A method for adjusting and optimizing the load force of floating slabs based on the force balance of vibration isolators

Abstract

This invention relates to the field of floating slab maintenance technology for rail transit, and particularly to a method for adjusting and optimizing the load force of floating slabs based on the force balance of vibration isolators. The method includes the following steps: S1: Divide the entire floating slab to be adjusted into continuous sections according to a preset maintenance window length, with the adjustment workload of each section adapted to a single maintenance window time; S2: During the first maintenance window time, collect static pressure data of the vibration isolators in the first section, apply statistical criteria to remove outliers, and calculate the weighted average of the remaining valid data as the virtual reference force for the first section. This invention, through zoning planning, benchmark transfer, overlap verification, and closed-loop control, achieves high-efficiency, high-precision, and globally balanced adjustment and optimization of the force on the vibration isolators of floating slabs under the dual constraints of limited maintenance window time and missing design parameters in existing line renovations.

Description

Technical Field

[0001] This invention relates to the field of maintenance technology for floating slabs in rail transit, and in particular to a method for adjusting and optimizing the load force of floating slabs based on the force balance of vibration isolators. Background Technology

[0002] In the operation and maintenance technology system of floating slabs, a long-standing but never fully recognized technical challenge is that, due to uneven settlement and differences in initial construction quality, the stress state of each vibration isolator varies significantly. In extreme cases, some vibration isolators may become suspended, while adjacent vibration isolators may be severely overloaded. This not only significantly reduces the vibration reduction performance of the floating slab and accelerates the fatigue damage of the vibration isolators, but also causes the track geometry to exceed the limits, posing a potential threat to traffic safety.

[0003] Existing technologies have significant shortcomings when dealing with the specific scenario of existing line renovation. Traditional adjustment methods based on absolute design loads, while mature in new lines, rely heavily on precise theoretical design values ​​and are almost completely ineffective in existing line renovation projects where original data is lacking. Local adjustment methods based on manual experience, such as simple suspended inspections and replacement after damage, do not rely on design parameters but lack a stress balancing strategy and cannot perceive the stress distribution of all vibration isolators, often leading to new stress imbalances in adjacent areas after local adjustments. Equalization adjustment methods based on synchronous data acquisition across the entire line can theoretically achieve relatively uniformity if they do not rely on a stress balancing strategy based on prior data, but this implicitly requires that data acquisition and processing across the entire line be completed within the same quiet window. However, existing line renovation is constrained by the extremely limited track time, i.e., a single operation window is usually only a few hours, making it impossible to complete synchronous data adjustment across the entire line for large-scale lines.

[0004] Faced with the dual constraints of missing design parameters and limited time windows, existing technologies cannot establish a benchmark transfer mechanism that can transcend time windows. They cannot ensure the efficiency of adjustments within a section while guaranteeing that the entire line will eventually converge to an overall equilibrium state. This results in a large number of existing lines having long-term abnormal operation of floating slabs. There is an urgent need for a floating slab load force adjustment and optimization method that can adapt to the time window operation mode, does not rely on prior data, and can guarantee global equilibrium. Summary of the Invention

[0005] To overcome the shortcomings of existing technologies that rely on absolute design loads and are limited by skylight time, this invention provides a method for adjusting and optimizing the load force of floating slabs based on the force balance of vibration isolators.

[0006] The technical implementation scheme of the present invention is as follows: a method for adjusting and optimizing the load force of a floating slab based on the force balance of a vibration isolator, comprising the following steps:

[0007] S1: Based on the preset window time length, the entire floating slab to be adjusted is divided into continuous sections, and the adjustment workload of each section is adapted to the single window time.

[0008] S2: During the first window of time, collect the static pressure data of the vibration isolators in the first section, apply statistical criteria to remove outliers, and calculate the weighted average of the remaining valid data as the virtual reference force of the first section;

[0009] S3: The virtual reference force of the first section is determined as the reference force. The reference force is used as the target reference for the adjustment of subsequent sections. During the subsequent window time, the subsequent sections are adjusted sequentially so that the force on the vibration isolators of the subsequent sections converges to the reference force.

[0010] S4: An overlapping verification zone is set at the junction of adjacent sections. The overlapping verification zone includes two vibration isolators that cross the boundary of the adjacent sections. When adjusting the next section, the force value of the vibration isolators in the overlapping verification zone is re-measured. When the deviation from the reference force exceeds the first preset threshold, the adjustment target of the next section is corrected according to the re-measurement result.

[0011] S5: After all sections have been adjusted, conduct a full-line sampling inspection to verify the overall stress balance.

[0012] Preferably, the step of dividing the entire floating slab to be adjusted into continuous sections according to the preset skylight time length, with the adjustment workload of each section adapted to a single skylight time, includes: obtaining the available working time of a single skylight as the single skylight time; calculating the maximum number of vibration isolators that can be adjusted within the single skylight time based on the single skylight time and the working efficiency of synchronous jacking; and dividing the entire line of vibration isolators into continuous sections with the maximum number of vibration isolators as the upper limit and the slab gap position of the floating slab as the natural boundary.

[0013] Preferably, the step of calculating the maximum number of vibration isolators adjusted within a single skylight based on the single skylight duration and the synchronous jacking efficiency includes: the formula for calculating the maximum number of adjustments is: ;in, To improve the efficiency of synchronous lifting; The available working time for a single skylight is expressed in minutes. The preparation time before each skylight maintenance period begins, in minutes; The standard adjustment time for a single vibration isolator includes the average time required for load testing, jacking adjustment, and retesting, in minutes. The number of vibration isolators adjusted simultaneously for synchronous jacking; The average time, in minutes, required to move the equipment to the next batch of vibration isolators after completing one batch of adjustments; This is the floor function.

[0014] Preferably, the step of collecting static pressure data of the vibration isolators in the first section, applying statistical criteria to remove outliers, and calculating the weighted average of the remaining valid data as the virtual reference force for the first section includes: during quiet periods when no trains are passing, using load testing equipment to sequentially perform jacking tests on each vibration isolator in the first section, recording the static load force values, and forming an initial dataset. , The total number of vibration isolators in the first section is recorded; the location, number, and detection time of each vibration isolator are recorded synchronously to construct a dataset containing spatial location labels; the 3σ criterion is applied to initially identify and remove obvious outliers caused by sensor failures or local abnormal operating conditions, resulting in a preliminarily cleaned dataset. , The initial number of valid data points to retain; based on a deep autoencoder, the dataset after initial cleaning... Secondary anomaly detection and data repair are performed to obtain the repaired dataset. The repaired dataset As the remaining valid data, each vibration isolator is assigned a different weighting coefficient according to its historical reliability in the floating plate; the weighted average of the remaining valid data is calculated, and the weighted average is used as the virtual reference force of the first segment.

[0015] Preferably, the dataset after initial cleaning is processed using a deep autoencoder. Secondary anomaly detection and data repair are performed to obtain the repaired dataset. This includes: constructing a deep autoencoder network consisting of an encoder and a decoder, where the encoder processes the input data. Mapped to low-dimensional hidden layer features The decoder reconstructs the hidden layer features into output values, denoted as... ,satisfy ,in, This represents the encoder mapping function, used to compress input data into a low-dimensional feature space; This represents the decoder mapping function, used to reconstruct data from low-dimensional features into the original space; For the first part of the pre-cleaned dataset One data point; For the data points Low-dimensional hidden layer features obtained after encoder mapping; and The network parameters are defined using historical vibration isolator force data collected under normal operating conditions to train the autoencoder, with the optimization objective being to minimize the reconstruction error. After training is completed, for the input data Calculate reconstruction error ;when Greater than the preset reconstruction error threshold ,determination For outliers, use the output value. Replace the input data Repair; when The original values ​​are preserved; after the above processing, all data yields the repaired dataset. , among which when hour, ,when hour, .

[0016] Preferably, the repaired dataset As the remaining valid data, different weighting coefficients are assigned to each vibration isolator according to its historical reliability in the floating slab; the weighted average of the remaining valid data is calculated, including: for each repaired data... Assign weight coefficients The weighting coefficient The determination includes the spatial weight of the location of the vibration isolator. The vibration isolators are positioned according to their location within the floating slab, specifically at the slab end, slab center, and slab edge, with higher weighting for isolators at the slab end than those at the slab center; historical reliability indicators. The confidence level of the self-encoder reconstruction is calculated based on the variance of the historical test data of the vibration isolator and its correlation with adjacent vibration isolators. The smaller the variance and the stronger the correlation, the higher the weight. Reconstruction error The smaller the value, the higher the confidence level; weighting coefficient The weighted combination after normalization is obtained, and the calculation formula is: ,in , , The preset combination coefficients satisfy... Calculate the weighted average. ,Will As a virtual reference force for the first segment.

[0017] Preferably, determining the virtual reference force of the first segment as the benchmark reference force, and using the benchmark reference force as the target benchmark for adjusting subsequent segments, and sequentially adjusting the subsequent segments within the subsequent window time so that the force on the vibration isolators in the subsequent segments converges to the benchmark reference force, includes: collecting static pressure data of all vibration isolators in the subsequent segments within the corresponding window time; using the benchmark reference force as the target value, driving synchronous jacking through a closed-loop feedback control algorithm to perform batch jacking adjustments on each vibration isolator in the segment; repeating the detection-adjustment-retest process until the deviation between the measured force value of all vibration isolators in the subsequent segments and the benchmark reference force is less than a second preset threshold.

[0018] Preferably, the step of using the reference force as the target value and driving synchronous jacking through a closed-loop feedback control algorithm to adjust each vibration isolator in batches includes: constructing a closed-loop control system with the reference force as input and the real-time force value of each vibration isolator as feedback; and using a proportional-integral-derivative control algorithm to calculate the required shim adjustment amount for each vibration isolator, with the calculation formula as follows: ;in, For the first The required shim adjustment amount for each vibration isolator is in millimeters. , As a reference force, For the first The current measured force value of each vibration isolator; , , These are the proportional coefficient, integral coefficient, and derivative coefficient, determined based on the stiffness characteristics of the vibration isolator. The lifting height is determined based on the adjustment amount of the pad, and the lifting height is converted into a synchronous lifting control command to drive the jack assembly to perform the lifting operation.

[0019] Preferably, an overlapping verification zone is set at the boundary of adjacent sections. The overlapping verification zone includes two vibration isolators that cross the boundary of adjacent sections. When adjusting the next section, the force value of the vibration isolators in the overlapping verification zone is re-measured. When the deviation from the reference force exceeds a first preset threshold, the adjustment target of the next section is corrected according to the re-measurement results. This includes: selecting two vibration isolators that cross the boundary at the boundary between the first section and the next section as overlapping verification points; after completing the adjustment of the first section, recording the final force value of each vibration isolator in the overlapping verification points as the reference anchor point value; when starting to adjust the next section, first re-measuring the current force value of each vibration isolator in the overlapping verification points, and calculating the deviation between the current force value and the reference anchor point value; when the deviation exceeds the first preset threshold, using the weighted average of the current measured values ​​of each vibration isolator in the overlapping verification points as the correction target reference for the adjustment of the next section.

[0020] Preferably, the step of conducting full-line sampling inspection to verify the overall stress balance includes: after all sections have been adjusted, selecting vibration isolators at the junctions of each section, in the middle of the slab, and at the ends of the slab as sampling objects; re-measuring the sampling objects using load testing equipment and recording the static load force values; calculating the difference and ratio of the force values ​​of any two adjacent vibration isolators in the sampling objects to determine if there is a sudden change in stress; when the force values ​​of all sampling objects are within the preset fluctuation threshold and there are no obvious sudden changes, it is determined that the stress on the entire line has reached a balanced state.

[0021] The beneficial effects of this invention are:

[0022] 1. This invention divides the entire floating slab into continuous sections based on the length of the maintenance window. A virtual reference force constructed in the first section is used as a benchmark to transfer to subsequent sections. An overlapping verification zone is set at the boundary of adjacent sections. When adjusting the next section, the force value of the verification point is re-measured. If the deviation exceeds a first preset threshold, the adjustment target is corrected based on the re-measurement results. This achieves the decomposition of large-scale adjustment tasks into multiple maintenance windows for sequential completion without relying on the original design parameters. Furthermore, the overlapping verification mechanism eliminates the sudden changes in force between sections introduced by batch construction. This method effectively overcomes the technical defects of traditional methods, which are limited by the duration of a single maintenance window and cannot achieve synchronous adjustment across the entire line or achieve intra-section balance and inter-section imbalance in batch adjustments. It has outstanding advantages such as strong robustness, low dependence on initial parameters, and suitability for large-scale rapid surveying and remediation of existing lines.

[0023] 2. This invention improves the accuracy and data quality of virtual reference force construction by employing the 3σ criterion to eliminate outliers, using a deep autoencoder for secondary anomaly detection and data repair during the construction of the first segment virtual reference force, and calculating a weighted average value after assigning weights based on the importance of the vibration isolator's position. In subsequent segment adjustments, a closed-loop feedback control algorithm with the benchmark reference force as the target value is used to drive synchronous jacking. This enables precise and quantitative adjustment of the force on the vibration isolator. The method improves upon the low accuracy and poor consistency of traditional manual experience-based adjustments, demonstrating significant advantages of data-driven, precise control, and standardized operation.

[0024] 3. This invention provides a quantifiable method for verifying the overall stress balance by conducting full-line sampling inspections after all sections have been adjusted, and judging whether there are sudden changes in stress based on the difference and ratio of the stress values ​​of adjacent vibration isolators. The floating slabs adjusted in batches can achieve a relatively uniform stress balance, reducing the phenomenon of suspension and overload. It realizes the high-efficiency, high-precision and global balance adjustment and optimization of the stress on the floating slab vibration isolators under the dual constraints of limited time for existing line renovation and lack of original design parameters. Attached Figure Description

[0025] Figure 1This is a flowchart of the floating slab load force adjustment and optimization method based on vibration isolator force balance according to the present invention;

[0026] Figure 2 This is a flowchart of the overlap verification benchmark correction method of the present invention.

[0027] Figure 3 The flowchart illustrates the method for collecting static pressure data of vibration isolators in the first section, applying statistical criteria to remove outliers, and calculating the weighted average of the remaining valid data as the virtual reference force for the first section.

[0028] Figure 4 The flowchart illustrates the method for conducting full-line sampling and testing to verify the overall force balance state of this invention. Detailed Implementation

[0029] The technical solutions of 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.

[0030] Example: A method for adjusting and optimizing the load force of floating slabs based on the force balance of vibration isolators, such as... Figure 1 and Figure 2 As shown, it includes the following steps:

[0031] S1: Based on the preset window time length, the entire floating slab to be adjusted is divided into continuous sections, and the adjustment workload of each section is adapted to the single window time.

[0032] S2: During the first window of time, collect the static pressure data of the vibration isolators in the first section, apply statistical criteria to remove outliers, and calculate the weighted average of the remaining valid data as the virtual reference force of the first section;

[0033] S3: The virtual reference force of the first section is determined as the reference force. The reference force is used as the target reference for the adjustment of subsequent sections. During the subsequent window time, the subsequent sections are adjusted sequentially so that the force on the vibration isolators of the subsequent sections converges to the reference force.

[0034] S4: An overlapping verification zone is set at the junction of adjacent sections. The overlapping verification zone includes two vibration isolators that cross the boundary of the adjacent sections. When adjusting the next section, the force value of the vibration isolators in the overlapping verification zone is re-measured. When the deviation from the reference force exceeds the first preset threshold, the adjustment target of the next section is corrected according to the re-measurement result.

[0035] S5: After all sections have been adjusted, conduct a full-line sampling inspection to verify the overall stress balance.

[0036] The available working time for a single skylight is obtained as the single skylight time; based on the single skylight time and the working efficiency of synchronous jacking, the maximum number of vibration isolators that can be adjusted within the single skylight time is calculated; taking the maximum number of vibration isolators as the upper limit and combining the plate gap position of the floating plate as the natural boundary, the vibration isolators of the entire line are divided into continuous sections.

[0037] It should be explained that the available working time for a single track maintenance window is during the off-peak operating period of the railway line. It is a fixed time window reserved for construction work, after deducting the necessary auxiliary time for personnel to enter and exit the track, equipment deployment, and evacuation. This represents the actual effective working time for vibration isolator adjustment. The available working time for a single track maintenance window is obtained by extracting the start and end times of the nighttime comprehensive maintenance window based on the target line's operating timetable and construction and maintenance work plan, and calculating the difference between the two to obtain the total window duration. This is done according to railway engineering safety regulations and on-site conditions. For operational safety requirements, pre-set preparation and completion times are defined. These include the walking time for personnel from the safe zone to the work site, the time for moving the synchronous lifting equipment, the erection time, and the time for removing the equipment from the line and restoring the line to its normal state after the operation. The difference between these preparation and completion times and the total duration of the skylight is the usable working time for a single skylight. Using the maximum adjustment quantity of the vibration isolators as the upper limit and the location of the gaps between the floating slabs as natural boundaries, the entire line's vibration isolators are divided into continuous sections. The division method is as follows: Based on the on-site survey data of the floating slabs, the mileage coordinates of all slab joints along the entire line are obtained. A slab joint position sequence is formed according to the spatial order of these mileage coordinates. Taking the first vibration isolator after the first slab joint in the position sequence as the starting point, vibration isolators are traversed sequentially along the increasing direction of the line mileage, accumulating the number of traversed vibration isolators. When the accumulated number reaches the maximum adjustment number of the vibration isolators, the current position is recorded as a candidate division point. The overlap between the candidate division point and the slab joint coordinates in the slab joint position sequence is determined. When they overlap, the slab joint... As the end boundary of the segment; when they do not coincide, backtrack to the position of the previous plate gap closest to the candidate dividing point, and use the previous plate gap as the end boundary of the segment. At the same time, the starting boundary of the next segment is determined as the first vibration isolator after the previous plate gap; reset the cumulative count and continue to traverse from the starting boundary of the next segment, repeat the above process until all vibration isolators are covered, and finally output the continuous segment. The number of vibration isolators in each segment does not exceed the maximum adjustment number of vibration isolators, and the segment boundaries are all located at the plate gaps of the floating plate.

[0038] The formula for calculating the maximum adjustment amount is: ;in, To improve the efficiency of synchronous lifting; The available working time for a single skylight is expressed in minutes. The preparation time before each skylight maintenance period begins, in minutes; The standard adjustment time for a single vibration isolator includes the average time required for load testing, jacking adjustment, and retesting, in minutes. The number of vibration isolators adjusted simultaneously for synchronous jacking; The average time, in minutes, required to move the equipment to the next batch of vibration isolators after completing one batch of adjustments; This is the floor function.

[0039] It should be explained that the maximum adjustment quantity mentioned above... The maximum number of vibration isolators that can be adjusted within a single skylight window is expressed in units of individual isolators. The efficiency of the synchronous lifting is reflected in the average cycle time required to adjust a batch of vibration isolators. In parallel operation mode, due to… The equipment works simultaneously, a batch The adjustment time for each vibration isolator is the standard adjustment time for a single vibration isolator. In addition, the inter-batch transfer time, which reflects the efficiency of equipment turnover in the work process, is also considered. ,Right now This constitutes a complete work cycle, which is the time it takes to complete each task. Total time required to adjust a vibration isolator; numerator The net operation time after deducting preparation time is the ratio of this to the operation cycle time, rounded down to obtain the number of complete operation batches that can be completed, and then multiplied by... The maximum number of vibration isolator adjustments that can be completed within a single sunroof is obtained. Preparation time before each skylight opening begins; To define a fixed preparation time independent of the actual work time, a standardized preparation time quota is formed by accumulating the time required for each step—from moving personnel from their work station to the work site, transporting equipment from storage to the work location and completing its installation and commissioning, to removing equipment from the track and restoring the track to its normal condition—based on railway safety rules and on-site work protection requirements. In the formula, and The logical relationship that constitutes the allocation of skylight time. The formula subtracts the net duration after deducting preparation time. The operation is for Defined calibration; select representative vibration isolators from the line as test samples, and have the work team use a standardized synchronous jacking operation procedure to record the time taken for each vibration isolator from load testing to jacking adjustment and then to passing the retest; collect valid test data of no less than a preset sample size threshold, calculate the arithmetic mean of the time taken for all test samples, and set the average value as the standard adjustment time for the individual vibration isolator. The preset sample size threshold is determined based on the minimum sample size requirement in mathematical statistics to ensure the accuracy of mean estimation; for example, it is usually no less than 30 valid samples. The system can be periodically reviewed and updated based on improvements in operator proficiency or process modifications; the number of vibration isolators adjusted simultaneously during the synchronous lifting is also considered. To configure the number of equipment for parallel operations, based on the pump station output flow rate, the number of control valve group channels, and the number of jacks equipped for synchronous jacking hydraulics, the upper limit of the number of jacks that a single configuration can drive simultaneously is determined. Considering the structural characteristics of the floating slab and the spacing of the vibration isolators, the spatial arrangement feasibility of the jacks between adjacent vibration isolators is evaluated. The smaller value between the upper limit of equipment capacity and the upper limit of spatial arrangement feasibility is taken as the number of vibration isolators that can be adjusted simultaneously during synchronous jacking. Measure the center-to-center distance between adjacent vibration isolators, and measure the minimum footprint width of the jack base and cylinder. Determine if the minimum footprint width is less than or equal to the center-to-center distance. If it is satisfied, the spatial arrangement is deemed feasible; otherwise, it is deemed infeasible and the arrangement is subsequently reduced accordingly. The value is taken until the spatial constraints are met; simulating actual operation scenarios, after the work team completes the adjustment of a batch of vibration isolators, the synchronous lifting equipment is disassembled, transferred to the next predetermined work position, and re-erected according to the standardized process. The time taken from the completion time of the previous batch of work to the time when the conditions for the start of the next batch of work are met is recorded; effective transfer tests of no less than the preset sample size threshold are repeated, and the arithmetic mean of the time taken of all test samples is calculated and set as the inter-batch transfer average time. The preset sample size threshold is determined based on the minimum sample size requirement in mathematical statistics to ensure the accuracy of mean estimation; for example, it is usually greater than or equal to 30 valid samples. Periodic reviews and updates can be performed based on workflow optimization; numerator items Divide by the denominator The theoretically achievable number of work batches is obtained, which may be a non-integer number; since only complete work batches can be completed in actual construction, it is necessary to... The function is rounded down to obtain the actual number of complete work batches that can be completed; then multiplied by the number of vibration isolators completed in each batch. Finally, the maximum number of vibration isolator adjustments that can be completed within a single skylight period is obtained. .

[0040] refer to Figure 3 During quiet periods when no trains pass through, load testing equipment is used to sequentially lift and test each vibration isolator in the first section, recording the static load values ​​to form an initial dataset. , The total number of vibration isolators in the first section is recorded; the location, number, and detection time of each vibration isolator are recorded synchronously to construct a dataset containing spatial location labels; the 3σ criterion is applied to initially identify and remove obvious outliers caused by sensor failures or local abnormal operating conditions, resulting in a preliminarily cleaned dataset. , The initial number of valid data points to retain; based on a deep autoencoder, the dataset after initial cleaning... Secondary anomaly detection and data repair are performed to obtain the repaired dataset. The repaired dataset As the remaining valid data, each vibration isolator is assigned a different weighting coefficient according to its historical reliability in the floating plate; the weighted average of the remaining valid data is calculated, and the weighted average is used as the virtual reference force of the first segment.

[0041] It should be explained that the so-called quiet window period is the period during the comprehensive maintenance window of the railway line at night, after confirmation by the train dispatcher and the implementation of safety protection measures, during which no trains are running on the railway line. The time interval between the time when the last train passes through the work section during the maintenance window period and the time when the first train enters the work section the next day is set as the quiet window period. The process of using load testing equipment to lift and test the vibration isolators in the first section is as follows: the operators move the synchronous lifting equipment to the position of the vibration isolator to be tested, set up jacks to make the lifting surface in close contact with the bearing surface of the vibration isolator; apply a preload to eliminate mechanical gaps, and slowly lift at a preset constant rate, while simultaneously collecting the lifting force value in real time through a high-precision pressure sensor integrated in the equipment; when the lifting height reaches a point where the vibration isolator is about to detach from the superstructure but has not yet produced significant displacement, the quiet window period is defined as follows. At the critical state, the stable readings of the sensor are recorded as the static load force value of the vibration isolator; after recording, the load is unloaded and returned to its initial bearing state, and the device is moved to the next vibration isolator to be tested; the data set containing spatial location tags, based on the field survey data of the floating slab, assigns a unique mileage marker coordinate to each vibration isolator, combined with the up / down line, slab joint number, and the sequential number of the vibration isolator within the slab, to form a multi-dimensional spatial positioning identifier, the spatial location tag; the format of the spatial location tag is a structured composite value, such as "down line K1024+356.2-3rd slab-5th vibration isolator", the spatial location tag is used to determine the physical location of the vibration isolator in the line, and establishes a mapping relationship between the load detection data and the spatial distribution of the vibration isolator; the number of initially retained valid data... The meaning is raw collected data. The number of data points remaining after initial anomaly identification and removal of obvious outliers using the 3σ criterion; the calculation logic of the 3σ criterion is to calculate the number of data points remaining in the dataset before initial cleaning. mean with standard deviation , each data point with the mean absolute difference and When comparing, Values ​​identified as outliers are removed. Data points retained after filtering constitute the dataset. ; The numerical range satisfies , and The difference is the number of obviously outlier data points that were removed.

[0042] Construct a deep autoencoder network consisting of an encoder and a decoder. The encoder takes the input data... Mapped to low-dimensional hidden layer features The decoder reconstructs the hidden layer features into output values, denoted as... ,satisfy ,in, This represents the encoder mapping function, used to compress input data into a low-dimensional feature space; This represents the decoder mapping function, used to reconstruct data from low-dimensional features into the original space; For the first part of the pre-cleaned dataset One data point; For the data points Low-dimensional hidden layer features obtained after encoder mapping; and The network parameters are defined using historical vibration isolator force data collected under normal operating conditions to train the autoencoder, with the optimization objective being to minimize the reconstruction error. After training is completed, for the input data Calculate reconstruction error ;when Greater than the preset reconstruction error threshold ,determination For outliers, use the output value. Replace the input data Repair; when The original values ​​are preserved; after the above processing, all data yields the repaired dataset. , among which when hour, ,when hour, .

[0043] It should be explained that the encoder The decoder is composed of multiple stacked fully connected hidden layers, with each hidden layer progressively reducing dimensionality, compressing the high-dimensional input data into a low-dimensional hidden layer feature space. Mathematically, this involves extracting the inherent correlation patterns and statistical regularities from the data. The structure is symmetrical to the encoder, and the dimensionality is recovered step by step by fully connected hidden layers, reconstructing the low-dimensional hidden layer features back to the original data space; the network parameters and This includes the weight matrix and bias vector for each hidden layer. The dimensionality is determined by the network structure, i.e., the number of network layers and the number of neurons in each layer. Based on the dimensionality and sample size of the input data, a grid search is used to pre-determine candidate network structures. Cross-validation is then used to select the structure with the smallest reconstruction error on the validation set as the final network structure. The input data... For the dataset after initial cleaning The first in Data points, For the data points The feature representation obtained after encoding; the decoder converts the hidden layer features The reconstructed output data is used as the reconstructed value. A training sample set is constructed using vibration isolator force data collected under historical normal operating conditions. These historical normal operating conditions refer to load data collected during the initial operation of the line or during stable operation periods confirmed to be free of any abnormalities. Each sample in the training sample set... The input is fed into an autoencoder, and the reconstructed value is obtained through forward propagation. Calculate the reconstruction error The optimization objective The cumulative sum of reconstruction errors for all training samples is used to calculate the loss function with respect to the network parameters via backpropagation. and The gradient is calculated, and the Adam optimizer is used to iteratively update the parameters until the loss function converges or the preset maximum number of iterations is reached. The preset maximum number of iterations is determined based on the convergence curve of the loss function during training. The change of the loss function value with each iteration is monitored in real time. When the decrease in the loss function value in several consecutive iterations is less than a low percentage of the initial loss value, the model can be determined to have converged, and the iteration number at this time is recorded. For example, the model is usually determined to have converged when the decrease in the loss function value in 50 consecutive iterations is less than 0.1% of the initial loss value. In the anomaly detection and data repair stage, the initially cleaned dataset is... Each data point in Input the trained autoencoder to obtain the corresponding reconstructed values. ; Calculate reconstruction error The physical meaning of the reconstruction error is the degree of deviation between the input data and the inherent patterns learned by the autoencoder based on normal patterns. A larger error indicates that the data points do not conform to the statistical characteristics under normal operating conditions. Compared with the preset reconstruction error threshold The preset reconstruction error threshold is compared. The preset reconstruction error threshold is a boundary value used to determine whether a data point is abnormal. The method for obtaining the mean value of the reconstruction error of the training set is to perform statistical analysis based on the reconstruction error of all samples in the training set. with standard deviation ,Will Set as the preset reconstruction error threshold ,in The preset confidence coefficient is used to control the balance between the sensitivity and false alarm rate of anomaly detection. It is based on the statistical distribution of reconstruction error of all samples in the training set, under the assumption of normal distribution. Value corresponds to mean The positions of several times the above standard deviations; when At that time, the judgment For outliers that do not conform to the normal distribution pattern, the reconstructed output of the autoencoder is used. As a repair value, it replaces the original input; when At that time, the judgment The dataset conforms to a normal distribution pattern and retains its original values; after the above judgment and processing, all data points constitute the repaired dataset. Each of them The value selection rule follows the condition judgment result.

[0044] For each repaired data Assign weight coefficients The weighting coefficient The determination includes the spatial weight of the location of the vibration isolator. The vibration isolators are positioned according to their location within the floating slab, specifically at the slab end, slab center, and slab edge, with higher weighting for isolators at the slab end than those at the slab center; historical reliability indicators. The confidence level of the self-encoder reconstruction is calculated based on the variance of the historical test data of the vibration isolator and its correlation with adjacent vibration isolators. The smaller the variance and the stronger the correlation, the higher the weight. Reconstruction error The smaller the value, the higher the confidence level; weighting coefficient The weighted combination after normalization is obtained, and the calculation formula is: ,in , , The preset combination coefficients satisfy... Calculate the weighted average. ,Will As a virtual reference force for the first segment.

[0045] It should be explained that the virtual reference force The weighting coefficient is a comprehensive characterization value of the stress state of the first section of the vibration isolator, expressed in kN. For the repaired data The weights in the weighted average calculation, the weight coefficients are dimensionless; the spatial weights Based on the structural stress characteristics of the vibration isolators in the floating slab, and based on the design drawings of the floating slab, the end, middle, and edge regions of each floating slab are identified. The end region refers to the area near the connection between the two ends of the floating slab and the adjacent slab; the middle region is the central area of ​​the floating slab; and the edge region refers to the two edge regions along the track direction. Referring to the theoretical support reaction distribution coefficients in the standard drawings of similar floating slab structures, the design load ratios of the vibration isolators at the end, edge, and middle positions are obtained. These ratios are used as the basis for weight assignment. For example, if the design load ratio of end:edge:middle is 1.0:0.85:0.65, the corresponding spatial weights are set to 1.0, 0.85, and 0.65, respectively. The load force value sequence of the vibration isolators in multiple historical tests is extracted, and the variance is calculated. The smaller the variance, the more stable the stress state of the vibration isolator and the higher its reliability; at the same time, the Pearson correlation coefficient between the load force value sequence of the vibration isolator and that of the adjacent vibration isolators is calculated. The closer the correlation coefficient is to 1, the more consistent the trend of force changes and the better the overall coordination; combining variance and correlation coefficient, according to the formula... The historical reliability index was calculated. This allows vibration isolators with small variance and strong correlation to receive high weights; the historical detection data comes from existing records of previous skylight periods; the self-encoder reconstruction confidence level In the middle, reconstruction error The formula, calculated from the preceding steps, physically maps the reconstruction error to the (0,1) interval. The smaller the error, the closer the confidence level is to 1; the larger the error, the closer the confidence level is to 0. Based on historical detection data, the entropy weight method is used to determine the relative importance of each weighting factor, and the weights are determined according to the degree of variation of each factor in the historical data. , , The greater the variation, the higher the weight; normalized weights In the calculation, the denominator is the sum of the comprehensive weight values ​​of all vibration isolators, which is used to normalize the weight coefficients of each vibration isolator to a total of 1.

[0046] During the window time corresponding to each subsequent section, static pressure data of all vibration isolators in each subsequent section are collected; using the reference force as the target value, synchronous jacking is driven through a closed-loop feedback control algorithm to adjust each vibration isolator in the section in batches; the detection-adjustment-retest process is repeated until the deviation between the measured force value of all vibration isolators in each subsequent section and the reference force is less than the second preset threshold.

[0047] It should be explained that the "detection-adjustment-retesting" process refers to a cyclical operation of adjusting the vibration isolators in batches within a section, using the reference force as the target value. Detection involves measuring the current force value of each vibration isolator in the current batch before adjustment to obtain real-time load data. Adjustment involves comparing the detected current force value with the reference force, calculating the deviation, and outputting an adjustment command through a closed-loop feedback control algorithm to drive the synchronous lifting device to lift or lower the batch of vibration isolators, bringing the force value closer to the reference force. Retesting involves measuring the force value of each vibration isolator in the batch again after adjustment to verify the adjustment effect and obtain updated force data. If the retesting results show that the deviation still exceeds the allowable range, the detection-adjustment steps are repeated. The process involves iterative corrections until the requirements are met, after which the next batch of work begins. After all batches are completed sequentially, the deviation between the measured force value and the reference force of all vibration isolators in the section is less than the second preset threshold. The second preset threshold is the allowable deviation limit value used to determine whether the force adjustment of the vibration isolator meets the qualified standard, and the unit is kN. The preset method is to extract the allowable deviation range of the static load of the vibration isolator in the relevant specifications for track smoothness maintenance of the floating slab, and set the upper limit of the absolute value of the allowable deviation range as the second preset threshold. If no relevant provisions are found in the specification documents, refer to the construction quality acceptance standards of similar projects, take ±3% to ±5% of the reference force as the allowable deviation range, and set the upper limit of the absolute value as the second preset threshold.

[0048] A closed-loop control system is constructed, using the reference force as input and the real-time force value of each isolator as feedback. A proportional-integral-derivative control algorithm is used to calculate the required shim adjustment for each isolator; the calculation formula is as follows: ;in, For the first The required shim adjustment amount for each vibration isolator is in millimeters. , As a reference force, For the first The current measured force value of each vibration isolator; , , These are the proportional coefficient, integral coefficient, and derivative coefficient, determined based on the stiffness characteristics of the vibration isolator. The lifting height is determined based on the adjustment amount of the pad, and the lifting height is converted into a synchronous lifting control command to drive the jack assembly to perform the lifting operation.

[0049] It should be explained that the adjustment amount of the pad is... For the first The thickness of the pads required to be increased or decreased for each vibration isolator, in millimeters; the force deviation... As a reference force and the first The difference in the currently measured force values ​​of each vibration isolator, in kN; the physical meaning of each part in the proportional-integral-derivative control algorithm is as follows: the proportional term... It provides an immediate response to the current force deviation; the larger the deviation, the greater the adjustment. (Integral term) It compensates for the cumulative effect of deviation over time, and is used to eliminate steady-state residual deviations caused by factors such as friction and leakage; differential term Predictive adjustments are made based on the trend of deviation changes; when the deviation increases, a reverse action is applied in advance to suppress overshoot and oscillations during the adjustment process; the differential term... Indicates force deviation Regarding time The derivative of the deviation, i.e., the rate of change of the deviation, is expressed in kN / min. It characterizes the tendency of the current force value to approach or move away from the reference force. The differential term applies a reverse adjustment in advance based on the rate of change, suppressing overshoot and oscillation during the adjustment process. The linear superposition of the three terms yields the first term. Adjustment amount of shims required for each vibration isolator The unit is millimeters; the proportionality coefficient Integral coefficient Differential coefficients The calibration method involves obtaining the load-displacement stiffness curve of the vibration isolator through static load tests and calculating the stiffness coefficient of the vibration isolator near the target load. The unit is kN / mm, meaning the change in load caused by a unit displacement; the reciprocal of the stiffness coefficient. proportionality coefficient The reference value, in physical terms, is the value that causes the force deviation to be... The basic conversion factor is used to convert the displacement adjustment to the required amount; the Ziegler-Nichols empirical formula is used to... , , Tuning is performed, and the value is taken according to the Ziegler-Nichols empirical formula. , , As the initial setpoint, where This is the critical proportionality coefficient. The oscillation period is defined; the setpoint is fine-tuned and optimized during field tests, with the optimization objectives being the shortest force adjustment time and the smallest overshoot, to determine the final value. , , Get the value; set the first Adjustment amount of the pad for each vibration isolator Establish a mapping relationship with the actual lifting height at the current vibration isolator. A positive value indicates that the thickness of the shim needs to be increased, i.e., the shim needs to be lifted upwards. A negative value indicates that the shim thickness needs to be reduced, i.e., the shim needs to be lowered. Based on the mechanical transmission characteristics and the conversion relationship between hydraulic cylinder stroke and lifting height, the target lifting height is determined. The target displacement of the hydraulic cylinder is converted into millimeters. The synchronous lifting control command is a set of digital commands containing parameters such as target displacement, lifting speed, and acceleration. These commands are sent to the hydraulic valve group controllers of each jack via a control communication bus. Each jack controller adjusts the opening of the proportional valve in real time according to the command. During the lifting process, the displacement sensor provides real-time feedback of the actual displacement value, forming a closed-loop position control.

[0050] At the boundary between the first and next sections, two vibration isolators crossing the boundary are selected as overlapping verification points. After the adjustment of the first section is completed, the final force value of each vibration isolator within the overlapping verification point is recorded as the reference anchor value. When the adjustment of the next section begins, the current force value of each vibration isolator within the overlapping verification point is re-measured, and the deviation between the current force value and the reference anchor value is calculated. When the deviation exceeds a first preset threshold, the weighted average of the current measured values ​​of each vibration isolator within the overlapping verification point is used as the correction target reference for the adjustment of the next section.

[0051] It should be explained that the reference anchor point value , The final force value of the two vibration isolators within the overlapping verification point after the first section adjustment is completed, in kN; the current remeasured value , The force values, in kN, are the force values ​​obtained by re-measuring the two vibration isolators within the overlapping verification point when starting to adjust the next section; the absolute deviation is... , The absolute value of the difference between the current remeasured value of each vibration isolator and the reference anchor point value, in kN; the maximum deviation The value is the maximum of the two absolute deviations, expressed in kN. The overlapping verification point is selected by choosing at least one vibration isolator closest to the plate joint on both sides of the plate joint at the junction of the first and next sections, ensuring that the two verification points belong to two different sections but are geographically adjacent. Based on the plate joint coordinates and vibration isolator layout diagram determined in the previous steps, the number information of the first vibration isolator on both sides of the plate joint is extracted, set as the overlapping verification point, and stored in the configuration. The reference anchor point value is recorded as follows: after all vibration isolators in the first section have been adjusted and the force deviation has been confirmed to be less than the second preset threshold, the final force value of the two vibration isolators within the overlapping verification point is immediately measured. The measurement result is used as the reference anchor point value and stored in the verification database in association with the vibration isolator number and detection time information. Let the reference anchor point values ​​of the two vibration isolators within the overlapping verification point be... and The current retest values ​​are respectively and Calculate the absolute deviation of each vibration isolator separately. and Take the maximum of the two as The deviation is compared with a first preset threshold, wherein... The first preset threshold is the allowable deviation limit value used to determine whether the stress state of the overlapping verification point has changed significantly. It is determined based on the control accuracy of the synchronous jacking adjustment operation and the repeatability error of the vibration isolator force measurement, and is taken as the larger of three times the sensor measurement accuracy value and the jacking control accuracy value. The sensor measurement accuracy value comes from the maximum allowable error calibrated by the load detection equipment at the factory, for example, ±0.5kN. The jacking control accuracy value comes from the force change corresponding to the positioning accuracy of the synchronous jacking closed-loop control, based on the stiffness coefficient. With displacement control accuracy product Calculations show that, for example, a displacement control accuracy of ±0.1 mm and a stiffness coefficient of 10 kN / mm correspond to a force change of ±1.0 kN. The larger of the displacement control accuracy and the stiffness coefficient is taken as the first preset threshold, for example, 1.0 kN. When the deviation... When the first preset threshold is exceeded, the weighted average of the current measured values ​​of the two vibration isolators within the overlapping verification point is used as the correction target benchmark for the next segment adjustment. The calculation process of the weighted average is based on the historical reliability index of each of the two vibration isolators. and Determine the weights and calculate the weighting coefficients for the two vibration isolators. , The current measured values ​​of the two vibration isolators and The corrected target benchmark value is obtained by performing a weighted average based on the above weights. In the adjustment work of the next section, with The original reference force is replaced as the target value for closed-loop control.

[0052] refer to Figure 4 After all sections have been adjusted, vibration isolators at the junctions of each section, in the middle of the slab, and at the ends of the slab are selected as inspection targets. Load testing equipment is used to re-measure the inspection targets and record the static load force values. The difference and ratio of the force values ​​of any two adjacent vibration isolators in the inspection targets are calculated to determine whether there is a sudden change in force. When the force values ​​of all inspection targets are within the preset fluctuation threshold and there is no obvious sudden change, it is determined that the force of the entire line has reached a state of equilibrium.

[0053] It should be explained that the selection method for the sampling objects is as follows: based on the section division results and the vibration isolator layout diagram, one vibration isolator on each side of the slab gap at the junction of each section and the adjacent section, one vibration isolator in the middle of the slab in each section, and vibration isolators at the slab ends at the start and end points of the entire line are automatically extracted to form the sampling object set; the retesting process includes, after all section adjustment work is completed, during the subsequent maintenance window, the operators carry load testing equipment to the location of each sampling object, record the stable reading values ​​of the sensors as the static load force value of the sampling object; each sampling object is measured twice, and the average of the two measurements is taken as the final retest value; the difference and ratio of the force values ​​of any two adjacent vibration isolators in the sampling object are calculated to determine the existence of sudden force changes. Specifically, the sampling objects are arranged in order of line mileage to form a retest value sequence. ,in The total number of samples; for any two adjacent test values ​​in the sequence and Calculate the absolute difference and relative ratio The absolute difference The unit is kN, and the relative ratio is... Dimensionless; when the absolute difference The relative ratio is greater than the preset difference threshold and When the force exceeds a preset ratio threshold, a sudden change in force is determined to exist at the adjacent location. If only one condition is met while the other is within limits, the stiffness characteristics of adjacent isolators are reviewed. A large stiffness difference indicates a normal force gradient, while similar stiffness indicates a sudden change. After traversing all adjacent pairs, the number of locations with sudden changes is counted. The preset fluctuation threshold is used to determine whether the force value of a single isolator meets the expected target within the allowable deviation range. It is set based on the second preset threshold determined in the previous steps, i.e., the value of the second preset threshold is taken as the preset fluctuation threshold. Physically, this means that the allowable deviation of the force value of a single isolator relative to the reference force is also applicable in the sampling and retesting stage. The preset difference threshold and the preset ratio threshold are used to determine whether there is a sudden change in the force values ​​of adjacent isolators. The preset difference threshold is in kN, and the preset ratio threshold is dimensionless. The preset difference threshold is determined based on the allowable limit for load differences between adjacent vibration isolators in the floating slab design specification. When no specific regulations exist, the preset difference threshold is set to 5% to 10% of the reference force, based on the vibration isolator stiffness characteristics and track smoothness requirements. For example, if the reference force is 100 kN, the preset difference threshold is 5 to 10 kN. The preset ratio threshold is set based on experience from similar projects and force distribution patterns, setting it to 1.10 to 1.20. That is, a difference of no more than 10% to 20% in the force values ​​of adjacent vibration isolators is considered a normal force gradient; exceeding this range is considered a sudden change in force. When the retest values ​​of all sampled objects meet the requirements… If no sudden change in force is detected after traversing adjacent pairs, it is determined that the force on the entire line has reached a state of equilibrium.

[0054] The above embodiments are provided for those skilled in the art to implement or use the present invention. Those skilled in the art can make various modifications or changes to the above embodiments without departing from the inventive concept of the present invention. Therefore, the protection scope of the present invention is not limited to the above embodiments, but should be the maximum scope that conforms to the innovative features mentioned in the claims.

Claims

1. A method for adjusting and optimizing the load force of a floating slab based on the force balance of vibration isolators, characterized in that, Includes the following steps: S1: Based on the preset window time length, the entire floating slab to be adjusted is divided into continuous sections, and the adjustment workload of each section is adapted to the single window time. S2: During the first window of time, collect the static pressure data of the vibration isolators in the first section, apply statistical criteria to remove outliers, and calculate the weighted average of the remaining valid data as the virtual reference force of the first section; S3: The virtual reference force of the first section is determined as the reference force. The reference force is used as the target reference for the adjustment of subsequent sections. During the subsequent window time, the subsequent sections are adjusted sequentially so that the force on the vibration isolators of the subsequent sections converges to the reference force. S4: An overlapping verification zone is set at the junction of adjacent sections. The overlapping verification zone includes two vibration isolators that cross the boundary of the adjacent sections. When adjusting the next section, the force value of the vibration isolators in the overlapping verification zone is re-measured. When the deviation from the reference force exceeds the first preset threshold, the adjustment target of the next section is corrected according to the re-measurement result. S5: After all sections have been adjusted, conduct a full-line sampling inspection to verify the overall stress balance.

2. The method for adjusting and optimizing the load force of a floating slab based on the force balance of a vibration isolator, as described in claim 1, is characterized in that... The process involves dividing the entire floating slab to be adjusted into continuous sections based on a preset skylight duration. The adjustment workload for each section is adapted to the duration of a single skylight, including: The available working time for a single skylight is obtained as the single skylight time; Based on the single skylight time and the synchronous jacking efficiency, calculate the maximum number of vibration isolators that can be adjusted within the single skylight time. Using the maximum number of vibration isolators as the upper limit and the position of the plate gap of the floating plate as the natural boundary, the entire vibration isolator is divided into continuous sections.

3. The method for adjusting and optimizing the load force of a floating slab based on the force balance of a vibration isolator, as described in claim 2, is characterized in that... The calculation of the maximum number of vibration isolators adjusted within a single skylight, based on the single skylight duration and the synchronous jacking efficiency, includes: The formula for calculating the maximum adjustment amount is: ;in, To improve the efficiency of synchronous lifting; The available working time for a single skylight is expressed in minutes. The preparation time before each skylight maintenance period begins, in minutes; The standard adjustment time for a single vibration isolator includes the average time required for load testing, jacking adjustment, and retesting, in minutes. The number of vibration isolators adjusted simultaneously for synchronous jacking; The average time, in minutes, required to move the equipment to the next batch of vibration isolators after completing one batch of adjustments; This is the floor function.

4. The method for adjusting and optimizing the load force of a floating slab based on the force balance of a vibration isolator according to claim 1, characterized in that, The static pressure data of the vibration isolators in the first section is collected, and after outliers are removed using statistical criteria, the weighted average of the remaining valid data is calculated as the virtual reference force for the first section, including: During quiet periods when no trains pass, load testing equipment is used to sequentially perform lifting tests on each vibration isolator in the first section, and the static load values ​​are recorded to form an initial dataset. , This represents the total number of vibration isolators in the first section. The location information, number information and detection time information of each vibration isolator are recorded synchronously to construct a dataset containing spatial location labels; The 3σ criterion was applied to initially identify and remove obvious outliers caused by sensor malfunctions or local abnormal operating conditions, resulting in a preliminarily cleaned dataset. , The initial number of valid data points to retain; Based on deep autoencoder analysis of the pre-cleaned dataset Secondary anomaly detection and data repair are performed to obtain the repaired dataset. The repaired dataset As the remaining valid data, different weighting coefficients are assigned to each vibration isolator according to its historical reliability in the floating plate; Calculate the weighted average of the remaining valid data, and use the weighted average as the virtual reference force for the first segment.

5. The method for adjusting and optimizing the load force of a floating slab based on the force balance of a vibration isolator according to claim 4, characterized in that, The dataset after initial cleaning is based on a deep autoencoder. Secondary anomaly detection and data repair are performed to obtain the repaired dataset. ,include: Construct a deep autoencoder network consisting of an encoder and a decoder. The encoder takes the input data... Mapped to low-dimensional hidden layer features The decoder reconstructs the hidden layer features into output values, denoted as... ,satisfy ,in, This represents the encoder mapping function, used to compress input data into a low-dimensional feature space; This represents the decoder mapping function, used to reconstruct data from low-dimensional features into the original space; For the first part of the pre-cleaned dataset One data point; For the data points Low-dimensional hidden layer features obtained after encoder mapping; and For network parameters; The autoencoder was trained using vibration isolator force data collected under historical normal operating conditions, with the optimization objective being to minimize the reconstruction error. ; After training is completed, for the input data Calculate reconstruction error ;when Greater than the preset reconstruction error threshold ,determination For outliers, use the output value. Replace the input data Repair; when Preserve the original values; After the above processing, all data is obtained as the repaired dataset. , among which when hour, ,when hour, .

6. The method for adjusting and optimizing the load force of a floating slab based on the force balance of a vibration isolator, as described in claim 4, is characterized in that... The repaired dataset As the remaining valid data, different weighting coefficients are assigned to each vibration isolator according to its historical reliability in the floating slab. Calculating the weighted average of the remaining valid data includes: For each repaired data Assign weight coefficients The weighting coefficient The determination includes the spatial weight of the location of the vibration isolator. The vibration isolators are positioned according to their location within the floating slab, specifically at the slab end, slab center, and slab edge, with higher weighting for isolators at the slab end than those at the slab center; historical reliability indicators. The variance of the historical test data of the vibration isolator and its correlation with adjacent vibration isolators are calculated. The smaller the variance and the stronger the correlation, the higher the weight. Autoencoder Reconstruction Confidence Reconstruction error The smaller the value, the higher the confidence level; weighting coefficient The weighted combination after normalization is obtained, and the calculation formula is: ,in , , The preset combination coefficients satisfy... ; Calculate the weighted average ,Will As a virtual reference force for the first segment.

7. The method for adjusting and optimizing the load force of a floating slab based on the force balance of a vibration isolator according to claim 1, characterized in that, The step of determining the virtual reference force of the first section as the reference reference force, using the reference reference force as the target reference for adjusting subsequent sections, and sequentially adjusting the subsequent sections during the subsequent window period to make the force on the vibration isolators of the subsequent sections converge towards the reference reference force includes: During the window time corresponding to each subsequent section, static pressure data of all vibration isolators in each subsequent section are collected. Using the aforementioned reference force as the target value, a closed-loop feedback control algorithm is used to drive synchronous jacking, and each vibration isolator in the section is jacked and adjusted in batches. Repeat the detection-adjustment-retest process until the deviation between the measured force value of all vibration isolators in each subsequent section and the reference force is less than the second preset threshold.

8. The method for adjusting and optimizing the load force of a floating slab based on the force balance of a vibration isolator according to claim 1, characterized in that, The process of using the reference force as the target value and driving synchronous jacking through a closed-loop feedback control algorithm to adjust each vibration isolator in the section in batches includes: Construct a closed-loop control system with the reference force as input and the real-time force value of each vibration isolator as feedback; The required shim adjustment for each vibration isolator is calculated using a proportional-integral-derivative (PI-DE) control algorithm. The calculation formula is as follows: ;in, For the first The required shim adjustment amount for each vibration isolator is in millimeters. , As a reference force, For the first The current measured force value of each vibration isolator; , , These are the proportional coefficient, integral coefficient, and derivative coefficient, respectively, determined based on the stiffness characteristics of the vibration isolator. The lifting height is determined based on the adjustment amount of the pad, and the lifting height is converted into a synchronous lifting control command to drive the jack assembly to perform the lifting operation.

9. The method for adjusting and optimizing the load force of a floating slab based on the force balance of a vibration isolator according to claim 1, characterized in that, An overlapping verification zone is set at the boundary of adjacent sections. The overlapping verification zone includes two vibration isolators that cross the boundary of the adjacent sections. When adjusting the next section, the force value of the vibration isolators in the overlapping verification zone is re-measured. When the deviation from the reference force exceeds a first preset threshold, the adjustment target of the next section is corrected according to the re-measurement results, including: At the boundary between the first section and the next section, two vibration isolators that cross the boundary are selected as overlapping verification points; After completing the adjustment of the first section, record the final force value of each vibration isolator within the overlapping verification point as the reference anchor point value; When starting to adjust the next section, first remeasure the current force value of each vibration isolator in the overlapping verification point, and calculate the deviation between the current force value and the reference anchor point value. When the deviation exceeds the first preset threshold, the weighted average of the current measured values ​​of each vibration isolator within the overlapping verification points is used as the correction target benchmark for the next segment adjustment.

10. The method for adjusting and optimizing the load force of a floating slab based on the force balance of a vibration isolator according to claim 1, characterized in that, The process of conducting full-line sampling and testing to verify the overall stress balance includes: After all sections have been adjusted, vibration isolators at the junctions of each section, in the middle of the slab, and at the ends of the slab are selected for random inspection. The static load force value was recorded by retesting the sampled objects using load testing equipment. Calculate the difference and ratio of the force values ​​of any two adjacent vibration isolators in the sampled objects to determine whether there is a sudden change in force. When the stress values ​​of all sampled objects are within the preset fluctuation threshold and there are no obvious sudden changes, it is determined that the stress on the entire line has reached a state of equilibrium.

Images