Self-adaptive synchronous degradation control method and system for bridge jacking

Abstract

The invention relates to the technical field of bridge jacking, and discloses a self-adaptive synchronous degradation control method and system for bridge jacking. A multi-dimensional synchronous monitoring system is initialized, and dynamic reference parameters of all jacking execution units are established. And constructing a dynamic load balancing model, and realizing cooperative control of multiple jacking execution units, including self-adaptive adjustment of a speed compensation coefficient and a pressure threshold range. And when the performance deviation of the jacking execution unit is detected, activating a multi-stage fault-tolerant mechanism to carry out topology reconstruction. And when the system stability is lower than a critical threshold value, starting a grading and degrading control mode, and redistributing a control weight based on the optimal combination of the remaining units. According to the method, the synchronism, the self-adaptability and the robustness in the bridge jacking process are improved, and technical support is provided for bridge jacking engineering.

Description

Technical Field

[0001] The present invention relates to the technical field of bridge jacking, and in particular to an adaptive synchronous degradation control method and system for bridge jacking. Background Art

[0002] Bridge jacking is a complex and technically demanding project, during which the synchronization of each jacking point needs to be ensured to avoid uneven stress on the bridge structure and damage. Traditional bridge jacking control methods often rely on fixed control parameters and strategies, which are difficult to cope with the various uncertainties and dynamic changes that occur in actual projects. For example, the performance of the jacking execution unit may change due to wear, failure or environmental factors, and the delay and packet loss of the communication system may also affect the real-time transmission of instructions. In addition, traditional control methods lack effective degradation control strategies when facing the failure or performance degradation of some execution units, which may lead to the failure of the entire jacking process. Therefore, a method that can adaptively adjust control parameters and strategies, with a multi-level fault-tolerant mechanism and a hierarchical degradation control mode, is needed to improve the synchronization and stability of the bridge jacking process. Summary of the Invention

[0003] In view of the above-mentioned technical deficiencies, the purpose of the present invention is to provide an adaptive synchronous degradation control method and system for bridge jacking, so as to solve the problems in the prior art that fixed control parameters and strategies cannot adapt to dynamic changes and lack effective degradation control strategies.

[0004] In order to solve the above technical problems, the present invention adopts the following technical solutions: In a first aspect, the present invention provides an adaptive synchronous degradation control method for bridge jacking, the method comprising: Step S100: Based on the bridge structure characteristics and jacking target parameters, initialize the multi-dimensional synchronous monitoring system and establish the dynamic benchmark parameters of each jacking execution unit; Step S200: constructing a dynamic load balancing model based on dynamic reference parameters, and implementing coordinated control of multiple lifting execution units through adaptive adjustment of the speed compensation coefficient and the pressure threshold range; Step S300: activating a multi-level fault-tolerance mechanism with topology reconstruction capability when the load balancing model detects that the performance of the jacking execution unit deviates; Step S400: When the system stability output by the load balancing model is lower than a critical threshold, a hierarchical degradation control mode is initiated, and control weights are redistributed based on the optimal combination of the remaining jacking execution units.

[0005] Preferably, in a possible implementation manner of the first aspect, the dynamic reference parameters include a set of jacking point coordinates acquired by a three-dimensional laser scanner and a local stiffness coefficient matrix measured by a fiber optic sensor.

[0006] Preferably, in a possible implementation manner of the first aspect, a regional grouping strategy is adopted to divide the lifting execution units into a master control group and a slave group, wherein the master control group adopts a full closed-loop feedback control, and the slave group adopts a feedforward-feedback composite control mode; The regional grouping strategy adopts an improved K-means clustering algorithm, and its objective function is:

[0007] in Group the lifting execution units into groups. is the total number of clusters, For the A sample set, each cluster corresponds to a lifting execution unit group, , is the weight coefficient, express The set of adjacent samples of is the sample point, are adjacent sample points, For the The clustering center of each cluster, the clustering dimensions include real-time load rate, displacement deviation and adjacent unit coupling coefficient; The comprehensive distance between each unit and the cluster center is calculated, and the master control group is determined based on the comprehensive clustering dimension. The remaining lifting execution units are classified into subordinate groups.

[0008] Preferably, in a possible implementation of the first aspect, the dynamic load balancing model includes: The data fusion layer uses a time series alignment algorithm and a graph-based heterogeneous data fusion method to output a fused feature tensor with spatiotemporal consistency; The state assessment layer uses a hybrid model of convolutional neural networks and long short-term memory networks. It generates a dynamic health index that includes displacement deviation, stress fluctuation, and hydraulic anomalies based on the fused feature tensor through multi-channel feature extraction. It also builds an anomaly detector based on the self-attention mechanism and outputs the performance indicators of the lifting execution unit. The decision optimization layer integrates the multi-objective particle swarm optimization algorithm and the distributed gradient descent strategy to construct a loss function that includes displacement synchronization error, pressure balance, and energy consumption coefficient, and outputs a system stability evaluation index that includes the optimal displacement compensation amount and hydraulic adjustment parameters; The instruction distribution layer builds a priority queue control mechanism to generate a control instruction sequence including jacking speed instructions, oil pressure adjustment amplitude and actuator compensation, and outputs a dynamic adjustment parameter set including instruction priority identification and timing constraints.

[0009] Preferably, in a possible implementation of the first aspect, the coordinated control is specifically: Based on the similarity matching results between the real-time displacement deviation and the historical working conditions, a fuzzy PID controller is used to dynamically adjust the speed compensation coefficient, where the displacement deviation weight factor is adaptively updated as the local stiffness of the bridge changes. The pressure threshold range uses a long short-term memory network to predict the structural stress distribution within the next 3 seconds, and combines the nonlinear characteristic curve of the hydraulic actuator to generate a dynamic pressure safety boundary; In view of the collaborative relationship between the master control group and the slave group, a feedforward-feedback composite control channel is constructed. The full closed-loop feedback signal of the master control group serves as the feedforward input of the slave group. At the same time, the pressure fluctuation data of the slave group reversely corrects the pressure threshold tolerance range of the master control group to achieve global pressure balance.

[0010] Preferably, in a possible implementation of the first aspect, the multi-level fault-tolerant mechanism includes a three-level response strategy: the first level is local PID parameter self-tuning, the second level starts adjacent unit coupling compensation, and the third level performs topology reorganization, wherein the second level compensation amount , express i The set of adjacent cells of a cell, For adjacent units, 、 are the stiffness coefficient and the damping coefficient, is the displacement difference between the adjacent unit and the current unit, is the speed difference between the adjacent unit and the current unit.

[0011] Preferably, in a possible implementation manner of the first aspect, the topology structure reorganization comprises the following steps: Construct a weighted connection graph, vertex weight Reflects unit health, edge weight Indicates communication reliability; The modified Dijkstra algorithm is used to obtain the optimal pressure transmission path, and the path scoring function , where Pressure transmission path, Representative Path The edge passing through Representative Path The vertices passed through; The topology structure is reorganized according to the optimal pressure transmission path.

[0012] Preferably, in a possible implementation manner of the first aspect, the hierarchical degradation control mode includes four levels of standards, corresponding to communication system degradation, execution unit degradation, sensor system degradation and comprehensive degradation mode respectively; The communication system degradation mode is to switch to the redundant wireless mesh network when the backbone network delay exceeds 50ms or the packet loss rate is greater than 5%, and adopt a priority filtering mechanism to ensure command transmission; The actuator unit degradation mode is to automatically isolate the faulty unit and maintain the lifting plane through force coupling compensation of adjacent units when a hydraulic cylinder leakage rate greater than or equal to 0.5L / min or a displacement deviation continuously exceeds the limit for 10 seconds is detected; The sensor system degradation mode is to activate the fusion compensation algorithm of lidar point cloud data and strain gauge redundant measurement values to reconstruct the deformation monitoring system when the number of fiber optic sensor failures exceeds 30% of the total number of clusters; The comprehensive degradation mode is to synchronously trigger a joint degradation strategy of the communication, execution, and sensing systems when the system stability index is lower than a preset minimum threshold.

[0013] Preferably, in a possible implementation manner of the first aspect, the reallocating the control weights includes: Construct an evaluation matrix that includes unit health, load capacity margin, and location strategic value, and calculate the comprehensive performance score of each remaining unit; The Hungarian algorithm is used to solve the optimal task allocation plan, and the lifting execution units with comprehensive efficiency scores exceeding the preset score threshold are allocated to the key support points; A dynamic weight adjustment strategy is introduced to update the control weight distribution ratio in real time based on the Kalman filter.

[0014] In a second aspect, the present invention provides an adaptive synchronous degradation control system for bridge jacking, the system comprising: The multi-dimensional monitoring initialization module initializes the multi-dimensional synchronous monitoring system based on the bridge structure characteristics and jacking target parameters, and establishes the dynamic benchmark parameters of each jacking execution unit; The dynamic load coordination module builds a dynamic load balancing model based on dynamic benchmark parameters, and realizes the coordinated control of multiple lifting execution units through adaptive adjustment of the speed compensation coefficient and pressure threshold range; A fault-tolerant topology reconstruction module activates a multi-level fault-tolerant mechanism with topology reconstruction capabilities when the load balancing model detects performance deviation of the jacking execution unit; The hierarchical weight optimization module starts the hierarchical degradation control mode when the system stability output by the load balancing model is lower than the critical threshold, and redistributes the control weights based on the optimal combination of the remaining jacking execution units.

[0015] The beneficial effects of the present invention are: by initializing a multi-dimensional synchronous monitoring system, dynamic benchmark parameters of each lifting execution unit are established, providing a basis for subsequent collaborative control. The construction of a dynamic load balancing model realizes the collaborative control of multiple lifting execution units, and improves the synchronization of the lifting process through adaptive adjustment of the speed compensation coefficient and the pressure threshold range. The introduction of a multi-level fault-tolerant mechanism enables the topology reconstruction capability to be quickly activated when the performance of the lifting execution unit deviates, thereby ensuring the continuity of the lifting process. The activation of the hierarchical degradation control mode redistributes the control weights based on the optimal combination of the remaining units, ensuring that when the system stability decreases, a high lifting efficiency and safety can still be maintained.

[0016] Overall, the present invention improves the adaptability and robustness of the bridge jacking process, and provides strong technical support for bridge jacking projects. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 A flow chart of an adaptive synchronous degradation control method for bridge jacking is provided for this application.

[0019] Figure 2 A structural diagram of an adaptive synchronous degradation control system for bridge jacking is provided for this application.

[0020] Explanation of the accompanying symbols: 1-multi-dimensional monitoring initialization module, 2-dynamic load coordination module, 3-fault-tolerant topology reconstruction module, 4-hierarchical weight optimization module. DETAILED DESCRIPTION

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0022] Example 1: Figure 1 As shown, the present invention provides an adaptive synchronous degradation control method for bridge jacking, comprising: Step S100: Based on the structural characteristics of the bridge and the jacking target parameters, a multi-dimensional synchronous monitoring system is initialized to establish dynamic benchmark parameters of each jacking execution unit.

[0023] In this embodiment, a 3D laser scanner is used to collect high-precision spatial coordinates of the bridge's jacking points, obtaining a coordinate set of these jacking points. Specifically, the 3D laser scanner scans the bridge's support structure from multiple angles at a preset scanning resolution, generating point cloud data containing the 3D coordinates of the jacking points. A point cloud registration algorithm is used to eliminate differences in scanning angles, and the ICP (Iterative Closest Point) algorithm is used to optimize coordinate accuracy. Ultimately, the geometric center coordinates of each jacking execution unit are extracted as reference positioning parameters. Simultaneously, a distributed fiber optic sensor network is deployed along the bridge's key stress-bearing areas, measuring the strain distribution of the concrete structure around each jacking point in real time at a preset sampling frequency. A local stiffness coefficient matrix is calculated using a finite element inverse analysis algorithm. This matrix characterizes the dynamic stiffness changes of the bridge during the jacking process. Based on this 3D coordinate set and stiffness coefficient matrix, an initial dynamic reference parameter set is constructed.

[0024] The improved K-means clustering algorithm is used to implement the regional grouping strategy. Specifically, each lifting execution unit is used as a sample point, and the clustering dimensions are defined to include real-time load rate, displacement deviation, and adjacent unit coupling coefficient. The real-time load rate is the ratio of the hydraulic system pressure value to the rated load, the displacement deviation is calculated by the absolute difference between the real-time acquisition value of the laser rangefinder and the target displacement, and the adjacent unit coupling coefficient is determined based on a weighted function of the correlation between the spatial distance of the lifting point and the stiffness. The objective function of the improved algorithm is: ,in Group the lifting execution units into groups. is the total number of clusters, For the A sample set, each cluster corresponds to a lifting execution unit group, is the sample point, are adjacent sample points, For the The clustering center of each cluster, the clustering dimension includes real-time load rate, displacement deviation and adjacent unit coupling coefficient, spatial distance weight coefficient Consistency weight coefficient with adjacent units Determined by genetic algorithm optimization, Represents sample points The range of adjacent units is dynamically adjusted by the continuity characteristics of the bridge structure.

[0025] During the clustering process, the data for each dimension is first Z-score standardized, and the comprehensive Euclidean distance between each unit and the cluster center is calculated. The real-time load factor accounts for 40%, the displacement deviation accounts for 35%, and the coupling coefficient between adjacent units accounts for 25%. Through iterative optimization of the cluster center position, the top three units with the smallest comprehensive distance are assigned to the master control group, and the remaining units are assigned to the slave groups. The master control group adopts a fully closed-loop feedback control strategy, whose feedback signal includes the real-time displacement and pressure dual variables. The slave groups form a composite control mode based on the master control group's feedforward instructions and local pressure feedback.

[0026] Step S200: constructing a dynamic load balancing model based on dynamic reference parameters, and realizing coordinated control of multiple lifting execution units through adaptive adjustment of the speed compensation coefficient and the pressure threshold range.

[0027] In this embodiment, the dynamic load balancing model includes a data fusion layer, a state evaluation layer, a decision optimization layer, and an instruction distribution layer.

[0028] The data fusion layer uses a time series alignment algorithm to perform spatiotemporal alignment on the multi-source heterogeneous data, combining the jacking point coordinates acquired by a 3D laser scanner with the local stiffness coefficient matrix measured by fiber optic sensors. By establishing a graph structure model with jacking execution units as nodes and structural coupling relationships as edges, the hydraulic system pressure values, displacement sensor data, and bridge strain distribution data are mapped to a unified spatiotemporal coordinate system. A graph convolutional network is then used to extract cross-domain correlation features, generating a fused feature tensor with spatiotemporal consistency constraints. Its dimensions include timestamp indexes, spatial position encodings, and multimodal data channels.

[0029] This tensor serves as input to the state assessment layer. In a hybrid model consisting of a convolutional neural network and a long short-term memory network, a three-dimensional convolution kernel slides along the time dimension to extract local spatiotemporal features. This is combined with a bidirectional LSTM to capture long-range dependencies, generating a multi-channel feature map representing displacement deviations, stress fluctuations, and hydraulic anomalies. An anomaly detector, built using a self-attention mechanism, recalibrates the feature map, amplifying the response values in areas of abnormal fluctuations. The final output is a dynamic health index matrix containing a standardized 0-1 score.

[0030] At the decision optimization layer, a multi-objective loss function is constructed that includes displacement synchronization error, pressure balance, and energy consumption coefficient. The displacement synchronization error uses the Hausdorff distance to calculate the maximum local deviation of the displacement sequence of each execution unit from the ideal trajectory. The pressure balance is quantified by calculating the Gini coefficient of the hydraulic system pressure value. The energy consumption coefficient is dynamically estimated by integrating the motor power curve and the hydraulic pipeline loss model. A multi-objective particle swarm optimization algorithm is used to generate a set of candidate parameters. Each particle represents a combination of a set of displacement compensation, hydraulic adjustment parameters, and speed compensation coefficients. The gradient of the loss function is solved in parallel at the local computing nodes through a distributed gradient descent strategy, and the Pareto optimal solution set is finally iterated to output. The calculation formula of the system stability evaluation index is to construct a multi-objective loss function by integrating the displacement synchronization error, pressure balance, and energy consumption coefficient. The specific expression is: ,in represents the displacement synchronization error based on the Hausdorff distance, which quantifies the maximum local deviation between the actual displacement set and the target trajectory. is the Gini coefficient of pressure balance, is the energy consumption coefficient, 、 、 For dynamic weights, real-time adjustment is made through multi-objective particle swarm optimization algorithm to meet and Based on the real-time system stability evaluation index, the optimal parameter combination that takes into account both synchronization accuracy and energy efficiency is selected to generate a control instruction primitive including the actuator compensation amount and oil pressure adjustment range.

[0031] The instruction distribution layer establishes an event-triggered priority queue control mechanism, classifying control instructions into three levels: critical, important, and ordinary instructions, based on their impact on system stability. Critical instructions include emergency braking signals for the hydraulic system and faulty unit isolation instructions, using a preemptive transmission channel to ensure delivery to the execution terminal within 50ms. Important instructions cover speed compensation coefficient updates and dynamic pressure threshold adjustment parameters, transmitted deterministically with timing constraints via a time-sensitive network. Ordinary instructions include health status polling signals and data acquisition instructions, using a best-effort transmission mode. The instruction sequence generation module logically sorts coupled control instructions based on spatiotemporal consistency constraints, ensuring that adjustment instructions for adjacent lifting units are executed sequentially according to the structural force transmission path to avoid stress concentration.

[0032] During the coordinated control process, a fuzzy PID controller dynamically adjusts the speed compensation coefficient based on similarity matching between real-time displacement deviation and a historical operating condition database. A fuzzy control rule is established by constructing a three-dimensional membership function that includes displacement deviation, deviation change rate, and local stiffness coefficient, allowing for real-time adjustment of proportional, integral, and differential coefficient weights. The displacement deviation weight factor is adaptively updated based on the local stiffness coefficient measured by the fiber optic sensor. When a decrease in local bridge stiffness is detected, the displacement deviation weight is automatically reduced to prevent overshoot. The pressure threshold range uses a long-short-term memory network to predict the structural stress distribution trend over the next three seconds. Combined with the nonlinear characteristic curve of the hydraulic actuator, a three-dimensional safety boundary surface containing the pressure-flow-displacement relationship is constructed, dynamically generating a pressure tolerance band that changes with the jacking stage. A bidirectional data exchange channel is established to coordinate the master and slave groups. The master group's fully closed-loop feedback signal serves as the slave group's feedforward input, enabling preemptive compensation of expected displacement deviations through model predictive control. Simultaneously, the slave group's pressure fluctuation data is transmitted back to the master group's control node. A sliding window variance analysis method is used to identify abnormal fluctuation patterns and dynamically adjust the master group's pressure threshold tolerance range.

[0033] Step S300: When the load balancing model detects that the performance of the jacking execution unit deviates, a multi-level fault tolerance mechanism with topology reconstruction capability is activated.

[0034] In this embodiment, when it is detected that the displacement deviation of a certain execution unit exceeds ±1.5mm for two consecutive sampling periods or the hydraulic pressure fluctuation amplitude exceeds the dynamic safety boundary by 15%, the system triggers the first-level fault-tolerant response. For the control loop where the abnormal unit is located, a PID parameter self-tuning strategy based on fuzzy inference rules is adopted: according to the real-time displacement deviation and its rate of change , dynamically adjust the proportional coefficient through the membership function , integration time and differential time Specifically, a quantization interval of input variables containing 7 fuzzy levels is constructed. and Activate the quick correction mode and Increased to 1.2 times the baseline value, Shortened to 80ms, The time interval is extended to 120ms, and the overlapping factor of the membership function is optimized online by the gradient descent method, so that the time integral square error index of the regulation process is reduced to below 0.8.

[0035] If the performance deviation does not converge to the allowable range within 5 seconds after the first-level response is implemented, the second-level adjacent unit coupling compensation mechanism will be activated. i The set of adjacent cells For an associated unit with a spatial distance less than 2.5m and a stiffness coupling coefficient greater than 0.6, the compensation calculation model is: ,in, express The set of adjacent cells of a cell, For adjacent units, is the displacement difference between the adjacent unit and the current unit, is the velocity difference between the adjacent unit and the current unit, and the stiffness coupling coefficient The node stiffness ratio is determined based on the finite element model, and the value range is [0.5, 1.2]. The damping coefficient According to the hydraulic system response characteristics, it is set to .

[0036] If the system still cannot return to a stable state after the second level compensation (e.g., the pressure balance degree is continuously lower than 0.75 for more than 8 seconds), the third level topology reorganization is activated. First, a weighted connection graph is constructed. , where the vertex set Represents the lifting execution unit, vertex weight 0.7×health index + 0.3×load margin, the health index is normalized to the [0,1] interval by the dynamic eigenvalue output by the convolutional neural network; the edge set E represents the communication link between units, and the edge weight 0.6×communication reliability + 0.4×pressure transmission efficiency. Communication reliability is calculated based on the historical packet loss rate. (Number of packet losses / total number of transmissions), the preset communication reliability threshold is not less than 0.85. The path scoring function is , where the product term Strengthen path communication reliability, cumulative items Reflects the overall performance of the node. During the reorganization process, the parameter set of the dynamic load balancing model is updated synchronously to ensure that the system energy consumption coefficient after pressure redistribution does not exceed 1.25 times the initial value.

[0037] Step S400: When the system stability output by the load balancing model is lower than a critical threshold, a hierarchical degradation control mode is initiated, and control weights are redistributed based on the optimal combination of the remaining jacking execution units.

[0038] In this embodiment, when the system stability index drops to a critical threshold (0.65 in this example), a distributed decision-making module automatically activates a hierarchical degradation control protocol, matching the optimal response strategy from the four degradation criteria based on real-time diagnostic results. In the communication system degradation mode, if the end-to-end latency of the backbone fiber network exceeds 50ms for three sampling periods or the data loss rate exceeds 5%, the control center automatically switches to a pre-configured redundant wireless mesh network and reconfigures the communication topology using a dynamic routing algorithm based on the IEEE 802.11s protocol. In this mode, a priority filtering mechanism is used to hierarchically process control command flows, raising the transmission priority of critical data such as displacement compensation commands and hydraulic control signals to QoS level 7, while downgrading non-real-time tasks such as sampled data backhaul to QoS level 3. Forward error correction coding (FEC) is also enabled to limit command retransmissions to three times, ensuring reliable delivery of critical commands within 100ms.

[0039] For actuator degradation mode, when the leakage rate sensor of the hydraulic actuator detects a leakage Or the laser displacement sensor detects displacement deviation exceeding When the fault occurs, the safety interlock module immediately cuts off the power supply to the faulty unit and maintains the balance of the lifting plane through the force coupling compensation algorithm of the adjacent units. With the faulty unit as the center, three spare units are selected within a radius of 2 meters to form a compensation group. The force balance equation group is constructed based on the spatial position relationship: ,in is the theoretical output value of the faulty unit, is the position angle between each compensation unit and the fault unit, and the target output increment of each compensation unit is solved by the least square method. , and at the same time introduce a 5% safety margin factor to prevent overload.

[0040] In the sensing system degradation mode, when the number of failed nodes in the fiber optic sensor cluster reaches 30% of the total, the multi-source data fusion compensation mechanism is activated. A 16-line laser radar is used to scan the bridge bottom surface at a frequency of 20 Hz to generate high-density point cloud data. The deformation characteristics of the jacking area are extracted using the RANSAC algorithm. The redundant measurement values of the embedded strain gauges are simultaneously collected. The weighted average fusion algorithm is used to reconstruct the deformation monitoring system. The deformation compensation value is calculated as follows: ,in For laser radar The displacement value of the feature point, For strain gauges Calculated displacement compensation amount, weight coefficient 、 Dynamically adjusted based on sensor confidence, confidence evaluation indicators include data fluctuation variance in the last minute and hardware health status parameters.

[0041] When the system stability index falls below the minimum threshold (0.5 in this example) for five consecutive sampling cycles, the integrated degradation mode is triggered, and communication link switching, faulty unit isolation, and sensor compensation mechanisms are synchronously executed. Cross-module timing synchronization is achieved through the degradation strategy coordinator to ensure that the degradation operations of each subsystem are phase-aligned within a 500ms time window.

[0042] In the control weight redistribution stage, first construct a three-dimensional evaluation matrix ,in represents the unit health index, is the load capacity margin, which is the inverse of the ratio of the current load to the maximum allowable load. The strategic value of the location, that is, the contribution of the jacking point of the unit to the overall bending moment of the bridge is calculated to calculate the comprehensive performance score The Hungarian algorithm is used to solve the optimal solution for task allocation. Units with a score greater than or equal to 0.8 are preferentially allocated to key support points. A bipartite graph model is constructed: the left vertex set represents the tasks to be assigned, the right vertex set represents the available execution units, and the edge weights are The degree of matching with the task requirements is jointly determined. During the algorithm iteration process, virtual vertices are introduced to handle the imbalance between supply and demand, and finally the minimum cost allocation solution is obtained to ensure that at least 80% of the key support points are covered by high-performance units. The dynamic weight adjustment module realizes real-time optimization of control weights based on the Kalman filter and establishes the state equation and the observation equation , where the state vector Contains the weight distribution coefficient of each unit, control input Reflects the real-time load change rate, observation matrix It is composed of the displacement synchronization error gradient. Through a prediction-correction loop iteration, the weight distribution ratio is updated every 200ms, reducing the dynamic response time of the control system to 60% of the original baseline while improving pressure balance by over 15%. Throughout the degradation control process, the system continuously monitors the recovery of the stability index. When the index rises above 0.75 and remains so for 10 sampling periods, the degradation mode is gradually lifted, and the control weights of each unit are restored in steps of 0.1 to avoid secondary oscillations caused by sudden changes in state.

[0043] Example 2: Figure 2 As shown, the present invention provides an adaptive synchronous degradation control system for bridge jacking, comprising: Multi-dimensional monitoring initialization module 1 initializes the multi-dimensional synchronous monitoring system based on the bridge structure characteristics and jacking target parameters, and establishes the dynamic benchmark parameters of each jacking execution unit; Dynamic load coordination module 2 builds a dynamic load balancing model based on dynamic benchmark parameters, and realizes coordinated control of multiple lifting execution units through adaptive adjustment of speed compensation coefficient and pressure threshold range; Fault-tolerant topology reconstruction module 3 activates a multi-level fault-tolerant mechanism with topology reconstruction capability when the load balancing model detects performance deviation of the jacking execution unit; The hierarchical weight optimization module 4 starts the hierarchical degradation control mode when the system stability output by the load balancing model is lower than the critical threshold, and redistributes the control rights based on the optimal combination of the remaining jacking execution units.

[0044] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. An adaptive synchronous degradation control method and system for bridge jacking, characterized in that: The method comprises: Step S100: Based on the bridge structure characteristics and jacking target parameters, initialize the multi-dimensional synchronous monitoring system and establish the dynamic benchmark parameters of each jacking execution unit; Step S200: constructing a dynamic load balancing model based on dynamic reference parameters, and implementing coordinated control of multiple lifting execution units through adaptive adjustment of the speed compensation coefficient and the pressure threshold range; Step S300: activating a multi-level fault-tolerance mechanism with topology reconstruction capability when the load balancing model detects that the performance of the jacking execution unit deviates; Step S400: When the system stability output by the load balancing model is lower than a critical threshold, a hierarchical degradation control mode is initiated, and control weights are redistributed based on the optimal combination of the remaining jacking execution units.

2. The adaptive synchronous degradation control method for bridge jacking according to claim 1, characterized in that: The dynamic reference parameters include a set of jacking point coordinates acquired by a three-dimensional laser scanner and a local stiffness coefficient matrix measured by an optical fiber sensor.

3. The adaptive synchronous degradation control method for bridge jacking according to claim 2, characterized in that: A regional grouping strategy is adopted to divide the lifting execution units into a master control group and a slave group. The master control group adopts full closed-loop feedback control, while the slave group adopts a feedforward-feedback composite control mode. The regional grouping strategy adopts an improved K-means clustering algorithm, and its objective function is: in Group the lifting execution units into groups. is the total number of clusters, For the A sample set of clusters, each cluster corresponds to a lifting execution unit group, , is the weight coefficient, express The set of adjacent samples of is the sample point, are adjacent sample points, For the The clustering center of each cluster, the clustering dimensions include real-time load rate, displacement deviation and adjacent unit coupling coefficient; The comprehensive distance between each unit and the cluster center is calculated, and the master control group is determined based on the comprehensive clustering dimension. The remaining lifting execution units are classified into subordinate groups.

4. The adaptive synchronous degradation control method for bridge jacking according to claim 3, characterized in that: The dynamic load balancing model includes: The data fusion layer uses a time series alignment algorithm and a graph-based heterogeneous data fusion method to output a fused feature tensor with spatiotemporal consistency; The state assessment layer uses a hybrid model of convolutional neural networks and long short-term memory networks. It generates a dynamic health index that includes displacement deviation, stress fluctuation, and hydraulic anomalies based on the fused feature tensor through multi-channel feature extraction. It also builds an anomaly detector based on the self-attention mechanism and outputs the performance indicators of the lifting execution unit. The decision optimization layer integrates the multi-objective particle swarm optimization algorithm and the distributed gradient descent strategy to construct a loss function that includes displacement synchronization error, pressure balance, and energy consumption coefficient, and outputs a system stability evaluation index that includes the optimal displacement compensation amount and hydraulic adjustment parameters; The instruction distribution layer builds a priority queue control mechanism to generate a control instruction sequence including jacking speed instructions, oil pressure adjustment amplitude and actuator compensation, and outputs a dynamic adjustment parameter set including instruction priority identification and timing constraints.

5. The adaptive synchronous degradation control method for bridge jacking according to claim 4, characterized in that: The collaborative control is specifically as follows: Based on the similarity matching results between the real-time displacement deviation and the historical working conditions, a fuzzy PID controller is used to dynamically adjust the speed compensation coefficient, where the displacement deviation weight factor is adaptively updated as the local stiffness of the bridge changes. The pressure threshold range uses a long short-term memory network to predict the structural stress distribution within the next 3 seconds, and combines the nonlinear characteristic curve of the hydraulic actuator to generate a dynamic pressure safety boundary; In view of the collaborative relationship between the master control group and the slave group, a feedforward-feedback composite control channel is constructed. The full closed-loop feedback signal of the master control group serves as the feedforward input of the slave group. At the same time, the pressure fluctuation data of the slave group reversely corrects the pressure threshold tolerance range of the master control group to achieve global pressure balance.

6. The adaptive synchronous degradation control method for bridge jacking according to claim 1, characterized in that: The multi-level fault-tolerant mechanism includes three-level response strategies: the first level is local PID parameter self-tuning, the second level starts adjacent unit coupling compensation, and the third level performs topology reorganization, where the second level compensation amount , express The set of adjacent cells of a cell, For adjacent units, 、 are the stiffness coefficient and the damping coefficient, is the displacement difference between the adjacent unit and the current unit, is the speed difference between the adjacent unit and the current unit.

7. The adaptive synchronous degradation control method for bridge jacking according to claim 6, characterized in that: The topology reorganization comprises the following steps: Construct a weighted connection graph, vertex weight Reflects unit health, edge weight Indicates communication reliability; The modified Dijkstra algorithm is used to obtain the optimal pressure transmission path, and the path scoring function , where Pressure transmission path, Representative Path The edge passing through Representative Path The vertices passed through; The topology structure is reorganized according to the optimal pressure transmission path.

8. The adaptive synchronous degradation control method for bridge jacking according to claim 1, characterized in that: The hierarchical degradation control mode includes four levels of standards, corresponding to communication system degradation, execution unit degradation, sensor system degradation and comprehensive degradation mode; The communication system degradation mode is to switch to the redundant wireless mesh network when the backbone network delay exceeds 50ms or the packet loss rate is greater than 5%, and adopt a priority filtering mechanism to ensure command transmission; The actuator unit degradation mode is to automatically isolate the faulty unit and maintain the lifting plane through force coupling compensation of adjacent units when a hydraulic cylinder leakage rate greater than or equal to 0.5L / min or a displacement deviation continuously exceeds the limit for 10 seconds is detected; The sensor system degradation mode is to activate the fusion compensation algorithm of lidar point cloud data and strain gauge redundant measurement values to reconstruct the deformation monitoring system when the number of fiber optic sensor failures exceeds 30% of the total number of clusters; The comprehensive degradation mode is to synchronously trigger a joint degradation strategy of the communication, execution, and sensing systems when the system stability index is lower than a preset minimum threshold.

9. The adaptive synchronous degradation control method for bridge jacking according to claim 8, characterized in that: The reallocation of control weights includes: Construct an evaluation matrix that includes unit health, load capacity margin, and location strategic value, and calculate the comprehensive performance score of each remaining unit; The Hungarian algorithm is used to solve the optimal task allocation plan, and the lifting execution units with comprehensive efficiency scores exceeding the preset score threshold are allocated to the key support points; A dynamic weight adjustment strategy is introduced to update the control weight distribution ratio in real time based on the Kalman filter.

10. Adaptive synchronous degradation control system for bridge jacking, characterized by: The system comprises: The multi-dimensional monitoring initialization module initializes the multi-dimensional synchronous monitoring system based on the bridge structure characteristics and jacking target parameters, and establishes the dynamic benchmark parameters of each jacking execution unit; The dynamic load coordination module builds a dynamic load balancing model based on dynamic benchmark parameters, and realizes the coordinated control of multiple lifting execution units through adaptive adjustment of the speed compensation coefficient and pressure threshold range; A fault-tolerant topology reconstruction module activates a multi-level fault-tolerant mechanism with topology reconstruction capabilities when the load balancing model detects performance deviation of the jacking execution unit; The hierarchical weight optimization module starts the hierarchical degradation control mode when the system stability output by the load balancing model is lower than the critical threshold, and redistributes the control weights based on the optimal combination of the remaining jacking execution units.

Images