Construction method and construction monitoring system for preventing terrace settlement in backfill soil area

Abstract

The invention discloses a construction method and a construction monitoring system for preventing terrace settlement in a backfill soil area, and relates to the technical field of civil engineering foundation treatment and construction monitoring. The construction method comprises the following steps which are executed in sequence: foundation gridding investigation and directional pretreatment, three-dimensional active reinforced network construction, intelligent layered backfilling and compaction based on feedback control, curing maintenance and top interface layer construction. According to the construction method and the construction monitoring system, by constructing a three-dimensional active reinforcement network and an intelligent compaction feedback system and utilizing continuous compaction control and a multi-source sensor network, real-time sensing and dynamic regulation and control of the compaction degree, the water content and the stress state of the backfill soil body are achieved, the uniformity of the construction quality is guaranteed, and the construction efficiency is improved. Advanced prediction and active early warning of the settlement trend are achieved through the monitoring system, an intelligent closed loop is formed through early warning linkage intervention measures, settlement is prevented from the source, the long-term stability of the terrace is remarkably improved, and the maintenance cost of the whole life cycle is greatly reduced.

Description

Technical Field

[0001] This invention relates to foundation treatment and construction monitoring technology in civil engineering, specifically to a construction method and construction monitoring system for preventing ground settlement in backfilled areas. Background Technology

[0002] In construction engineering, ground settlement in backfill areas is a common and complex engineering problem. This settlement is mainly caused by the consolidation and compression of the backfill itself, softening of the foundation soil due to changes in groundwater levels, and the long-term effects of dynamic loads during operation. Settlement often manifests as unevenness, easily leading to cracking and misalignment of the floor slab, severely affecting ground flatness and functionality, and posing a potential threat to the operational accuracy of superstructure equipment and the durability of the building. This problem is highly insidious and has a long development period, making it one of the key hidden dangers affecting the long-term quality of the project.

[0003] Traditional construction methods for preventing ground settlement in backfilled areas typically employ layered compaction, relying on the experience of construction workers to control compaction quality. This lacks real-time perception and control of the internal state of the backfill soil (such as compaction uniformity, moisture content distribution, and stress state). After construction, settlement problems can only be passively detected through periodic settlement observations, followed by remedial measures such as grouting and reinforcement. This approach is not only costly and affects usability but also fails to address the root cause of the problem. Summary of the Invention

[0004] The purpose of this invention is to provide a construction method and construction monitoring system for preventing ground settlement in backfill areas, in order to solve the problem that existing construction methods for preventing ground settlement in backfill areas rely on the experience of construction personnel to control compaction quality, lack real-time perception and control of the internal state of the backfill soil, and can only be passively dealt with after the fact.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a construction method for preventing ground settlement in backfill areas, comprising the following steps performed sequentially:

[0006] S1. Foundation grid-based investigation and directional pretreatment: The site is scanned with a pre-set grid using ground-penetrating radar, and a light dynamic penetration test is conducted at the grid nodes to generate a foundation bearing capacity cloud map; for areas with bearing capacity characteristic values ​​below 80 kPa or with hidden cavities, precise over-excavation is carried out to 0.5 meters below the old soil layer, and graded crushed stone is replaced and compacted in layers until the compaction coefficient is not less than 0.97;

[0007] S2. Construction of three-dimensional active reinforcement network: The first layer of bidirectional geogrid is laid on the treated foundation; vertical micropiles are constructed at the designed location, the micropiles are pressed into and penetrate the backfill layer and enter the old soil layer by no less than 1.5 meters, the pile body is injected with micro-expansion cement grout, and the steel bar head reserved at the top of the pile is mechanically locked to the geogrid.

[0008] S3. Intelligent layered backfilling and compaction based on feedback control:

[0009] S31. Preparation of backfill materials: Cement-modified soil is used, and its construction moisture content is dynamically determined;

[0010] S32. Layered paving and sensor pre-embedding: The backfill material is paved in layers, and monitoring sensors are pre-embedded simultaneously;

[0011] S33, Closed-loop intelligent compaction: Rolling is carried out using equipment equipped with a continuous compaction control system, and a compaction field cloud map is generated in real time.

[0012] S34. Dynamic reinforcement of weak areas: Based on the compaction field cloud map, perform targeted compaction on weak areas;

[0013] S35. Network Deepening: Repeat steps S32 to S34 until backfilling reaches the design elevation, and lay new geogrids at preset intervals, and tie them to the lower geogrids and micropiles with connectors to form a three-dimensional active reinforcement network consisting of multiple horizontal reinforcement layers and vertical micropiles.

[0014] S4. Curing and Top Interface Layer Construction: Lay a steel-plastic composite geogrid on the top layer, spray diluted soil curing agent and cover for 7 days; after curing, construct a 150mm thick C20 concrete cushion layer and implant ground settlement monitoring markers before the concrete initially sets.

[0015] Furthermore, in step S2, the nominal tensile strength of the bidirectional geogrid in both the longitudinal and transverse directions is not less than 80kN / m, and the inter-width connection adopts special plastic steel fasteners with an overlap width of not less than 200mm.

[0016] Furthermore, the micropile described in step S2 uses a steel pipe with an outer diameter of 168mm and a wall thickness of 8mm, and is injected with M30 micro-expansion cement grout at a grouting pressure of 0.5~1.0MPa.

[0017] Furthermore, the continuous compaction control system described in step S33 collects data via an acceleration sensor on the roller, wherein the compaction value... The calculation formula is: ,in This represents the root mean square value of the acceleration of the vibrating wheel. The frequency is the vibration frequency.

[0018] A construction monitoring system for preventing ground settlement in backfilled areas includes:

[0019] The multi-source sensor network includes: vertical settlement monitoring pipes pre-embedded in the backfill, with a magnetic ring settlement probe installed every 2 meters in height inside the pipes; fiber optic strain sensors attached to the ribs of each layer of geogrid; and pore water pressure gauges deployed in the site.

[0020] The data acquisition and transmission module uses a LoRa wireless network to transmit the data from the multi-source sensor network to the cloud in real time;

[0021] The intelligent analysis and early warning platform includes:

[0022] Digital twin model building unit, used to create a three-dimensional finite element model of the site based on construction parameters;

[0023] The settlement prediction unit incorporates a machine learning-based settlement prediction model, which is as follows: ,in The predicted settlement at time t. For the initial instantaneous settlement, , These are soil rheological parameters that are dynamically updated through machine learning. This is an empirical coefficient. For the monitored changes in additional stress, For the compaction value The converted soil compression modulus; this unit trains the model every 6 hours using the latest data to achieve rolling prediction of the settlement trend for the next 7 days;

[0024] The early warning release unit triggers a level 3 early warning and pushes it to relevant personnel when the predicted settlement exceeds 10mm or when the sensor data is abnormal.

[0025] The active intervention execution terminal receives instructions from the early warning release unit and executes the corresponding intervention plan, including: generating a precise grouting reinforcement map and guiding the grouting equipment operation, automatically starting and stopping preset dewatering well points, or sending load scheduling suggestions to the operation and management party.

[0026] Furthermore, the soil compression modulus Compaction value The conversion relationship is as follows ,in and The material correlation coefficient is determined through on-site calibration tests.

[0027] Furthermore, in the three-level early warning system, the first-level early warning is blue and is pushed to the project group; the second-level early warning is yellow and an SMS notification is sent to the technical person in charge; the third-level early warning is red and an alarm and handling suggestions are automatically pushed to the supervision and owner's APP.

[0028] Furthermore, the precise grouting reinforcement map includes the three-dimensional coordinates of the grouting points, the grouting depth, and the grouting volume information, which is automatically generated by the settlement prediction unit based on the predicted settlement distribution.

[0029] Compared with existing technologies, the present invention provides a construction method and construction monitoring system for preventing ground settlement in backfill areas. By constructing a three-dimensional active reinforcement network and an intelligent compaction feedback system, it changes the traditional construction mode. First, by utilizing continuous compaction control and a multi-source sensor network, it achieves real-time perception and dynamic control of the compaction degree, moisture content, and stress state of the backfill soil, ensuring the uniformity of construction quality from the source. Second, the intelligent early warning platform based on digital twins and machine learning algorithms can predict settlement trends in advance, transforming passive monitoring into proactive early warning. Finally, by directly linking early warning with preset intervention measures such as precise grouting and automatic dewatering, a complete closed loop of perception-analysis-decision-execution is formed, achieving the prevention of settlement from the root, significantly improving the long-term stability of the ground, and greatly reducing the maintenance cost throughout the entire life cycle. Attached Figure Description

[0030] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0031] Figure 1 A schematic flowchart of the construction method provided in an embodiment of the present invention;

[0032] Figure 2 This is a schematic diagram illustrating the specific process of intelligent layered backfilling and compaction based on feedback control in the construction method provided in this embodiment of the invention.

[0033] Figure 3 This is a flowchart illustrating the construction monitoring system provided in an embodiment of the present invention.

[0034] Figure 4 This is a schematic diagram illustrating the specific process of the intelligent analysis and early warning platform in the construction monitoring system provided in this embodiment of the invention. Detailed Implementation

[0035] To enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0036] As attached Figure 1 To be continued Figure 4 As shown:

[0037] Example:

[0038] This invention provides a construction method for preventing ground settlement in backfilled areas, comprising the following steps performed in sequence:

[0039] S1. Foundation grid-based exploration and directional pretreatment:

[0040] Grid scanning: Using ground-penetrating radar equipment, a full-coverage scan of the cleared site is performed in a 2m x 2m grid to detect the planar distribution of underground anomalies such as hidden cavities and weak interlayers;

[0041] Quantitative penetration test: At each of the above grid nodes, a light dynamic penetration test was conducted, the number of hammer blows was recorded, and converted into the characteristic value of the foundation bearing capacity;

[0042] Cloud map generation and decision-making: By fusing radar imagery with probe data, a digital "foundation bearing capacity cloud map" is created. The cloud map uses different colors to intuitively display the bearing capacity at various locations, and automatically marks red warning zones with bearing capacity characteristic values ​​below 80 kPa and void areas identified by radar.

[0043] Targeted pretreatment: Construction is carried out only in the warning area shown in the cloud map. An excavator is used to precisely excavate 1.5 meters outward from the boundary of the warning area until the weak layer is penetrated and the underlying solid old soil layer (such as dense clay or weathered rock) is reached. Then, the excavation is over-excavated by 0.5 meters. Subsequently, graded crushed stone (particle size 5-40mm) is backfilled in layers, with each layer having a loose thickness of ≤300mm. A small vibratory roller or plate compactor is used to compact the layer, ensuring that the compaction coefficient of each layer is ≥0.97. Only after the layer passes the test can the next layer be backfilled until it is level with the original ground.

[0044] The traditional "blind exploration" method, which relies on sporadic drilling, has been upgraded to a comprehensive quantitative diagnosis. Through data visualization, the "precise location" and "quantitative operation" of foundation defects have been achieved, avoiding the high cost and risks of inadequate local treatment associated with traditional large-scale blind replacement. This has established a uniform and high-bearing-capacity starting platform for the entire project.

[0045] S2, Construction of 3D Actively Reinforced Network:

[0046] First-layer horizontal reinforcement: On the pre-treated flat foundation, lay the first layer of bidirectional high-density polyethylene geogrid. When laying, it must be stretched taut and without wrinkles. Adjacent strips are connected with special high-strength plastic steel fasteners. The longitudinal overlap width is ≥200mm, and the transverse joint is fixed to the foundation with U-shaped nails.

[0047] Vertical micropile construction: According to the design drawings, key load points such as column positions and equipment foundations are laid out. A static pressure pile driver is used to vertically press an open steel pipe with an outer diameter of 168mm and a wall thickness of 8mm into the soil until the pile tip penetrates all backfill layers and enters the lower old soil layer ≥1.5 meters. After the pile is formed, M30 micro-expansion cement grout is pumped through the pile pipe. The grouting pressure is maintained at 0.5~1.0MPa until the grout returns steadily at the borehole opening, forming a cement-soil composite pile.

[0048] Mechanical locking: Before the initial setting of the concrete at the top of the pile, the Φ18 or larger steel bar head (about 300mm in length) pre-welded to the top of the steel pipe is vertically passed through the mesh of the first layer of geogrid, then bent at 90° and pressed tightly on the grid, or a special clamp is used to lock the steel bar head to the grid ribs, thus completing the rigid spatial connection between the vertical pile and the horizontal grid.

[0049] By actively constructing a spatial load-bearing skeleton and mechanically locking it, the micropiles and grid are forced to form a rigid connection node, making the horizontal grid a "pillar platform" on the pile top, which can effectively spread the load; the vertical piles become "anchors" of the grid, restricting its displacement, which greatly improves the integrity and deformation resistance of the composite foundation.

[0050] S3. Intelligent layered backfilling and compaction based on feedback control:

[0051] S31. Backfill material preparation: Well-graded gravelly soil is selected, and 2% PO 42.5 cement is added at the mixing plant. Rapid compaction tests are conducted daily on the incoming raw materials, and the results are measured... Substitute into the formula Dynamically calculate the optimal construction moisture content for the day. And guide the amount of water used for mixing, among which, For the optimal moisture content, The standard maximum dry density, This represents the actual maximum dry density of the material on site.

[0052] S32. Spreading and pre-embedding: Using dump trucks and bulldozers, the improved soil material is spread and leveled at a loose thickness of 400mm per layer. Then, workers dig shallow trenches on the leveled soil surface according to a 20m×20m grid and bury wireless temperature and humidity sensors and soil pressure boxes. The top of the sensors is flush with the soil surface, and the cables are connected to the data collection box at the edge.

[0053] S33, Closed-loop intelligent compaction:

[0054] Start the vibratory roller equipped with a continuous compaction control system (CMC);

[0055] CMC collects vibration signals in real time through acceleration sensors on the steel drum of the road roller, and then uses the formula... Calculate the compaction value at the current location. (in The root mean square of the acceleration, (for setting the frequency) The value is displayed in real time on the tablet in the driver's cab;

[0056] The system intelligently determines: when the road roller continuously compacts the same location, its If the rate of change of the value is less than 5% within 3 seconds, the system determines that the point has been "compacted and stabilized" and issues a prompt sound.

[0057] After the entire layer has been compacted, the system automatically integrates all the compacted materials from all locations. The value generates a color "compaction field cloud map" (red represents...). Low value / soft, green indicates (High value / Dense);

[0058] S34. Dynamic Reinforcement: Construction technicians review the cloud map and delineate all... For "weak areas" (typically indicated by yellow or red patches) where the value is 85% below the overall average, instruct the roller to return to these specific areas and perform an additional 2-3 passes of static or low-vibration compaction until a retest is conducted. The value meets the standard, and the color of the cloud map patch is updated to green;

[0059] S35. Network Deepening: Repeat steps S32-S34 to carry out the next layer of backfill. When the cumulative backfill thickness reaches 1.2 meters (e.g., the 4th layer, the 8th layer, etc.), after the current layer is compacted and sensors are installed, and before the next layer is laid, lay a new layer of bidirectional geogrid. The new geogrid is tied to the geogrid already laid in the lower layer with connectors at the overlap, and tied or welded to all micropiles (or reserved connecting bars) that cross this elevation, so that the three-dimensional reinforced network can be continued and densified vertically.

[0060] By establishing a quality closed loop through "real-time CMC perception – cloud map visualization diagnosis – dynamic targeted reinforcement," blind spots in quality are eliminated, and the cloud map exposes any weak points. On-demand compaction is achieved, avoiding over- or under-compaction and generating high-value data. The data field is a digital archive of the foundation stiffness across the entire site, fundamentally contributing to the overall design: it ensures that the "skeleton" of the S2 structure is encased in uniform, high-strength "muscle" (soil), allowing the mechanical properties of the "skeleton" to be fully utilized. Simultaneously, a massive amount of real-world data... Value data is the core input data source for subsequent intelligent monitoring systems to make accurate settlement predictions.

[0061] S4. Curing and Top Interface Layer Construction:

[0062] After backfilling to the design elevation and completing the compaction of the top layer, lay a layer of steel-plastic composite geogrid with an ultimate tensile strength ≥100kN / m;

[0063] Using a spraying device, the soil stabilizer is diluted at a ratio of 1:5 (by volume) and then sprayed evenly onto the top layer of grid and the surface of the soil. The amount of spraying should be such that the surface is moistened but does not form runoff.

[0064] Immediately cover with geotextile and regularly spray water to keep it moist for at least 7 days to promote the hardening of the surface soil into a hard shell.

[0065] After curing, remove the geotextile, tie the concrete cushion reinforcement mesh (if necessary), pour 150mm thick C20 concrete, and before the concrete sets, accurately insert stainless steel settlement monitoring markers with forced centering plates at the designed locations, with the top of the markers flush with the concrete surface.

[0066] By using "reinforcement-chemical curing," a seepage-proof, crack-proof, and erosion-resistant "protective shell" is formed on the top layer, effectively isolating external water damage and mechanical damage. The poured concrete subbase provides a smooth and durable final surface layer. Standardized monitoring points are implanted to establish a permanent and reliable physical benchmark for long-term health monitoring. Its contribution to the overall solution lies in: consolidating and encapsulating the high-quality foundation achievements created by S1-S3, and seamlessly connecting it with the monitoring system, thus achieving a smooth transition between the construction and operation and maintenance periods.

[0067] A construction monitoring system for preventing ground settlement in backfilled areas includes:

[0068] Multi-source sensor networks:

[0069] Vertical settlement chain: a rigid PVC pipe pre-embedded in the early stage of construction, with a magnetic ring settlement probe installed every 2 meters in height inside the pipe. As the backfill rises, the magnetic ring settlement probe settles synchronously with the soil. The ground receiver measures the stratified compression of soil layers at different depths by detecting the position changes of the magnetic ring settlement probe.

[0070] Fiber Bragg grating strain sensor: In the factory or on site, optical fibers are attached to specific ribs of each roll of geogrid in a zigzag pattern. After backfilling, the change in grating wavelength is monitored by a demodulator, and the tensile strain of the geogrid is calculated in real time, thereby knowing the stress state of the three-dimensional network.

[0071] Pore ​​water pressure gauges are installed in boreholes drilled in low-lying areas or at key sections of groundwater flow to monitor changes in pore water pressure caused by fluctuations in groundwater levels over a long period.

[0072] Data Acquisition and Transmission: All sensor data is collected by each acquisition terminal via LoRa wireless network to the site-side gateway, and then transmitted in real time to the cloud server via 4G / 5G network. LoRa technology has the characteristics of long distance, low power consumption and strong penetration, making it very suitable for construction sites without mains power and with many obstructions.

[0073] Intelligent Analysis and Early Warning Platform:

[0074] Digital twin model construction: The platform imports real-time data from surveying, design, and construction, automatically generates a 1:1 three-dimensional finite element model corresponding to the physical site, and can be dynamically updated as construction progresses;

[0075] Settlement prediction unit: This unit runs the core algorithm. ,in The predicted settlement at time t. For the initial instantaneous settlement, , The soil rheological parameters are dynamically updated through machine learning and are not fixed values. Every 6 hours, the platform automatically uses actual settlement monitoring data from a past period (e.g., 30 days) as a training set, and uses machine learning algorithms (such as gradient descent) to back-optimize and solve for the best-fitting parameters. , This value enables the model to adapt to the actual creep characteristics of the soil. For the monitored changes in additional stress, For the compaction value The model can output settlement curves and contour maps of points for the next 7 days based on the calculated soil compression modulus.

[0076] The early warning unit continuously compares the prediction results with the threshold (e.g., 10mm). Once the limit is exceeded, or the sensor data (e.g., a sudden 20% increase in water pressure) is abnormal, an early warning is triggered. The early warning information is pushed according to the graded standards: Level 1 (blue) to the project group chat, Level 2 (yellow) to the technical person in charge via SMS, and Level 3 (red) to the owner and supervisor's APP with handling suggestions.

[0077] Active intervention execution terminal:

[0078] The proactive intervention execution terminal is activated simultaneously. For example, in response to a settlement warning, the system immediately calls the algorithm to automatically generate a precise grouting reinforcement map (PDF or BIM model) based on the predicted settlement basin morphology. The map clearly marks the area to be reinforced, the three-dimensional coordinates of each grouting hole, the recommended depth and grouting volume (e.g., 1.5m³). This map can be directly sent to the tablet computer of the grouting team to guide them in precise construction at night.

[0079] This monitoring system achieves a leap from "diagnosis to prognosis to treatment." Its core advantages are: comprehensive perception dimensions, from stress and deformation to environmental factors; strong predictive capabilities, achieving early warning through a self-learning model; and rapid closed-loop execution, with early warning directly driving intervention measures. Its contribution to the overall solution is like the finishing touch: it transforms a static, completed "high-quality project" into a dynamic, self-sensing, self-early warning, and "self-healing" intelligent infrastructure that can be implemented with the help of external execution, extending settlement control from the construction stage to the entire life cycle.

[0080] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A construction method for preventing floor settlement in a backfill area, characterized by, The method comprises the following steps executed in sequence: S1, foundation grid surveying and directional preprocessing: scanning the site with a geological radar in a preset grid, and performing a light dynamic sounding test at the grid nodes to generate a foundation bearing capacity cloud chart; for areas with a bearing capacity characteristic value lower than 80 kPa or with hidden cavities, accurately excavating 0.5 m below the old soil layer, replacing the soil with graded gravel, and compacting the graded gravel in layers to a compaction coefficient not lower than 0.97; S2, three-dimensional active reinforcement network construction: laying a first layer of bidirectional geogrid on the processed foundation; constructing vertical micropiles at the design positions, the micropiles penetrating the backfill layer and entering the old soil layer by not less than 1.5 m, the micropile body being filled with micro-expanding cement slurry, and the micropile top being reserved with a steel head to be mechanically locked with the geogrid; S3, intelligent layered backfilling and compaction based on feedback control: S31, backfill material preparation: using cement modified soil material, and dynamically determining the construction moisture content thereof; S32, layered paving and sensor pre-embedding: paving the backfill material in layers, and synchronously embedding the monitoring sensors; S33, closed-loop intelligent compaction: using a device equipped with a continuous compaction control system to perform rolling, and generating a compaction degree field cloud chart in real time; S34, dynamic reinforcement of weak areas: according to the compaction degree field cloud chart, performing point reinforcement on the weak areas; S35, network deepening: repeating steps S32 to S34 until the backfilling reaches the design elevation, and laying new geogrids at preset intervals, and binding the new geogrids with the lower geogrids and micropiles through connectors to form a three-dimensional active reinforcement network composed of multiple horizontal reinforcement layers and vertical micropiles; S4, curing and maintenance and top interface layer construction: laying a steel-plastic composite geogrid on the topmost layer, spraying a diluted soil stabilizer, and covering and maintaining for 7 days; after the maintenance, constructing a 150 mm thick C20 concrete cushion, and implanting a floor settlement monitoring marker before the concrete initial setting.

2. The construction method for preventing floor settlement in a backfill area according to claim 1, characterized by, In step S2, the bidirectional geogrid has a nominal tensile strength not lower than 80 kN / m in the longitudinal and transverse directions, and a special plastic steel fastener is used for interlaced connection with a lap width not less than 200 mm.

3. The construction method for preventing floor settlement in a backfill area according to claim 1, characterized by, In step S2, the micropile uses a steel pipe with an outer diameter of 168 mm and a wall thickness of 8 mm, fills the steel pipe with M30 micro-expanding cement slurry, and the grouting pressure is 0.5-1.0 MPa.

4. The construction method for preventing floor settlement in a backfill area according to claim 1, characterized by, The continuous compaction control system in step S33 collects data through an acceleration sensor on the roller, and the compaction value The calculation formula is: Wherein is the root mean square value of the vibration wheel acceleration, is the vibration frequency.

5. The construction monitoring system for preventing floor settlement in a backfill area according to any one of claims 1 to 4, characterized in that, The method comprises: a multi-source sensor network, comprising: a vertical settlement monitoring pipe pre-embedded in the backfill body, a magnetic ring type settlement meter probe being arranged in the pipe at every 2 m height; a fiber bragg grating strain sensor pasted on each layer of geogrid rib; and a pore water pressure gauge arranged in the site; a data acquisition and transmission module, which uses a LoRa wireless network to transmit data of the multi-source sensor network to the cloud in real time; an intelligent analysis and early warning platform, comprising: a digital twin model construction unit for establishing a three-dimensional finite element model of the site according to construction parameters; The settlement prediction unit incorporates a machine learning-based settlement prediction model, which is as follows: ,in The predicted settlement at time t. For the initial instantaneous settlement, , These are soil rheological parameters that are dynamically updated through machine learning. This is an empirical coefficient. For the monitored changes in additional stress, For the compaction value The converted soil compression modulus; this unit trains the model every 6 hours using the latest data to achieve rolling prediction of the settlement trend for the next 7 days; a warning issuing unit for triggering a three-level warning and pushing to relevant personnel when the predicted settlement amount exceeds 10 mm or the sensor data is abnormal. The active intervention execution terminal receives instructions of the early warning issuing unit and executes corresponding intervention plans, including generating a precise grouting reinforcement map and guiding grouting equipment operation, automatically starting and stopping preset dewatering well points, or sending load scheduling suggestions to an operation management party.

6. The construction monitoring system of claim 5, wherein, The soil compression modulus The conversion relationship with the compaction value Wherein And The material correlation coefficient determined by the field calibration test.​ 7. The construction monitoring system of claim 5, wherein, In the three-level early warning, the first-level early warning is blue and is pushed to a project group; the second-level early warning is yellow and increases short message notification of a technical person in charge; and the third-level early warning is red and automatically pushes an alarm and a disposal suggestion to a supervision and an owner APP.

8. The construction monitoring system of claim 5, wherein, The precise grouting reinforcement map contains grouting point three-dimensional coordinates, grouting depth, and grouting quantity information, which are automatically generated by the settlement prediction unit according to a predicted settlement distribution.

Images