Automatic foundation load detection system integrating intelligent sensing and remote cooperation

Abstract

The invention relates to the technical field of foundation detection, and discloses an automatic foundation load detection system integrating intelligent sensing and remote collaboration, which comprises a field sensing execution layer, an edge computing control layer and a cloud remote collaboration layer. The field sensing execution layer is responsible for executing physical loading and collecting environment parameters; the edge computing control layer serves as a local processing center, forms closed-loop control independent of a remote network according to local logic, and uploads data in real time; the cloud remote collaboration layer is used for visually monitoring and issuing an intervention instruction, and the edge computing control layer is configured to preferentially execute the intervention instruction. According to the invention, through a layered architecture and a dual-mode communication mechanism, in combination with dynamic attitude self-stabilization and intelligent data analysis functions, cooperation of local millisecond-level response and remote safety supervision is realized, the problems of equipment stability and foundation damage early warning under an unattended working condition are effectively solved, and the detection efficiency and safety are improved.

Description

Technical Field

[0001] This invention relates to the field of foundation testing technology, specifically to an automated foundation load testing system that integrates intelligent sensing and remote collaboration. Background Technology

[0002] Currently, the assessment of foundation bearing capacity in building engineering relies heavily on static load tests, a crucial step in ensuring the safety of the superstructure. As infrastructure construction extends into remote mountainous areas and regions with complex geological conditions, higher demands are being placed on the real-time nature of testing data and the degree of automation in on-site operations. The industry is gradually promoting the use of automated devices to replace traditional manual meter reading and pressurization methods, aiming to reduce the labor intensity of on-site personnel and minimize human error in recording data.

[0003] For existing automated testing applications, a centralized control architecture based on microcontrollers or PLCs is typically employed. The system collects analog quantities from the field via pressure transmitters and displacement gauges, and aggregates the data to the control center via wired or digital transmission modules. The control logic, based on preset national standards, adjusts the output power of the hydraulic pump station in stages to maintain the load. Technicians primarily view real-time curves through remote host computer software and confirm the entry into the next loading stage after reaching a preset stability standard, or control the unloading process at the end of the test.

[0004] However, such technologies have limitations in practical applications. Architectures relying solely on remote commands are highly sensitive to network conditions. If communication in remote locations fluctuates or is interrupted, cloud commands cannot be issued in a timely manner, causing the hydraulic system to continuously pressurize uncontrollably, leading to safety hazards. Furthermore, existing loading devices are mostly rigid support structures, lacking the physical adaptive adjustment capability to cope with uneven settlement of the foundation soil. Eccentricity of the load can easily cause equipment to overturn. Traditional control logic is mostly based on current numerical judgments, lacking trend prediction of the foundation's ultimate failure state. When the foundation suddenly experiences plastic failure, the system often cannot make a millisecond-level unloading response, easily causing equipment damage.

[0005] Therefore, the present invention provides an automated foundation load detection system that integrates intelligent sensing and remote collaboration to address the shortcomings of the prior art. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides an automated foundation load detection system that integrates intelligent sensing and remote collaboration. This system solves the problems of safety lapses caused by excessive reliance on remote networks in existing automated foundation detection technologies, equipment overturning risks due to uneven settlement, and the lack of predictive early warning of ultimate foundation failure states.

[0007] To achieve the above objectives, the present invention provides the following technical solution: an automated foundation load detection system integrating intelligent sensing and remote collaboration, comprising: an on-site sensing and execution layer, an edge computing control layer, and a cloud-based remote collaboration layer;

[0008] The field perception execution layer is deployed at the foundation detection site and includes a loading component and a sensing component. The loading component is used to apply a vertical load to the foundation, and the sensing component is used to collect physical environment parameters and convert the collected analog signals into digital signals, which are then sent to the edge computing control layer through a first communication protocol. The field perception execution layer is configured to receive and execute drive instructions from the edge computing control layer.

[0009] The edge computing control layer serves as a local processing center, including a local AI control terminal and a wireless communication gateway. The local AI control terminal establishes a communication connection with the field perception execution layer through the first communication protocol, and is used to run the loading control algorithm and send control commands directly to the loading component according to the local preset program logic to form a local closed-loop control. At the same time, it encapsulates the processed data into data packets and uploads them to the cloud remote collaboration layer in real time through the second communication protocol, and receives the commands issued by the cloud remote collaboration layer.

[0010] The cloud-based remote collaboration layer includes a cloud platform server and a remote client. The cloud-based remote collaboration layer establishes a bidirectional data channel with the edge computing control layer through the second communication protocol. The cloud platform server is used to parse the data packets and display real-time monitoring images and data curves through the remote client. The remote client is used to issue intervention commands. The intervention commands are sent to the edge computing control layer via the cloud platform server. The edge computing control layer is configured to prioritize the execution of the intervention commands, thereby changing the operating state of the field perception execution layer.

[0011] By adopting the above technical solution, and utilizing a layered architecture combining a field perception execution layer, an edge computing control layer, and a cloud-based remote collaboration layer, along with a dual-mode communication mechanism of local closed-loop control and remote command intervention, the local AI control terminal can achieve millisecond-level loading control response under the support of the first communication protocol, ensuring the continuity and stability of the loading process without relying on the connection status of the remote network. Simultaneously, the cloud-based remote collaboration layer achieves centralized data visualization and remote monitoring through the second communication protocol, and the edge computing control layer is configured to prioritize the execution of remote intervention commands. Therefore, a detection system is obtained that possesses both independent on-site automated operation capabilities and remote real-time secure takeover capabilities, solving the problems of high risks associated with manual operation in traditional foundation detection and network latency security vulnerabilities in pure remote control, thus improving detection efficiency and safety.

[0012] Preferably, the edge computing control layer further includes a servo driver, and the automated foundation load detection system integrating intelligent sensing and remote collaboration includes an automated closed-loop loading control subsystem; the automated closed-loop loading control subsystem consists of the local AI control terminal, the servo driver, the execution unit, and the wireless pressure sensor; the local AI control terminal is used to compare the received actual loading pressure value with the currently set target pressure value in each control cycle, calculate the deviation, and process the deviation using a proportional-integral-derivative (PID) control algorithm to calculate the control quantity; the PID control algorithm is configured to calculate the product of the instantaneous magnitude of the deviation and the proportional coefficient, the product of the cumulative amount of the deviation over time and the integral coefficient, and the product of the rate of change of the deviation and the derivative coefficient, and the sum of the three is used as the control quantity; the local AI control terminal is used to send the control quantity to the servo driver, and the servo driver is used to adjust the output speed and torque of the servo motor in the execution unit according to the control quantity.

[0013] By adopting the above technical solution and using the proportional-integral-derivative control algorithm to comprehensively calculate the instantaneous value, cumulative value and rate of change of the deviation, the precise adjustment of the servo motor speed and torque is realized, and the pressure fluctuation caused by foundation settlement or mechanical clearance is automatically compensated, ensuring that the actual loading pressure strictly follows the preset loading curve, thereby improving the accuracy and stability of load control.

[0014] Preferably, the on-site perception execution layer further includes an automated centering adjustment mechanism, and the automated foundation load detection system integrating intelligent sensing and remote collaboration includes a dynamic attitude self-stabilization subsystem based on multi-dimensional perception; the dynamic attitude self-stabilization subsystem based on multi-dimensional perception includes an attitude calculation module, an anomaly determination module, and a correction control execution module integrated in the local AI control terminal; the attitude calculation module is used to receive the X-axis direction component data and Y-axis direction component data output by the triaxial tilt sensor, and the settlement value output by the wireless capacitive grating displacement sensor; the anomaly determination module is used to generate an attitude anomaly signal based on the calculation result of the attitude calculation module; the correction control execution module is used to respond to the attitude anomaly signal and determine the tilt based on the X-axis direction component data and the Y-axis direction component data. The oblique vector direction is used to calculate the correction displacement vector, and the correction displacement vector is converted into a drive pulse and sent to the automated centering adjustment mechanism. The automated centering adjustment mechanism is used to drive the micro-motion slide or multi-point support array to perform displacement adjustment according to the drive pulse, generating a reverse restoring torque. The attitude calculation module is used to calculate the comprehensive tilt and the maximum relative settlement difference. The value of the comprehensive tilt is equal to the square root of the sum of the squares of the X-axis direction component data and the squares of the Y-axis direction component data. The value of the maximum relative settlement difference is equal to the difference between the maximum and minimum values ​​among all valid readings of the wireless capacitive grating displacement sensors. The anomaly determination module is used to generate the attitude anomaly signal when the comprehensive tilt is greater than a preset tilt safety threshold or the maximum relative settlement difference is greater than a preset settlement difference threshold.

[0015] By adopting the above technical solution, the system can calculate the spatial attitude of the loading platform in real time, quantify the risk of overturning by calculating the comprehensive tilt and the maximum relative settlement difference, and automatically drive the automated centering adjustment mechanism to generate a reverse restoring torque when an anomaly is detected, thereby actively correcting the equipment tilt caused by uneven foundation settlement, and realizing the physical self-stabilization and active safety protection of the detection equipment under unattended working conditions.

[0016] Preferably, the local AI control terminal integrates an intelligent data analysis and edge computing module, which includes a data preprocessing unit, a regression analysis unit, and a limit state determination unit. The data preprocessing unit establishes a first-in-first-out queue, uses a moving average filtering algorithm to smooth the raw sampling data uploaded by the wireless pressure sensor and the wireless capacitive displacement sensor, and calculates the arithmetic mean of all data points in the current queue as the effective measurement value. The regression analysis unit receives the preprocessed effective pressure data and effective settlement data, and fits them using a hyperbolic constitutive model. The relationship curve between foundation pressure and settlement; the hyperbolic constitutive model used in the regression analysis unit is configured to describe the foundation settlement as the ratio of the applied load value to the denominator, where the denominator is the sum of the products of the first fitting parameter and the second fitting parameter and the load value; wherein, the first fitting parameter is related to the initial tangent modulus of the foundation soil, and the second fitting parameter is related to the ultimate bearing capacity of the foundation soil; the regression analysis unit is used to transform the hyperbolic constitutive model into a linear form, and to perform regression iteration calculations on multiple sets of load values ​​and load-settlement ratio data using the least squares method to solve for the first fitting parameter and the second fitting parameter.

[0017] By adopting the above technical solution, real-time data cleaning and modeling were achieved at the edge. The nonlinear relationship between pressure and settlement was fitted using a hyperbolic constitutive model, and key foundation parameters were solved by linear transformation and least squares method. This enabled an intelligent analysis function that could predict the mechanical properties of the foundation using only early loading data, reducing the impact of noise interference on data analysis.

[0018] Preferably, the limit state determination unit is used to calculate the settlement change rate, which is defined as the first derivative of the foundation settlement with respect to the load; the value of the settlement change rate is equal to the first fitting parameter divided by the divisor, where the divisor is the square of the sum of the products of the first fitting parameter, the second fitting parameter, and the load value; the limit state determination unit is used to monitor the numerical change of the settlement change rate, and generate a foundation failure early warning signal when the settlement change rate exceeds a preset change rate threshold, or when the difference between the theoretical ultimate bearing capacity calculated according to the second fitting parameter and the current actual applied load is less than a preset safety margin.

[0019] By adopting the above technical solution, the first derivative of settlement with respect to load is calculated based on the fitted parameters. This can accurately capture the trend of the foundation transitioning from elastic deformation to plastic failure, and automatically trigger early warning based on the rate of change threshold and safety margin. This effectively prevents sudden foundation failure caused by overload and provides an intelligent basis for judging the limit state for automated detection.

[0020] Preferably, the cloud platform server includes a communication gateway module, a data parsing and storage module, a visual monitoring interface, and an instruction distribution module; the communication gateway module is used to receive and verify data packets uploaded by the edge computing control layer; the data parsing and storage module is used to extract timestamped physical quantity data from the data packets and write it into a time-series database; the visual monitoring interface is used to display video streams and dynamic data charts drawn based on the time-series database, and provides control instruction input controls; the instruction distribution module is used to encapsulate the pause loading, immediate unloading, parameter correction, or test termination instructions input by the remote client into encrypted messages and send them to the edge computing control layer; the edge computing control layer is configured to interrupt the currently executing automated loading loop program when it receives the immediate unloading or test termination instruction.

[0021] By adopting the above technical solutions, a full-process data traceability system based on time-series database and a remote emergency control mechanism based on command priority were constructed. This ensures that remote technicians can handle on-site emergencies with the highest authority based on real-time video and data charts, thus guaranteeing the controllability and safety of the testing process.

[0022] This invention provides an automated foundation load detection system integrating intelligent sensing and remote collaboration. It offers the following advantages:

[0023] 1. This invention employs a layered architecture combining an edge computing control layer and a cloud-based remote collaboration layer to construct a dual-mode communication mechanism that integrates local real-time closed-loop control with remote asynchronous monitoring. The local AI control terminal independently executes servo loading logic based on sensor feedback, ensuring millisecond-level response and continuity in the loading process, unaffected by remote network fluctuations. Simultaneously, the cloud platform possesses the highest-priority command issuance authority, allowing remote technicians to intervene forcefully in emergencies. This effectively solves the problem of balancing real-time control and safety monitoring in traditional automated detection, improving the system's reliability under complex operating conditions.

[0024] 2. This invention integrates a dynamic attitude self-stabilization subsystem based on multi-dimensional perception, utilizing a three-axis tilt sensor and distributed displacement sensors to calculate the spatial attitude of the bearing plate in real time. The system can automatically identify overturning risks based on the comprehensive tilt angle and the maximum relative settlement difference, and drive an automated centering adjustment mechanism to generate a reverse restoring torque to correct the attitude. This proactively counteracts the lateral torque caused by uneven foundation settlement under unattended operating conditions, ensuring the physical stability of the detection device and preventing equipment overturning accidents.

[0025] 3. This invention utilizes an embedded intelligent data analysis module at the edge, employing a hyperbolic constitutive model regression algorithm and limit state determination logic to process collected data in real time. The system can calculate the settlement change rate based on fitted parameters, i.e., the first derivative of settlement with respect to load, thereby keenly capturing the characteristics of the foundation transitioning from elastic deformation to plastic failure. This enables dynamic prediction of the foundation's ultimate bearing capacity and intelligent early warning of potential failure precursors, avoiding sudden foundation failure caused by blind loading and providing a more forward-looking safety guarantee for engineering monitoring. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the overall system architecture of the present invention;

[0027] Figure 2 This is a signal flow diagram of the automated closed-loop loading control subsystem of the present invention;

[0028] Figure 3 This is a flowchart of the dynamic attitude self-stabilization subsystem based on multi-dimensional perception of the present invention.

[0029] Figure 4 This is a flowchart of the data processing of the intelligent data analysis and edge computing module of the present invention;

[0030] Figure 5 This is an interactive flowchart of the remote collaborative management and control platform of the present invention. Detailed Implementation

[0031] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] See attached document Figure 1 This invention provides an automated foundation load detection system integrating intelligent sensing and remote collaboration. The system mainly consists of a field sensing and execution layer, an edge computing control layer, and a cloud-based remote collaboration layer. These three layers are connected through a hierarchical communication network, forming an Internet of Things (IoT) architecture with distributed sensing and centralized management characteristics.

[0033] The on-site sensing and execution layer is deployed at the physical site of the foundation testing to perform physical loading actions and collect physical environmental parameters. This layer mainly includes loading components and sensing components. The loading components contain execution units integrating electric jacks and pressure servo systems for applying vertical loads to the foundation. The sensing components include several distributed wireless capacitive displacement sensors, wireless pressure sensors, and triaxial tilt sensors. The wireless pressure sensors are installed at the output end of the loading components or at the hydraulic chamber to detect the actual loading pressure value. The wireless capacitive displacement sensors are symmetrically arranged around the center of the loading components around the bearing plate to detect the foundation settlement. The triaxial tilt sensors are fixed to the geometric center or reference surface of the bearing plate to detect the tilt angle and attitude changes of the loading plane.

[0034] The edge computing control layer, serving as the system's local processing center, primarily comprises a local AI control terminal and a wireless communication gateway. The local AI control terminal has a built-in microprocessor and storage unit for running and loading control and attitude correction algorithms. The edge computing control layer establishes communication connections with various sensors and execution units in the field perception and execution layer via a first communication protocol. This first communication protocol employs low-power wide-area network (LPWAN) technology or industrial-grade short-range wireless communication technologies, such as LoRa, ZigBee, or industrial Wi-Fi. At this layer, each sensor acts as a data acquisition node, converting the acquired analog signals into digital signals and transmitting them to the local AI control terminal via a wireless channel at a preset sampling frequency. The local AI control terminal performs timestamp alignment and preliminary filtering on the received raw data.

[0035] The cloud-based remote collaboration layer, located in the remote management center, primarily comprises a cloud platform server and remote clients. This layer establishes a bidirectional data channel with the edge computing control layer via a second communication protocol. This second communication protocol utilizes mobile cellular network communication technology, such as 4G or 5G. The cloud platform server is equipped with a database and data analysis engine to store and visualize the entire process of monitoring data. The remote clients are used to display real-time monitoring images and data curves to technicians and provide a remote command input interface.

[0036] This embodiment employs a dual-mode communication architecture combining local real-time control and remote asynchronous monitoring. During normal loading, the edge computing control layer, based on locally preset program logic and feedback data from wireless pressure sensors and wireless capacitive displacement sensors, directly sends servo control commands to the electric jack, forming a local closed-loop control. This local closed-loop control does not depend on the network connection status of the second communication protocol, ensuring the continuity of the loading process.

[0037] Meanwhile, the edge computing control layer uses message queue telemetry transmission protocols or hypertext transfer protocols to encapsulate processed status data, sensor values, and on-site image data into data packets, which are then uploaded in real time to the cloud platform server of the remote collaboration layer. The cloud platform server parses the uploaded data packets and updates the display interface of the remote client. When the remote client issues an intervention command, the command is sent to the edge computing control layer via a second communication protocol. The edge computing control layer parses the command and executes it first, thereby changing the operating state of the on-site sensing execution layer. This architecture ensures both the millisecond-level response capability of on-site equipment to changes in pressure and displacement and enables remote personnel to monitor the detection process in real time and intervene as needed.

[0038] See attached document Figure 2 The automated closed-loop loading control subsystem provided by this invention mainly includes a local AI control terminal, a servo driver, a loading execution unit, and a pressure detection unit. The local AI control terminal, as the control core, is responsible for running the loading program and generating control commands. The servo driver is connected between the local AI control terminal and the loading execution unit, used to convert digital control signals into electrical power signals. The loading execution unit includes a servo motor and a mechanical transmission mechanism, used to convert electrical energy into mechanical thrust. The pressure detection unit, namely the aforementioned wireless pressure sensor, is used to monitor the actual pressure at the loading interface in real time.

[0039] During automated loading, the system executes electric servo drive logic. The local AI control terminal reads the target load value for the current stage according to preset detection standards. The servo driver receives instructions from the local AI control terminal and controls the output speed and torque of the servo motor. The rotational motion of the servo motor is converted into vertical linear motion through a mechanical transmission mechanism, thereby pushing the bearing plate to apply pressure to the foundation soil. When the servo motor rotates forward, the output pressure of the loading execution unit increases. When the servo motor rotates in reverse or stops, the output pressure decreases or remains constant. The servo driver precisely controls the amplitude and speed of the loading execution unit's movement by adjusting the magnitude and phase of the current supplied to the motor.

[0040] To ensure that the load applied to the foundation strictly follows the preset curve, the system operates a pressure closed-loop control algorithm. The pressure detection unit continuously collects the current actual pressure value at a fixed sampling period and feeds this data back to the local AI control terminal in real time via wireless signal. Within each control cycle, the local AI control terminal compares the received actual pressure value with the currently set target pressure value and calculates the deviation between the two. This deviation reflects the gap between the current system output and the desired target.

[0041] The local AI control terminal uses a proportional-integral-derivative (PID) control algorithm to process the aforementioned deviations and calculate the control input for adjusting the servo motor. This control algorithm performs comprehensive calculations based on the instantaneous magnitude, cumulative total, and trend of the deviations, thereby generating an adjustment signal that can quickly eliminate deviations and maintain system stability.

[0042] The mathematical model of the proportional-integral-derivative (PID) control algorithm is expressed as follows:

[0043]

[0044] Where u(t) represents the control quantity output to the servo driver at time t, which corresponds to adjusting the voltage value or pulse frequency of the motor; e(t) represents the deviation obtained by subtracting the actual pressure value from the target pressure value at time t; K p K represents the proportionality coefficient, used to make proportional adjustments based on the magnitude of the deviation; i K represents the integral coefficient, used to eliminate steady-state error based on the cumulative amount of deviation over time; d τ represents the differential coefficient, used to predict trends and suppress overshoot based on the rate of change of the deviation; τ represents the time variable in the integration process.

[0045] The local AI control terminal sends the calculated control quantity u(t) to the servo driver, thereby adjusting the servo motor's operating state in real time. If the actual pressure is less than the target pressure, the control quantity u(t) drives the motor to increase its output torque. If the actual pressure is greater than the target pressure, the control quantity u(t) drives the motor to decrease its output torque or slightly retract it. Through this continuous closed-loop feedback adjustment, the system can automatically compensate for pressure fluctuations caused by foundation settlement or mechanical clearances, ensuring that the actual loaded pressure remains stable within the allowable error range of the target value.

[0046] See attached document Figure 3 The dynamic attitude self-stabilization subsystem based on multi-dimensional perception provided by this invention mainly includes an attitude calculation module, an anomaly detection module, and a correction control execution module. These modules are integrated into a local AI control terminal and are electrically connected to a three-axis tilt sensor, a wireless capacitive displacement sensor, and an automated centering adjustment mechanism distributed on-site.

[0047] During system operation, the attitude calculation module receives multidimensional data streams from field sensors in real time to construct a virtual spatial attitude model of the support plate. This module first reads the X-axis and Y-axis component data output from the three-axis tilt sensor. These two components represent the longitudinal and lateral tilt degrees of the support plate relative to the horizontal reference plane, respectively. To quantify the overall tilt risk, the attitude calculation module calculates the comprehensive tilt degree based on these two orthogonal components.

[0048] The calculation logic for the overall tilt follows the following mathematical relationship:

[0049]

[0050] Where Φ represents the current overall tilt of the bearing plate, which reflects the total deviation of the loading axis from the vertical line of gravity; θ x This represents the tilt angle value along the X-axis measured by the triaxial tilt sensor; θ y This represents the tilt angle value along the Y-axis measured by the triaxial tilt sensor.

[0051] Simultaneously, the attitude calculation module reads the settlement values ​​from multiple wireless capacitive displacement sensors distributed around the bearing plate to monitor the uniformity of foundation settlement. To prevent misjudgments caused by local sensor failures or sudden changes at a single point, the system uses the range method to calculate the maximum relative settlement difference of the current loading plane.

[0052] The calculation logic for the maximum relative settlement difference follows the following mathematical relationship:

[0053] Δs=max(S i )-min(S i );

[0054] Where Δs represents the maximum relative settlement difference at the edge of the bearing plate at the current moment; S i This represents the settlement measured by the i-th wireless capacitive displacement sensor, where i is the sensor index, ranging from 1 to the total number of sensors; max(S i ) represents the maximum value among all valid sensor readings; min(S) i ) represents the minimum value among all valid sensor readings.

[0055] The anomaly detection module compares the calculated overall tilt angle Φ with a preset tilt safety threshold, and simultaneously compares the maximum relative settlement difference Δs with a preset settlement difference threshold. When Φ exceeds the tilt safety threshold or Δs exceeds the settlement difference threshold, the anomaly detection module generates an attitude anomaly signal. This signal triggers the system to enter active safety protection mode.

[0056] Upon receiving an attitude anomaly signal, the deviation correction control execution module immediately takes over control of the system. This module first sends a pause or pressure holding command to the automated closed-loop loading control subsystem to prevent the risk of overturning from escalating further under continuous loading. Subsequently, the deviation correction control execution module adjusts the current θ... x and θ yThe direction of the tilt vector is determined, and the corrective displacement vector used to generate the reverse restoring torque is calculated. The corrective control execution module converts the calculated corrective displacement vector into motor drive pulses and sends them to the automated centering adjustment mechanism. The automated centering adjustment mechanism drives the bottom micro-motion slide or multi-point support array to make corresponding displacement adjustments according to the instructions. This adjustment action changes the plane coordinates of the jack's force application point relative to the geometric center of the bearing plate, thereby artificially introducing a reverse corrective torque to counteract the lateral overturning torque caused by uneven foundation settlement. The system continuously monitors the change in the overall tilt angle Φ. When Φ falls back into the safe zone, the corrective control execution module unlocks, allowing the system to resume normal vertical loading. This process constitutes an active safety closed loop, ensuring the physical stability of the testing equipment under unattended operation.

[0057] See attached document Figure 4 The intelligent data analysis and edge computing module provided by this invention is integrated into the local AI control terminal of the edge computing control layer. This module mainly includes a data preprocessing unit, a regression analysis unit, and a limit state determination unit. These units work together to transform the discrete physical quantities collected by field sensors into a continuous mathematical model that reflects the mechanical properties of the foundation, and perform automated state assessment accordingly.

[0058] The data preprocessing unit is responsible for cleaning the raw data stream uploaded from the field sensing and execution layer. Due to the presence of engineering machinery vibrations, wind loads, and electromagnetic interference at the test site, the time-series data uploaded by wireless pressure sensors and wireless capacitive displacement sensors often contain high-frequency random noise. The data preprocessing unit uses a moving average filtering algorithm to smooth the raw sampled data. This algorithm establishes a fixed-length first-in-first-out queue in memory. Whenever a new sampling point is received, it is pushed to the tail of the queue, while the oldest sampling point at the head of the queue is removed. The data preprocessing unit calculates the arithmetic mean of all data points in the current queue and outputs this average as the current valid measurement value to subsequent units. This process filters out instantaneous spike interference and preserves the true low-frequency trends of load and settlement changes.

[0059] The regression analysis unit receives preprocessed effective pressure and settlement data and uses an artificial intelligence regression algorithm to fit the pressure-settlement relationship curve (Ps curve) of the foundation in real time. To accurately predict the nonlinear deformation characteristics and ultimate bearing capacity of the foundation, the regression analysis unit employs a hyperbolic constitutive model as the basic fitting function. This model can effectively describe the physical process of the foundation soil transitioning from the elastic deformation stage to the plastic failure stage.

[0060] The regression equation of the hyperbolic constitutive model follows the mathematical relationship:

[0061]

[0062] Where s represents the settlement of the foundation, which is the average value of the readings of multiple wireless capacitive displacement sensors; P represents the load applied to the bearing plate; a represents the fitting parameter related to the initial tangent modulus of the foundation soil, which physically corresponds to the reciprocal of the tangent slope at the origin of the Ps curve; b represents the fitting parameter related to the ultimate bearing capacity of the foundation soil, which physically corresponds to the reciprocal of the asymptote value of the Ps curve.

[0063] To solve for the unknown parameters in the above equation, the regression analysis unit transforms the nonlinear hyperbolic equation into a linear form, namely P / s = a + bP. Subsequently, the system uses the least squares method to perform linear regression iterations on multiple sets of (P, P / s) data pairs that have been collected, thereby obtaining the optimal solutions for parameters a and b with the minimum sum of squared residuals. As the loading stage increases, the regression analysis unit continuously updates the values ​​of parameters a and b, making the fitted Ps curve dynamically approximate the actual foundation response trajectory.

[0064] The limit state determination unit monitors the stability of the foundation based on the real-time updated Ps curve equation. This unit calculates the slope of the tangent line to the curve at the current load level, i.e., the settlement rate, by differentiating the regression equation.

[0065] The calculation logic for the rate of change of settlement follows the following mathematical relationship:

[0066]

[0067] in, This represents the first derivative of foundation settlement with respect to the load, which reflects the instantaneous flexibility of the foundation under the current load.

[0068] Limit State Determination Unit Continuously Monitors The numerical change. When the foundation approaches failure, the settlement increases sharply with increasing load, leading to... The values ​​exhibit an exponential upward trend. The system has a preset threshold for the rate of change; when the calculated... If the threshold is exceeded, or if the difference between the theoretical ultimate bearing capacity 1 / b calculated based on parameter b and the current actual applied load is less than the preset safety margin, the ultimate state determination unit generates a foundation failure early warning signal. This signal will immediately trigger the termination logic of the loading program and mark the relevant data as ultimate bearing capacity characteristic values ​​for final confirmation by technical personnel in the remote collaborative layer. This mechanism enables intelligent analysis functions that can predict the final failure load of the foundation using only early loading data.

[0069] See attached document Figure 5The remote collaborative management and control platform provided by this invention is deployed on a cloud server cluster and provides a visual operation terminal for remote technicians. The platform mainly consists of a communication gateway module, a data parsing and storage module, a visual monitoring interface, and a command distribution module. These modules work together to achieve centralized management and real-time intervention of multiple foundation testing sites distributed in different geographical locations.

[0070] The communication gateway module serves as the data interaction interface between the cloud and the field edge computing control layer, maintaining a persistent network connection with field devices. When the local AI control terminal uploads data packets via the mobile communication network, the communication gateway module is responsible for receiving and verifying the integrity and security of the data packets. The data parsing and storage module unpacks the verified data packets according to a predefined communication protocol. This module extracts key physical quantities with timestamps from the binary data stream, including the current applied pressure value, the settlement displacement value of each measuring point, the tilt angle value of the bearing plate, and the Ps curve fitting parameters calculated by the edge computing module. The extracted structured data is synchronously written to the time-series database in the cloud to ensure the immutability and traceability of the archived data.

[0071] The visual monitoring interface runs on a remote client and is used to display the real-time status of the site to technicians. This interface integrates multiple video stream playback windows and dynamic data chart windows. The video stream playback windows decode and render video data from on-site cameras, displaying the physical attitude of the loading device and the surrounding environment. The dynamic data chart windows subscribe to the latest data in the time-series database, plotting load-time curves, settlement-time curves, and pressure-settlement curves (Ps curves) in real-time. In addition, the interface includes an attitude dashboard that uses X-axis and Y-axis tilt angle data transmitted from the site to visually simulate and display the current planar tilt state of the load-bearing plate. This multi-dimensional information display method allows remote experts to fully grasp every detail of the inspection process.

[0072] The system features an event-driven anomaly alarm push mechanism. When the edge computing control layer detects a foundation damage warning or abnormal equipment posture, it includes a high-priority alarm status code in the uploaded data packet. Once the cloud platform recognizes this status code, it immediately triggers the alarm response process. The visual monitoring interface will forcefully attract the attention of remote operators by popping up dialog boxes, highlighting abnormal value areas, or playing alert sounds. Simultaneously, the system automatically captures and stores video clips from the scene before and after the alarm, assisting technicians in fault review and cause analysis.

[0073] In terms of expert collaboration and emergency intervention, this embodiment establishes a remote command issuance process with access control. The visual monitoring interface provides a set of control command input controls, including pause loading, immediate unloading, parameter correction, and test termination. When a remote expert is authenticated and issues a control command, the command distribution module encapsulates it into an encrypted control message and sends it to the designated field edge computing control layer via the downlink channel.

[0074] To ensure the effectiveness of emergency operations, the system's control logic prioritizes remote commands. Upon receiving a remote command to immediately unload or terminate the test, the local AI control terminal will immediately interrupt the currently executing automated loading loop and prioritize the unloading or shutdown action. This mechanism grants remote technicians the highest authority in emergency situations such as sudden foundation collapse, distorted sensor data, or impending collapse of video display equipment, thus forming the last line of defense for the automated testing system in unattended operation. All remotely issued commands, operation times, and operator IDs are automatically recorded by the system, forming a complete operation log.

Claims

1. An automated foundation load detection system integrating intelligent sensing and remote collaboration, characterized in that, include: The system comprises an on-site perception and execution layer, an edge computing control layer, and a cloud-based remote collaboration layer. The field perception execution layer is deployed at the foundation detection site and includes a loading component and a sensing component. The loading component is used to apply a vertical load to the foundation, and the sensing component is used to collect physical environment parameters and convert the collected analog signals into digital signals, which are then sent to the edge computing control layer through a first communication protocol. The field perception execution layer is configured to receive and execute drive instructions from the edge computing control layer. The edge computing control layer serves as a local processing center, including a local AI control terminal and a wireless communication gateway. The local AI control terminal establishes a communication connection with the field perception execution layer through the first communication protocol, and is used to run the loading control algorithm and send control commands directly to the loading component according to the local preset program logic to form a local closed-loop control. At the same time, it encapsulates the processed data into data packets and uploads them to the cloud remote collaboration layer in real time through the second communication protocol, and receives the commands issued by the cloud remote collaboration layer. The cloud-based remote collaboration layer includes a cloud platform server and a remote client. The cloud-based remote collaboration layer establishes a bidirectional data channel with the edge computing control layer through the second communication protocol. The cloud platform server is used to parse the data packets and display real-time monitoring images and data curves through the remote client. The remote client is used to issue intervention commands. The intervention commands are sent to the edge computing control layer via the cloud platform server. The edge computing control layer is configured to prioritize the execution of the intervention commands, thereby changing the operating state of the field perception execution layer.

2. The automated foundation load detection system integrating intelligent sensing and remote collaboration according to claim 1, characterized in that, The loading component includes an execution unit that integrates an electric jack and a pressure servo system; The sensing components include a wireless pressure sensor, a wireless capacitive displacement sensor, and a triaxial tilt sensor; The wireless pressure sensor is installed at the output end of the loading component or at the hydraulic cavity to detect the actual loading pressure value. The wireless capacitive displacement sensor is symmetrically arranged around the center of the loading component on the periphery of the bearing plate to detect the amount of foundation settlement; The triaxial tilt sensor is fixed to the geometric center or reference surface of the bearing plate and is used to detect the tilt angle and attitude change of the loading plane.

3. The automated foundation load detection system integrating intelligent sensing and remote collaboration according to claim 2, characterized in that, The local AI control terminal has a built-in microprocessor and storage unit; The first communication protocol uses low-power wide-area network technology or industrial-grade short-range wireless communication technology, with the wireless pressure sensor, the wireless capacitive displacement sensor, and the triaxial tilt sensor serving as data acquisition nodes; the second communication protocol uses mobile cellular network communication technology.

4. The automated foundation load detection system integrating intelligent sensing and remote collaboration according to claim 2, characterized in that, The edge computing control layer also includes a servo driver, and the automated foundation load detection system integrating intelligent sensing and remote collaboration includes an automated closed-loop loading control subsystem. The automated closed-loop loading control subsystem consists of the local AI control terminal, the servo driver, the execution unit, and the wireless pressure sensor. The local AI control terminal is used to compare the received actual loading pressure value with the currently set target pressure value in each control cycle, calculate the deviation, and use a proportional-integral-derivative control algorithm to process the deviation and calculate the control quantity. The proportional-integral-derivative (PID) control algorithm is configured to calculate the product of the instantaneous magnitude of the deviation and the proportional coefficient, the product of the cumulative amount of the deviation over time and the integral coefficient, and the product of the rate of change of the deviation and the derivative coefficient, and use the sum of the three as the control quantity. The local AI control terminal is used to send the control quantity to the servo driver, and the servo driver is used to adjust the output speed and torque of the servo motor in the execution unit according to the control quantity.

5. The automated foundation load detection system integrating intelligent sensing and remote collaboration according to claim 2, characterized in that, The field perception execution layer also includes an automated centering adjustment mechanism, and the automated foundation load detection system integrating intelligent sensing and remote collaboration includes a dynamic attitude self-stabilization subsystem based on multi-dimensional perception. The dynamic attitude self-stabilization subsystem based on multi-dimensional perception includes an attitude calculation module, an anomaly detection module, and a deviation correction control execution module integrated into the local AI control terminal. The attitude calculation module is used to receive the X-axis direction component data and Y-axis direction component data output by the triaxial tilt sensor, as well as the settlement value output by the wireless capacitive grating displacement sensor. The anomaly determination module is used to generate an attitude anomaly signal based on the calculation results of the attitude calculation module; The correction control execution module is used to respond to the attitude abnormality signal, determine the tilt vector direction based on the X-axis direction component data and the Y-axis direction component data, calculate the correction displacement vector, and convert the correction displacement vector into a drive pulse to send to the automatic centering adjustment mechanism; The automated centering adjustment mechanism is used to drive the micro-motion slide or multi-point support array to adjust the displacement according to the driving pulse, thereby generating a reverse restoring torque.

6. The automated foundation load detection system integrating intelligent sensing and remote collaboration according to claim 5, characterized in that, The attitude calculation module is used to calculate the overall tilt and the maximum relative settlement difference; The value of the overall tilt is equal to the square root of the sum of the squares of the X-axis component data and the squares of the Y-axis component data. The value of the maximum relative settlement difference is equal to the difference between the maximum and minimum values ​​among all valid readings of the wireless capacitive displacement sensor. The anomaly detection module is used to generate the attitude anomaly signal when the overall tilt is greater than a preset tilt safety threshold or the maximum relative settlement difference is greater than a preset settlement difference threshold.

7. The automated foundation load detection system integrating intelligent sensing and remote collaboration according to claim 2, characterized in that, The local AI control terminal integrates an intelligent data analysis and edge computing module, which includes a data preprocessing unit, a regression analysis unit, and a limit state determination unit. The data preprocessing unit is used to establish a first-in-first-out queue, and uses a moving average filtering algorithm to smooth the raw sampling data uploaded by the wireless pressure sensor and the wireless capacitive displacement sensor, and calculates the arithmetic mean of all data points in the current queue as the effective measurement value. The regression analysis unit is used to receive preprocessed effective pressure data and effective settlement data, and to fit the pressure-settlement relationship curve of the foundation using a hyperbolic constitutive model.

8. The automated foundation load detection system integrating intelligent sensing and remote collaboration according to claim 7, characterized in that, The hyperbolic constitutive model used in the regression analysis unit is configured to describe the settlement of the foundation as the ratio of the applied load value to the denominator, where the denominator is the sum of the products of the first fitting parameter and the second fitting parameter and the load value. The first fitting parameter is related to the initial tangent modulus of the foundation soil, and the second fitting parameter is related to the ultimate bearing capacity of the foundation soil. The regression analysis unit is used to transform the hyperbolic constitutive model into a linear form, and to perform regression iteration calculations on multiple sets of load values ​​and load-settlement ratio data pairs using the least squares method to solve for the first fitting parameter and the second fitting parameter.

9. The automated foundation load detection system integrating intelligent sensing and remote collaboration according to claim 8, characterized in that, The limit state determination unit is used to calculate the settlement change rate, which is defined as the first derivative of the foundation settlement with respect to the load. The value of the settlement change rate is equal to the first fitting parameter divided by the divisor, where the divisor is the square of the sum of the products of the first fitting parameter, the second fitting parameter, and the load value. The limit state determination unit is used to monitor the numerical change of the settlement change rate, and generates a foundation failure early warning signal when the settlement change rate exceeds a preset change rate threshold, or when the difference between the theoretical ultimate bearing capacity calculated according to the second fitting parameter and the current actual applied load is less than a preset safety margin.

10. The automated foundation load detection system integrating intelligent sensing and remote collaboration according to claim 1, characterized in that, The cloud platform server includes a communication gateway module, a data parsing and storage module, a visual monitoring interface, and an instruction distribution module. The communication gateway module is used to receive and verify data packets uploaded by the edge computing control layer; The data parsing and storage module is used to extract timestamped physical quantity data from the data packet and write it into the time series database; The visual monitoring interface is used to display video streams and dynamic data charts drawn based on the time-series database, and provides control command input controls; The instruction distribution module is used to encapsulate the instructions input by the remote client, such as pause loading, unload immediately, correct parameters, or terminate test, into encrypted messages and send them to the edge computing control layer. The edge computing control layer is configured to interrupt the currently executing automated loading loop when it receives the immediate unload or the terminate test instruction.

Images