoring unit within the target control period; Based on the backfill load, current pressure status, and bidirectional differential pressure safety boundary, determine the target support pressure that will maintain the predicted net differential pressure within a preset safety margin; Based on the pressure continuity between adjacent monitoring units and the allowable stress variation range of the wall, the target support pressure is collaboratively corrected to generate a target support pressure field. Based on the inversion of the support action relationship, the pressure contribution of each support chamber to the target support pressure field is calculated. Under the condition of satisfying the chamber pressure regulation boundary and the pressure difference constraint between adjacent chambers, the target pressure of each support chamber is determined.

8. The method according to claim 7, characterized in that, The pressure contribution of each support chamber to the target support pressure field is inverted based on the support action relationship. Under the condition of satisfying the chamber pressure regulation boundary and the pressure difference constraint between adjacent chambers, the target pressure of each support chamber is determined, including: Based on the weight of action and pressure transmission efficiency in the support action relationship, an action matrix is ​​constructed to reflect the change in pressure from the support chamber to the change in support pressure in the monitoring unit. By combining the uncertainty of the action matrix and the support action relationship, candidate action scenarios representing different pressure transmission states are generated; A pressure inversion model is constructed using the deviation between the support pressure and the target support pressure field under each candidate action scenario, the change in chamber pressure, and the pressure difference between adjacent chambers as joint constraint objectives. The pressure inversion model is solved within the feasible region defined by the bidirectional differential pressure safety boundary and the chamber pressure regulation boundary, and the corresponding chamber pressure is determined as the target pressure.

9. The method according to claim 1, characterized in that, The process of adjusting each support chamber according to the target pressure and updating the wall pressure state field based on the adjusted multi-source pressure data until the backfilling operation is completed includes: Based on the current pressure and target pressure of each support chamber, a zone pressure regulation command is generated, and pressurization or depressurization is performed under pressure change constraints. Collect multi-source pressure data after pressure adjustment, determine the actual support response based on the support action relationship, and update the internal support pressure, external backfill pressure, and net differential pressure of each monitoring unit. The updated net differential pressure is compared with the bidirectional differential pressure safety boundary. If the risk monitoring unit fails to return to a safe state, the target pressure of the support chamber is corrected and the pressure is adjusted again. When the risk range expands or the pressure regulation results do not meet the safety conditions, the backfilling of the area is suspended and the area enters an abnormal support state; when each monitoring unit continues to meet the safety conditions and the backfilling progress reaches the completion condition, the closed-loop pressure regulation ends.

10. An adaptive support monitoring system for thin-walled cavity archaeological sites based on pressure fusion, characterized in that, include: The acquisition module is used to acquire the spatial morphology and wall condition of the thin-walled cavity site, divide the site wall into multiple monitoring units, and establish the correlation between the support chambers, pressure sensing units and monitoring units of the partitioned flexible support system. The pressure regulating module is used to perform limited exploratory pressure regulation on each support chamber and determine the support relationship between the support chamber and the monitoring unit based on the chamber pressure change and contact pressure response. The calibration module is used to acquire chamber pressure, wall contact pressure and external backfill pressure during the backfilling process, and to perform spatiotemporal calibration and reliability assessment on multi-source pressure data; The fusion module is used to fuse multi-source pressure data based on the aforementioned correlation, support relationship, and load transfer relationship between monitoring units to generate a wall pressure state field. The comparison module is used to determine the bidirectional differential pressure safety boundary of each monitoring unit based on the wall condition, and compare the wall pressure state field with the bidirectional differential pressure safety boundary to determine the risk monitoring unit. The generation module is used to generate a target support pressure field based on the current pressure status of the risk monitoring unit and the prediction results of subsequent backfill load, and to convert the target support pressure field into the target pressure of each support chamber using the support action relationship. The update module is used to adjust each support chamber according to the target pressure and update the wall pressure state field according to the adjusted multi-source pressure data until the backfilling operation is completed.

Images