Method for optimizing vibration reduction parameters of heavy-load air floating floor

Abstract

The invention provides a method for optimizing vibration damping parameters of a heavy-load air floating terrace, and belongs to the technical field of heavy-load air floating terraces. An accurate mathematical analysis framework is constructed by establishing a system dynamics model of the heavy-load air floating terrace and an air floating system performance optimization equation set; an objective multi-dimensional feature data set is established by adopting a pressure sensor array and a vibration sensor network, intelligent parameter identification and optimization are performed by applying a self-adaptive parameter adjustment model, scientific parameter adjustment strategy selection is realized based on a comprehensive performance evaluation index, self-adaptive weight distribution adjustment is realized through a gating weight function, and the accuracy of parameter adjustment is improved. A closed-loop feedback control system is established to realize continuous automatic parameter optimization, and accurate mathematical calculation and an intelligent algorithm are used for replacing unreliable artificial experience judgment, so that the technical problem that the heavy-load air floating floor damping parameter optimization often depends on artificial experience and is not accurate enough is solved.

Description

Technical Field

[0001] This invention belongs to the field of heavy-duty air-floor flooring technology, and more specifically, relates to a method for optimizing vibration reduction parameters of heavy-duty air-floor flooring. Background Technology

[0002] Heavy-duty air-floored flooring, a precision support system based on the principle of gas hydrostatic bearings, is widely used in industrial settings with extremely high vibration isolation requirements, such as ultra-precision machining equipment, large measuring instruments, aerospace test platforms, precision optical equipment, and metrology and testing devices. Traditional methods for optimizing air-floored flooring vibration reduction parameters rely primarily on operators' engineering experience and repeated experiments to adjust key parameters such as air pressure, damping coefficient, and load-bearing stiffness. The final parameter configuration is determined through manual observation and subjective judgment of the system's vibration reduction effect and load-bearing capacity. However, in current applications of heavy-duty air-floored flooring, due to the numerous and interdependent system parameters, the optimal parameter combinations often vary significantly under different operating conditions. Traditional manual adjustment methods cannot accurately grasp the complex relationships between these parameters, making it difficult to find the true optimal solution in the multi-dimensional parameter space. This leads to parameter settings often deviating from the theoretical optimal values, and the system failing to achieve the designed vibration reduction performance and load-bearing capacity in actual operation. Traditional technologies struggle to address the subjectivity and inaccuracy of human experience in vibration reduction parameter optimization. They lack scientific mathematical models and intelligent algorithms to guide parameter setting, making it impossible to achieve parameter optimization based on objective data and precise calculations. This results in low precision in adjusting vibration reduction parameters for heavy-duty air-floored floors, preventing the system from fully realizing its performance. Summary of the Invention

[0003] In view of this, the present invention provides a method for optimizing vibration reduction parameters of heavy-duty air-floor pavement, which can solve the problem that the optimization of vibration reduction parameters of heavy-duty air-floor pavement often relies on human experience and is not accurate enough in the prior art.

[0004] This invention is implemented as follows: A method for optimizing vibration reduction parameters of heavy-duty air-floored pavement is provided. A system dynamic model of the heavy-duty air-floored pavement is established, and a mathematical model describing the coupling relationship between the load-bearing capacity and vibration reduction performance of the air-floored system is constructed. Operating data of the heavy-duty air-floored pavement under different working conditions are collected through a pressure sensor array and a vibration sensor network, establishing a multi-dimensional feature dataset containing air pressure distribution data, load weight data, vibration frequency data, vibration amplitude data, and system damping data. An adaptive parameter adjustment model is used to identify and optimize the nonlinear dynamic characteristics of the air-floored system, balancing the influence of different physical parameters on system performance by adjusting the hierarchical fusion weights. A performance optimization equation set for the air-floored system is established to jointly optimize the system's load-bearing capacity and vibration reduction effect, outputting the maximum load-bearing capacity parameter, optimal damping parameter, and coupling configuration parameter. Based on the maximum load-bearing capacity parameter, optimal damping parameter, and coupling configuration parameter, a comprehensive performance evaluation index of the air-floored system is calculated. The hierarchical fusion weight allocation of the adaptive parameter adjustment model is dynamically adjusted through a gated weight function. A closed-loop feedback control system for the vibration reduction parameters of the air-floored pavement is established to adjust the operating parameters of the heavy-duty air-floored pavement in real time, achieving precise optimization of the vibration reduction parameters.

[0005] Among them, heavy-duty air-floored pavement refers to a ground support system with a load-bearing capacity of more than 10 tons and which uses the principle of gas static pressure bearing to achieve load support and vibration isolation. The system dynamic model refers to a mathematical model that describes the dynamic response characteristics and internal parameter variation law of the air-floored system under external excitation. The mathematical model considers the comprehensive influence of key factors such as air pressure distribution, load weight, vibration transmission and system damping.

[0006] The establishment of the system dynamics model involves analyzing the interaction between the air-bearing mechanism and the vibration reduction characteristics. The air-bearing mechanism refers to the working principle of supporting the load by forming a static pressure field in the air-bearing gap with compressed gas. The vibration reduction characteristics refer to the air-bearing system's ability to isolate and attenuate external vibration excitation. The coupling relationship refers to the physical relationship between the air-bearing capacity and the vibration reduction performance of the air-bearing system, which influence and restrict each other.

[0007] Among them, the pressure sensor array refers to a combination of multiple pressure measuring devices arranged according to spatial distribution rules, which are used to monitor the air pressure distribution state inside the air flotation system in real time. The pressure sensor array measures the air pressure distribution data. The vibration sensor network refers to a distributed monitoring system composed of multiple vibration measuring devices, which is used to obtain the vibration response information of the heavy-duty air flotation floor at different locations and directions. The vibration sensor network measures the vibration frequency data and vibration amplitude data.

[0008] The steps for establishing the multidimensional feature dataset are as follows: load sensors measure load weight data, damping testing devices measure system damping data, and the multidimensional feature dataset refers to a data set containing multiple key physical parameters during the operation of the air flotation system, which is used to describe the comprehensive operating status and performance characteristics of the system.

[0009] The adaptive parameter adjustment model receives a multi-dimensional feature dataset as input and outputs an optimized system parameter configuration to achieve intelligent adjustment of the air flotation system parameters. The hierarchical fusion weight refers to the weight coefficients in the adaptive parameter adjustment model used to balance the importance of features at different abstract levels.

[0010] The air flotation system performance optimization equation set receives air pressure distribution data, load weight data, vibration frequency data, vibration amplitude data, and system damping data as inputs. By solving multiple interrelated mathematical equations, it determines the optimal combination of system operating parameters to achieve a coordinated balance between load-bearing capacity and vibration reduction performance.

[0011] The performance optimization equation set of the air flotation system includes the load-bearing capacity balance equation, the vibration reduction characteristic equation, and the parameter coupling equation. The load-bearing capacity balance equation is used to calculate the optimal air pressure distribution and load-bearing capacity of the air flotation system under a given load condition. The inputs include load weight data, air flotation area parameters, air pressure distribution data, air flotation gap height parameters, and friction coefficient parameters. The outputs are the maximum load-bearing capacity parameters and the optimized air pressure distribution parameters.

[0012] The vibration reduction characteristic equation is used to analyze the vibration transmission characteristics and optimize the vibration reduction effect of the air flotation system. The inputs include vibration frequency data, vibration amplitude data, system damping data, air pressure distribution data and elastic coefficient parameters. The outputs are vibration transmissibility parameters and optimal damping parameters.

[0013] The parametric coupling equation describes the interaction between bearing capacity and vibration reduction performance. The inputs include air pressure coupling coefficient parameters, load change rate parameters, vibration response coefficient parameters, and system stability factor parameters. The outputs are coupling configuration parameters and system comprehensive performance index parameters.

[0014] The calculation steps for the comprehensive performance evaluation index of the air flotation system specifically involve adopting corresponding parameter adjustment strategies based on the numerical range of the comprehensive performance evaluation index of the air flotation system. When the comprehensive performance evaluation index of the air flotation system is in different numerical ranges, the corresponding optimization algorithm and adjustment method are automatically selected.

[0015] The gating weight function receives the comprehensive performance evaluation index of the air flotation system as input, and automatically adjusts the air pressure distribution ratio, system damping parameter and control gain parameter according to the real-time monitored system status information and performance requirements. It outputs the adjusted hierarchical fusion weight. The air pressure distribution ratio refers to the distribution ratio of air pressure values ​​in different areas of the air flotation system, and the control gain parameter refers to the gain coefficient used to adjust the control response intensity in the closed-loop feedback control system.

[0016] The closed-loop feedback control system receives the adjusted hierarchical fusion weights and the comprehensive performance evaluation index of the air flotation system as inputs. Based on the system performance evaluation results and preset performance index requirements, it adjusts the operating parameters of the heavy-duty air flotation floor in real time, and maintains the best vibration reduction effect and load-bearing capacity of the system through continuous optimization.

[0017] The adaptive parameter adjustment model is structured as a multilayer perceptron based on a hierarchical attention network architecture, consisting of an input layer, three hidden layers, and an output layer. Each hidden layer contains 256 neurons. It employs the ReLU activation function and batch normalization technique, and achieves adaptive weight allocation for features at different abstraction levels through a hierarchical fusion weight mechanism.

[0018] Among them, the gated weight function is used to adjust the hierarchical fusion weight of the adaptive parameter adjustment model. The gated weight function calculates a balance value based on the comprehensive performance evaluation index of the air flotation system, the load change rate parameter, the vibration amplitude data and the air pressure stability parameter. When the balance value belongs to different intervals, the corresponding weight adjustment function is used to adjust the hierarchical fusion weight.

[0019] The comprehensive performance evaluation index of the air flotation system is calculated by weighted combination of bearing capacity performance factor, vibration reduction performance factor and system stability factor. When the comprehensive performance evaluation index of the air flotation system belongs to different ranges, the system performance status is different and corresponding parameter adjustment strategies need to be adopted.

[0020] This invention establishes a system dynamics model and a set of performance optimization equations for heavy-duty air-floored pavement, constructing a precise mathematical framework describing the system's physical characteristics and parameter relationships. It replaces subjective human experience with objective mathematical calculations, achieving scientific and precise optimization of vibration reduction parameters and overcoming the technical shortcomings of traditional human experience methods. This invention employs an intelligent analysis method using multi-dimensional feature datasets and adaptive parameter adjustment models. It collects system operation data in real time through pressure sensor arrays and vibration sensor networks, and uses deep learning algorithms to automatically identify the system's nonlinear dynamic characteristics and parameter patterns. This eliminates reliance on human experience and achieves parameter optimization based on objective data and intelligent algorithms, significantly improving the accuracy and reliability of parameter settings. The gated weighting function and closed-loop feedback control system established in this invention provide an adaptive adjustment mechanism for parameter optimization. It can automatically correct parameter configurations based on real-time monitoring data, maintaining the system's optimal performance state through continuous intelligent adjustment, thus completely solving the technical problem that the optimization of vibration reduction parameters for heavy-duty air-floored pavement often relies on inaccurate human experience. Attached Figure Description

[0021] Figure 1 This is a flowchart of the method of the present invention.

[0022] Figure 2 This is a schematic diagram of the neural network structure of the adaptive parameter adjustment model involved in the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0024] like Figure 1 The diagram shown is a flowchart of a method for optimizing vibration reduction parameters of heavy-duty air-floor pavement provided by the present invention. This method includes the following steps:

[0025] S01. Establish a system dynamic model for heavy-duty air-floor pavement. By analyzing the interaction between the air-floor bearing mechanism and vibration reduction characteristics, construct a mathematical model describing the coupling relationship between the bearing capacity and vibration reduction performance of the air-floor system. The mathematical model considers the comprehensive influence of key factors such as air pressure distribution, load weight, vibration transmission and system damping.

[0026] S02. Collect operational data of heavy-duty air-floor pavement under different working conditions through pressure sensor array and vibration sensor network. The pressure sensor array measures air pressure distribution data, the vibration sensor network measures vibration frequency data and vibration amplitude data, the load sensor measures load weight data, and the damping test device measures system damping data. Establish a multi-dimensional feature dataset containing air pressure distribution data, load weight data, vibration frequency data, vibration amplitude data, and system damping data.

[0027] S03. An adaptive parameter adjustment model is used to identify and optimize the nonlinear dynamic characteristics of the air flotation system. The adaptive parameter adjustment model receives a multi-dimensional feature dataset as input, adjusts the hierarchical fusion weights to balance the influence of different physical parameters on the system performance, and outputs the optimized system parameter configuration to realize the intelligent adjustment of the air flotation system parameters.

[0028] S04. Establish a set of performance optimization equations for the air flotation system to jointly optimize the system's load-bearing capacity and vibration reduction effect. The set of performance optimization equations for the air flotation system receives air pressure distribution data, load weight data, vibration frequency data, vibration amplitude data, and system damping data as inputs. By solving multiple interrelated mathematical equations, the optimal combination of system operating parameters is determined, and the maximum load-bearing capacity parameter, optimal damping parameter, and coupling configuration parameter are output to achieve the coordinated unity of load-bearing capacity and vibration reduction performance.

[0029] S05. Calculate the comprehensive performance evaluation index of the air flotation system based on the maximum bearing capacity parameter, the optimal damping parameter and the coupling configuration parameter. Adopt the corresponding parameter adjustment strategy according to the numerical range of the comprehensive performance evaluation index of the air flotation system. When the comprehensive performance evaluation index of the air flotation system is in different numerical ranges, automatically select the corresponding optimization algorithm and adjustment method.

[0030] S06. The hierarchical fusion weight allocation of the adaptive parameter adjustment model is dynamically adjusted by a gated weight function. The gated weight function receives the comprehensive performance evaluation index of the air flotation system as input, and automatically adjusts the air pressure distribution ratio, system damping parameters and control gain parameters according to the real-time monitored system status information and performance requirements, and outputs the adjusted hierarchical fusion weight.

[0031] S07. Establish a closed-loop feedback control system for the vibration reduction parameters of the air-floor pavement. The closed-loop feedback control system receives the adjusted hierarchical fusion weights and the comprehensive performance evaluation index of the air-floor system as inputs. Based on the system performance evaluation results and preset performance index requirements, it adjusts the operating parameters of the heavy-duty air-floor pavement in real time, and maintains the best vibration reduction effect and load-bearing capacity of the system through continuous optimization.

[0032] The heavy-duty air-floored pavement refers to a ground support system with a load-bearing capacity exceeding 10 tons, employing the principle of gas static pressure bearings to achieve load support and vibration isolation. The system dynamics model is a mathematical model describing the dynamic response characteristics and internal parameter changes of the air-floored system under external excitation. The air-floored load-bearing mechanism refers to the working principle of supporting the load by forming a static pressure field in the air-floor gap using compressed gas. The vibration reduction characteristics refer to the air-floored system's ability to isolate and attenuate external vibration excitation. The coupling relationship refers to the physical relationship of mutual influence and constraint between the air-floored system's load-bearing capacity and vibration reduction performance. The pressure sensor array refers to a combination of multiple pressure measuring devices arranged according to a spatial distribution pattern, used to monitor the internal air pressure distribution state of the air-floored system in real time. The vibration sensor network refers to a distributed monitoring system composed of multiple vibration measuring devices, used to acquire vibration response information of the heavy-duty air-floored pavement at different locations and directions. The multidimensional feature dataset refers to a data set containing multiple key physical parameters during the operation of the air-floored system, used to describe the system's comprehensive operating state and performance characteristics. The hierarchical fusion weight refers to the weight coefficients in the adaptive parameter adjustment model used to balance the importance of features at different abstract levels. The air pressure distribution ratio refers to the distribution ratio of air pressure values ​​in different areas of the air flotation system. The control gain parameter refers to the gain coefficient used to adjust the control response intensity in a closed-loop feedback control system.

[0033] The performance optimization equations for the air flotation system include a load-bearing capacity balance equation, a vibration reduction characteristic equation, and a parameter coupling equation. The load-bearing capacity balance equation is used to calculate the optimal air pressure distribution and load-bearing capacity of the air flotation system under a given load condition. The inputs include load weight data, air flotation area parameters, air pressure distribution data, air flotation gap height parameters, and friction coefficient parameters. The outputs are the maximum load-bearing capacity parameters and the optimized air pressure distribution parameters. The vibration reduction characteristic equation is used to analyze the vibration transmission characteristics and optimize the vibration reduction effect of the air flotation system. The inputs include vibration frequency data, vibration amplitude data, system damping data, air pressure distribution data, and elastic coefficient parameters. The outputs are the vibration transmissibility parameters and the optimal damping parameters. The parameter coupling equation is used to describe the interaction between load-bearing capacity and vibration reduction performance. The inputs include air pressure coupling coefficient parameters, load change rate parameters, vibration response coefficient parameters, and system stability factor parameters. The outputs are the coupling configuration parameters and the system comprehensive performance index parameters.

[0034] The specific structure of the adaptive parameter adjustment model is a multilayer perceptron based on a hierarchical attention network architecture, including an input layer, three hidden layers, and an output layer. Each hidden layer contains 256 neurons. It employs the ReLU activation function and batch normalization, and uses a hierarchical fusion weight mechanism to achieve adaptive weight allocation for features at different abstract levels. The steps for establishing the training dataset for the adaptive parameter adjustment model specifically include collecting operational data of the air-bearing system under different working conditions, including air pressure distribution data, load weight data, vibration frequency data, vibration amplitude data, and system damping data, to establish a training dataset containing 50,000 samples. 70% of the data is used for training, 20% for validation, and 10% for testing. The data undergoes normalization and noise filtering to improve model training effectiveness. The training steps for the adaptive parameter adjustment model specifically include using the Adam optimization algorithm for gradient descent optimization, setting the learning rate to 0.001, the batch size to 128, and the training period to 200 epochs. An early stopping mechanism is used to prevent overfitting. Model performance is evaluated through cross-validation, and finally, the model parameters with the smallest validation set error are selected as the optimal model configuration.

[0035] The gated weight function is used to adjust the hierarchical fusion weight of the adaptive parameter adjustment model. The gated weight function calculates an equilibrium value based on the comprehensive performance evaluation index of the air flotation system, the load change rate parameter, the vibration amplitude data, and the air pressure stability parameter. When the equilibrium value belongs to the interval [0, 0.3), a linear weight adjustment function is used to adjust the hierarchical fusion weight. When the equilibrium value belongs to the interval [0.3, 0.7], a nonlinear weight adjustment function is used to adjust the hierarchical fusion weight. When the equilibrium value belongs to the interval (0.7, 1.0], an exponential weight adjustment function is used to adjust the hierarchical fusion weight.

[0036] The comprehensive performance evaluation index of the air flotation system is calculated by weighted combination of bearing capacity performance factor, vibration reduction performance factor and system stability factor. When the comprehensive performance evaluation index of the air flotation system is in the range of [0, 0.4], the system performance is excellent and no parameter adjustment is required. When the comprehensive performance evaluation index of the air flotation system is in the range of (0.4, 0.7], the system performance is average and some parameters need to be optimized. When the comprehensive performance evaluation index of the air flotation system is in the range of (0.7, 1.0], the system performance is poor and the operating parameters need to be fully adjusted and the system configuration needs to be re-optimized.

[0037] The specific implementation methods of the above steps are described in detail below.

[0038] The specific implementation of step S01 involves establishing a system dynamic model for the heavy-duty air-floor pavement. The purpose of this step is to construct a mathematical framework describing the multi-physics coupling characteristics of the air-floor system. First, the system's equations of motion are established using Lagrange dynamics principles. By analyzing the flow characteristics of compressed gas within the gaps in the air-floor bearing mechanism, the Reynolds equation is used to describe the air film pressure distribution. A bearing capacity calculation model is established by combining load balance conditions and boundary constraints. Next, linearization theory and modal analysis methods are used to study the system's vibration characteristics. Damping and stiffness coefficients are introduced to describe the vibration reduction performance, and a vibration transfer function model is established. Then, multivariable coupling analysis is used to establish the interaction between bearing capacity and vibration reduction performance. A coupling coefficient matrix is ​​introduced to describe the influence of air pressure distribution on the system's stiffness and damping characteristics. Finally, a comprehensive mathematical model is formed that includes key factors such as air pressure distribution, load weight, vibration transmission, and system damping.

[0039] The specific implementation of step S02 involves collecting operational data of the heavy-duty air-floor pavement using multi-sensor fusion technology. The purpose of this step is to establish a data foundation encompassing the system's multi-dimensional physical parameters. First, a pressure sensor array is installed on the air-floor pavement surface in a grid layout, with a sensor spacing of 50–100 mm and a sampling frequency of 1000 Hz. Air pressure distribution data is acquired in real-time via a data acquisition card. Simultaneously, a vibration sensor network of at least 16 triaxial accelerometers is arranged around the perimeter and center of the pavement, with a sampling frequency of 2000 Hz to acquire vibration frequency and amplitude data. Next, strain gauge load cells are installed to measure load weight data, with a sensor range set to 1.2 times the maximum load value and an accuracy class of 0.1. Finally, a damping test device is used to measure system damping data using the free vibration decay method, with an excitation frequency range of 1–500 Hz. All collected data undergoes time synchronization processing and filtering preprocessing to establish a multi-dimensional feature dataset containing air pressure distribution data, load weight data, vibration frequency data, vibration amplitude data, and system damping data.

[0040] The specific implementation of step S03 involves using an adaptive parameter adjustment model to identify and optimize the parameters of the air flotation system. The purpose of this step is to achieve intelligent adaptive adjustment of the system parameters. First, a multi-dimensional feature dataset is input into the adaptive parameter adjustment model. The model employs a deep neural network architecture based on an attention mechanism, extracting multi-scale features from the input data through a feature extraction layer. Next, a hierarchical fusion algorithm is used to weight and fuse features at different abstraction levels. The weight coefficients are dynamically adjusted using a backpropagation algorithm, with the weight values ​​ranging from 0.1 to 0.9. Then, a parameter identification algorithm is used to model the nonlinear dynamic characteristics of the air flotation system. Online parameter identification and updating are achieved using the least squares method and Kalman filtering algorithm. Finally, a multi-objective optimization algorithm is used to optimize the system parameters. The optimization objectives include maximizing bearing capacity and minimizing vibration transmissibility, while the constraints include air pressure stability and system safety requirements. The optimized system parameter configuration is then output.

[0041] The specific implementation of step S04 involves establishing a set of performance optimization equations for the air flotation system for joint optimization. The purpose of this step is to achieve a coordinated balance between load-bearing capacity and vibration reduction performance. First, a load-bearing capacity balance equation is established. This equation is based on hydrostatic bearing theory and air film lubrication theory. Input parameters include load weight data, air flotation area parameters, air pressure distribution data, air flotation gap height parameters, and friction coefficient parameters. The supporting capacity of the air film pressure on the load is calculated using numerical integration methods, and the maximum load-bearing capacity parameters and optimized air pressure distribution parameters are output. Next, a vibration reduction characteristic equation is established. This equation is based on vibration transmission theory and dynamic analysis methods. Input parameters include vibration frequency data, vibration amplitude data, system damping data, air pressure distribution data, and elastic coefficient parameters. The vibration transfer function is calculated using frequency domain analysis methods, and the vibration transmissibility parameters and optimal damping parameters are output. Then, a parameter coupling equation is established. This equation describes the interaction between load-bearing capacity and vibration reduction performance. Input parameters include air pressure coupling coefficient parameters, load change rate parameters, vibration response coefficient parameters, and system stability factor parameters. The influence weights between parameters are determined using coupling analysis methods, and the coupling configuration parameters and system comprehensive performance index parameters are output.

[0042] The specific implementation of step S05 involves calculating the comprehensive performance evaluation index of the air-bearing system and adopting corresponding parameter adjustment strategies. The purpose of this step is to establish a system performance evaluation and adjustment strategy selection mechanism. First, based on the maximum bearing capacity parameter, optimal damping parameter, and coupling configuration parameters, the bearing capacity performance factor, vibration reduction performance factor, and system stability factor are calculated. The bearing capacity performance factor is calculated based on the ratio of actual bearing capacity to theoretical bearing capacity; the vibration reduction performance factor is calculated based on the reciprocal of the vibration transmissibility; and the system stability factor is calculated based on the convergence index of the system response. Next, the comprehensive performance evaluation index of the air-bearing system is calculated using a weighted average method, with the weight allocation ratio set to 0.4 for the bearing capacity performance factor, 0.4 for the vibration reduction performance factor, and 0.2 for the system stability factor. Then, according to the numerical range of the comprehensive performance evaluation index, corresponding parameter adjustment strategies are selected. When the index is in the range of 0 to 0.4, the system performance is excellent and no parameter adjustment is required. When the index is in the range of 0.4 to 0.7, a local optimization algorithm is used to adjust some parameters. When the index is in the range of 0.7 to 1.0, a global optimization algorithm is used to reconfigure all parameters.

[0043] The specific implementation of step S06 involves dynamically adjusting the hierarchical fusion weight allocation of the adaptive parameter adjustment model using a gated weight function. The purpose of this step is to achieve adaptive adjustment of the weight allocation. First, the gated weight function receives the comprehensive performance evaluation index of the air flotation system, load change rate parameters, vibration amplitude data, and air pressure stability parameters as input. It calculates the equilibrium value using a weighted summation method, with weight coefficients set to 0.4, 0.2, 0.2, and 0.2, respectively. Next, the appropriate weight adjustment function is selected based on the interval to which the equilibrium value belongs. When the equilibrium value is in the 0–0.3 interval, a linear weight adjustment function is used, with an adjustment range of ±10% of the original weight. When the equilibrium value is in the 0.3–0.7 interval, a nonlinear weight adjustment function is used, with an adjustment range of ±20% of the original weight. When the equilibrium value is in the 0.7–1.0 interval, an exponential weight adjustment function is used, with an adjustment range of ±30% of the original weight. Then, based on the real-time monitored system status information and performance requirements, the air pressure distribution ratio, system damping parameters, and control gain parameters are automatically adjusted. The adjustment range of the air pressure distribution ratio is set to 80% to 120% of the standard value, the adjustment range of the system damping parameters is set to 50% to 150% of the standard value, and the adjustment range of the control gain parameters is set to 70% to 130% of the standard value. Finally, the adjusted hierarchical fusion weights are output.

[0044] The specific implementation of step S07 involves establishing a closed-loop feedback control system for the vibration reduction parameters of the air-floor pavement. The purpose of this step is to achieve continuous optimization and automatic adjustment of the system parameters. First, a feedback control structure is established, using a proportional-integral-derivative (PID) controller as the main controller. The proportional gain is set to 1.2, the integral time constant to 0.5 seconds, and the derivative time constant to 0.1 seconds. Next, a parameter adjustment mechanism is established. The system calculates the control deviation based on the comprehensive performance evaluation index of the air-floor system and preset performance requirements. The performance threshold is set to 0.3. When the actual performance index exceeds the threshold, the parameter adjustment program is initiated. Then, an adaptive control algorithm is used to achieve real-time parameter adjustment. The algorithm is based on the model reference adaptive control principle, identifying system parameter changes online and adjusting the controller parameters. The parameter update step size is set to 0.01, and the convergence criterion is that the parameter change rate is less than 1%. Finally, a performance monitoring and evaluation mechanism is established. The system calculates the comprehensive performance evaluation index every 10 seconds. When the results of five consecutive calculations meet the performance requirements, the system is considered to have reached its optimal state. Continuous optimization maintains the system's optimal vibration reduction effect and load-bearing capacity.

[0045] like Figure 2 As shown, the detailed structure of the adaptive parameter adjustment model adopts a multilayer perceptron design based on a hierarchical attention network architecture. This model consists of a five-layer structure: an input layer, three hidden layers, and an output layer. The input layer receives a multidimensional feature vector containing air pressure distribution data, load weight data, vibration frequency data, vibration amplitude data, and system damping data, with a data dimension of 128. The first hidden layer contains 256 neurons, employs the ReLU activation function for nonlinear transformation, and applies batch normalization to improve training stability. The second hidden layer also contains 256 neurons, introducing an attention mechanism to weight the features; the attention weights are calculated using a learnable parameter matrix. The third hidden layer contains 256 neurons, employs a residual connection structure to avoid the gradient vanishing problem, and uses dropout to prevent overfitting, with a dropout ratio set to 0.3. The output layer contains 64 neurons, outputting the optimized system parameter configuration through a fully connected layer. The hierarchical fusion weight mechanism introduces learnable weight parameters between each hidden layer to achieve adaptive weight allocation for features at different abstract levels. The weights are normalized using a softmax function to ensure that the weight sum is 1.

[0046] The detailed steps for establishing the training dataset for the adaptive parameter adjustment model include four stages: data collection, data preprocessing, dataset partitioning, and data augmentation. In the data collection stage, a heavy-duty air-floor pavement system was operated under different conditions, including load variations from 5 to 15 tons, vibration frequencies from 1 Hz to 500 Hz, and air pressure variations from 0.2 MPa to 0.8 MPa. Data was continuously collected for 30 minutes under each condition, with a sampling interval of 0.01 seconds, ultimately resulting in a raw dataset containing 50,000 samples. In the data preprocessing stage, outlier detection and removal were first performed on the raw data. The 3σ criterion was used to identify outlier data points, and the proportion of outliers was controlled to within 5%. Next, data normalization was performed, using the maximum-minimum normalization method to scale the data to the 0-1 range, ensuring consistent numerical ranges across different physical quantities. Then, a moving average filter was used for noise filtering, with a filter window length of 10 sampling points, to remove high-frequency noise components. In the dataset partitioning phase, the preprocessed dataset is divided into training, validation, and test sets at proportions of 70%, 20%, and 10%, respectively, to ensure a uniform distribution of samples within each subset. In the data augmentation phase, Gaussian-distributed noise signals are added to the original data to augment the training samples. The noise amplitude is set to 5% of the signal amplitude, resulting in an augmented training set of 70,000 samples.

[0047] The adaptive parameter-adjusted model training process employs the Adam optimization algorithm for gradient descent optimization. This algorithm combines the advantages of momentum gradient descent and adaptive learning rate adjustment, effectively handling sparse gradients and non-stationary objective functions. The learning rate is set to 0.001, the momentum parameter to 0.9, and the exponential decay rate for second-moment estimation to 0.999. The batch size is set to 128, meaning 128 samples are used to calculate the gradient each time model parameters are updated. The training cycle is set to 200 epochs, with each epoch traversing the entire training set. To prevent overfitting, an early stopping mechanism is used to monitor the loss function value on the validation set; training is terminated early when the loss on the validation set no longer decreases after 20 consecutive epochs. The loss function uses a combination of mean squared error loss and a regularization term, with the regularization coefficient set to 0.0001. Model performance is evaluated using a 5-fold cross-validation method. The training set is divided into 5 subsets, with 4 subsets used for training and 1 subset used for validation in rotation. The average of the 5 validation results is calculated as the model performance metric. Finally, the model parameters with the smallest validation set error were selected as the optimal model configuration.

[0048] The adaptive parameter adjustment model possesses the ability to handle multivariable nonlinear coupled systems and adaptive learning characteristics. Compared to traditional proportional-integral-derivative (PID) control methods, which are based on linear control theory and struggle to handle the complex nonlinear coupling relationships between air pressure distribution, load changes, and vibration response in air-float systems, the control effect is limited by model accuracy and parameter fixity. The adaptive parameter adjustment model, however, leverages the nonlinear mapping capabilities of deep neural networks to learn and characterize the complex dynamic characteristics of the system. Through a hierarchical fusion weight mechanism, it adaptively adjusts the influence weights of different physical parameters, achieving intelligent parameter adjustment. Compared to fuzzy logic-based control methods, which rely on expert experience to formulate fuzzy rules (the number of rules increases exponentially with the dimensionality of input variables and cannot automatically optimize fuzzy membership functions), the adaptive parameter adjustment model automatically learns system characteristics through a data-driven approach, avoiding the subjectivity of manually formulated rules. It achieves automatic parameter adjustment and optimization through attention mechanisms and gradient optimization algorithms. The model's advantages lie in its ability to handle high-dimensional multivariable inputs, its strong nonlinear fitting capability, and its ability to adapt to system parameter changes through an online learning mechanism, achieving intelligent and automated parameter adjustment.

[0049] The key technical ideas of this invention include four aspects: multi-physics coupling modeling, adaptive intelligent parameter adjustment, dynamic gating weight allocation, and closed-loop feedback optimization control. The multi-physics coupling modeling technology, compared to traditional single-physics modeling methods, comprehensively describes the interaction relationships between multiple physical fields such as aerodynamics, mechanics, and vibration in the air-bearing system. By establishing a set of coupled equations, it accurately reflects the actual operating characteristics of the system, avoiding modeling errors caused by neglecting the coupling effects between physical fields in traditional methods, thus improving the accuracy and reliability of system parameter optimization. The adaptive intelligent parameter adjustment technology, compared to traditional fixed-parameter control methods, automatically adjusts control parameters according to system operating states and environmental changes. Through the learning ability of deep neural networks, it establishes a complex nonlinear mapping relationship between input and output, overcoming the poor adaptability problem caused by fixed parameters in traditional methods, and achieving intelligent adjustment of control parameters and continuous optimization of system performance. The dynamic gating weight allocation technology, compared to traditional static weight allocation methods, dynamically adjusts the importance weights of different parameters according to the real-time state of the system. Through a gating mechanism, it adaptively selects the weight adjustment strategy, avoiding the limitations of static weights in adapting to changes in system state, thus improving the accuracy of parameter adjustment and the speed of system response. The advantage of closed-loop feedback optimization control technology over traditional open-loop control methods lies in its ability to monitor system performance in real time and make feedback adjustments. Through continuous performance evaluation and parameter optimization, it maintains the optimal operating state of the system, overcomes the disadvantage of open-loop control that cannot automatically correct deviations, and ensures long-term stable operation and optimal performance of the system.

[0050] The synergistic effect of these four key technological approaches, compared to existing technologies, results in the construction of a complete intelligent air-floor pavement vibration reduction parameter optimization system. Each approach supports the others, forming an organic whole. Multiphysics coupling modeling provides an accurate mathematical foundation for the system; adaptive parameter intelligent adjustment, based on an accurate model, enables intelligent parameter configuration; dynamic allocation of gating weights further improves the accuracy of parameter adjustment; and closed-loop feedback optimization control ensures continuous optimization of system performance. Compared to the relatively independent and unconnected technical architecture of traditional methods, this invention achieves integrated optimization of the entire process from modeling, adjustment, allocation to control through the synergistic effect of these technological approaches. This significantly improves the system's adaptability, control accuracy, and operational stability, providing a more reliable and efficient technical solution for the engineering application of heavy-duty air-floor pavements.

[0051] Traditional manual adjustment methods often fail to maintain consistency and stability in parameter tuning under varying load conditions, vibration environments, and operational requirements. Adjustments by the same operator at different times can vary, and different operators employ different methods and standards, resulting in a lack of reproducibility and predictability in system performance. This invention, by establishing a unified mathematical model and standardized algorithm flow, ensures that the same scientific methods can be used for parameter optimization under various operating conditions. A comprehensive performance evaluation index provides a unified quantitative standard for parameter tuning under different conditions, and a closed-loop feedback control system guarantees the automation and standardization of the parameter tuning process, completely resolving the technical shortcomings of poor consistency in manual adjustments.

[0052] Traditional methods require experienced technicians to find suitable parameter combinations through repeated trials and gradual adjustments. The entire optimization process is time-consuming and inefficient, especially when there are many system parameters and complex operating conditions, where manual debugging often takes several days or even weeks. This invention employs intelligent parameter identification and automated optimization algorithms, enabling the calculation and evaluation of a large number of parameter combinations in a short time. The adaptive parameter adjustment model significantly improves the speed of parameter optimization through parallel computing, while the gated weight function and closed-loop feedback control system achieve real-time automatic parameter adjustment. This reduces the traditional manual debugging cycle from several days to several hours or even less, significantly improving the efficiency and response speed of parameter optimization.

[0053] Specifically, the principle of this invention is as follows: The core principle that enables this invention to solve the technical problem of insufficient accuracy in optimizing vibration reduction parameters for heavy-duty air-floor pavement lies in replacing unreliable human experience-based judgment with precise mathematical modeling and intelligent algorithms. Through a system dynamics model, the physical mechanism of air-floor pavement is transformed into calculable mathematical equations, providing a scientific theoretical basis for parameter optimization. The logical rationality of the technical solution is reflected in the use of multi-dimensional feature datasets to comprehensively obtain objective operational information of the system, the application of an adaptive parameter adjustment model to intelligently analyze the parameter patterns implicit in the data, and the precise calculation of the optimal parameter combination through the air-floor system performance optimization equation set, thus achieving a fundamental shift from subjective experience to objective science.

[0054] The adaptive parameter adjustment model, based on a hierarchical attention network architecture, can automatically learn and identify the nonlinear dynamic characteristics of the air-bearing system. Through machine learning on a large amount of training data, it acquires more accurate and comprehensive parameter optimization knowledge than human experience. The hierarchical fusion weight mechanism ensures that the influence of different physical parameters is accurately quantified and balanced. The performance optimization equations of the air-bearing system are solved simultaneously by the load-bearing capacity balance equation, vibration reduction characteristic equation, and parameter coupling equation, transforming the complex multi-parameter optimization problem into a rigorous mathematical calculation process, eliminating the subjectivity and arbitrariness of human judgment.

[0055] The gating weighting function automatically selects the most suitable parameter adjustment strategy based on the comprehensive performance evaluation index. The application of linear, nonlinear, and exponential weighting functions across different operating conditions ensures the most accurate optimization algorithm is used under various circumstances, avoiding judgment errors caused by human experience in complex situations. The closed-loop feedback control system forms a continuous intelligent optimization mechanism through real-time data monitoring and automatic parameter adjustment. It can dynamically track changes in system state and correct parameter deviations in a timely manner, ensuring the continuous accuracy of parameter optimization. This technical solution, based on precise modeling, intelligent algorithms, and automatic control, fundamentally eliminates the uncertainty and subjectivity of human experience, achieving precise optimization of vibration reduction parameters for heavy-duty air-floor flooring.

[0056] The following provides a specific embodiment 1 of the present invention, and the specific implementation of each step in this embodiment 1 is described in detail below.

[0057] The specific implementation of step S01 involves establishing a system dynamic model for the heavy-duty air-floor pavement. This model includes the coupling relationship between the air-floor bearing mechanism and vibration reduction characteristics. The system dynamic model is expressed as follows:

[0058]

[0059] In the formula, M is the equivalent mass matrix of the system, in kg; C is the damping matrix of the system, in N·s / m; K is the stiffness matrix of the system, in N / m; and z is the displacement vector of the system, in m. This is the system velocity vector, in m / s. This is the system acceleration vector, in m / s². 2 ;F load F is the external load force vector, in N; air This represents the air buoyancy support force vector, with units of N. The formula for calculating the air buoyancy support force is:

[0060] F air =∫ A p(x,y)dA;

[0061] In the formula, p(x, y) is the air film pressure distribution function, in Pa; A is the air flotation area, in m². 2 x and y are the coordinate variables of the air-floating surface, in meters (m). The air film pressure distribution is calculated based on the Reynolds equation:

[0062]

[0063] In the formula, h is the gas film thickness, in meters (m), ranging from 10 to 100 μm; μ is the aerodynamic viscosity, in Pa·s, and is taken as 1.8 × 10⁻⁶ under standard conditions. -5 Pa·s; Q is the gas supply flow rate, in m³ / s. 3 / s, with a value range of 0.001 to 0.01m 3 / s. This equation needs to be solved with boundary conditions, namely, the pressure at the edge of the air flotation is equal to atmospheric pressure.

[0064] The specific implementation methods for steps S02-S03 are the same as those described above, and will not be repeated in detail here.

[0065] The specific implementation of step S04 involves establishing a set of performance optimization equations for the air flotation system, including the load-bearing capacity balance equation, vibration reduction characteristic equation, and parameter coupling equation. The load-bearing capacity balance equation is expressed as:

[0066]

[0067] In the formula, F max Maximum load-bearing capacity, in N; L x L y The length and width of the air flotation floor are in meters (m); p opt (x, y) is the optimized pressure distribution function, in Pa; μ f ρ is the friction coefficient, dimensionless; W is the load weight, in N. The formula for calculating the optimized air pressure distribution is:

[0068]

[0069] In the formula, p0 is the reference pressure, in Pa, ranging from 200,000 to 800,000 Pa; α1 and α2 are pressure adjustment coefficients, dimensionless, ranging from -0.5 to 0.5; ε p The pressure distribution error term is expressed in Pa. The parameters are obtained as follows: α1 and α2 are obtained experimentally, including step 1: measuring the bearing capacity change under different pressure distribution patterns; step 2: obtaining the adjustment coefficient through least squares fitting; ε p The range is -50 to 50 Pa. The vibration reduction characteristic equation is expressed as:

[0070]

[0071] In the formula, T(ω) is the vibration transmissibility, dimensionless; ω is the vibration frequency, in rad / s; and H(ω) is the system transfer function. The formula for calculating the system transfer function is:

[0072]

[0073] In the formula, A ref For reference area, the unit is m. 2 ; i is the imaginary unit, dimensionless; M eq Equivalent mass per unit area, in kg / m² 2 C eq The equivalent damping per unit area is expressed in N·s / m. 3 ;K eq Equivalent stiffness per unit area, in N / m 3 K eq =K / A ref +K air / A ref A ref The value is taken as the total area of ​​the air-supported floor. The formula for calculating the air film stiffness is:

[0074]

[0075] In the formula, Φ is the partial derivative of the film pressure with respect to the thickness, expressed in Pa / m. The parameter coupling equation is expressed as: Φ coupling =β1ξ pressure +β2ξ vibration +β3ξ load +β4R stability ;

[0076] In the formula, Φ coupling ξ is a coupling configuration parameter, dimensionless; β1, β2, β3, β4 are coupling weight coefficients, dimensionless, ranging from 0 to 1, and β1+β2+β3+β4=1; pressureξ is the pressure coupling coefficient parameter, dimensionless; vibration ξ is a dimensionless parameter representing the vibration response coefficient. load R is the load change rate parameter, in N / s. stability This is the system stability factor parameter, dimensionless. Note that R here... stability With the system stability factor I in step S05 stability Different, R stability I is the stability parameter in the coupling analysis. stability This is the stability factor used in performance evaluation. The formula for calculating the pressure coupling coefficient parameter is:

[0077]

[0078] In the formula, p max This is the maximum air pressure value, in Pa; p min This represents the minimum air pressure value, expressed in Pa. The formula for calculating the vibration response coefficient parameter is:

[0079]

[0080] In the formula, C max C is the maximum damping coefficient, expressed in N·s / m; min The minimum damping coefficient is expressed in N·s / m. The load change rate parameter is calculated from the time-series data of the load sensor.

[0081]

[0082] In the formula, W max Maximum load weight, in N; W min Minimum load weight, in N; Let σ be the derivative of the load weight with respect to time, in N / s. response The variance of the system output signal is obtained by statistical analysis of the system output response; s poles The system poles are obtained by using a system identification method and by analyzing the characteristic equation of the system transfer function.

[0083] The specific implementation of step S05 is to calculate the comprehensive performance evaluation index of the air flotation system. The calculation formula is expressed as: I performance =γ1I bearing +γ2I damping +γ3I stability ;

[0084] In the formula, I performance I is a dimensionless index for evaluating the comprehensive performance of air flotation systems, ranging from 0 to 1. bearing I is the load-bearing capacity performance factor, dimensionless, with a value range of 0 to 1;damping I is the vibration reduction performance factor, dimensionless, with a value range of 0 to 1; stability γ1 is the system stability factor, dimensionless, ranging from 0 to 1; γ2, γ3 are weighting coefficients, dimensionless, ranging from 0 to 1, and γ1 + γ2 + γ3 = 1, with typical values ​​of γ1 = 0.4, γ2 = 0.4, and γ3 = 0.2. The formula for calculating the bearing capacity performance factor is:

[0085]

[0086] In the formula, F actual The actual load-bearing capacity is expressed in N, with a range of 50,000 to 150,000 N; F theoretical δ1 is the theoretical bearing capacity, in N, with a value ranging from 50,000 to 150,000 N; δ1 is the bearing capacity correction factor, dimensionless, with a typical value of 2.0; p var This represents the pressure variance, in Pa. 2 ;p mean The value represents the average air pressure, expressed in Pa, and ranges from 200,000 to 800,000 Pa. The formula for calculating the vibration reduction performance factor is:

[0087]

[0088] In the formula, T average δ is the average vibration transmissibility, dimensionless; δ2 is the vibration reduction correction coefficient, dimensionless; σ T Let be the standard deviation of vibration transmissibility, which is dimensionless. The system stability factor is calculated using the following formula:

[0089]

[0090] In the formula, ρ1 is the stability correction coefficient, which is dimensionless and ranges from 1 to 3; σ response denoted as σ0, representing the system response variance, dimensionless; κ is the pole weighting coefficient, dimensionless, ranging from 0.1 to 0.5; s poles Here is the system pole vector, in rad / s; ||s poles || represents the norm of the system's pole vector, expressed in rad / s.

[0091] The specific implementation of step S06 is to dynamically adjust the hierarchical fusion weights through a gating weight function, which is expressed as follows:

[0092]

[0093] In the formula, G(t) is the gating equilibrium value, which is dimensionless; t is the time variable, with the unit being seconds; R load (t) is the load change rate parameter, in N / s. Note that R here... load(t) is the time-dependent load change rate function, and ξ in step S04 load They are the same type of parameter but expressed in different ways. A vibration (t) represents the vibration amplitude data, in meters (m); S pressure (t) represents the pressure stability parameter, dimensionless; λ1, λ2, and λ3 are normalization coefficients, dimensionless. The parameters are obtained by standardizing λ1, λ2, and λ3. λ3=1, where R load，max A represents the maximum rate of change of load. vibration，max This represents the maximum amplitude of the vibration. The weighting adjustment function is selected based on the range of G(t): when G(t) ∈ [0, 0.3), a linear weighting adjustment function is used.

[0094] w new =w old +η1G(t);

[0095] When G(t)∈[0.3, 0.7], a nonlinear weight adjustment function is used:

[0096] w new =w old +η2sin(πG(t));

[0097] When G(t)∈(0.7, 1.0], the exponential weighting adjustment function is used:

[0098] w new =w old +η3e 2G(t)-1 ;

[0099] In the formula, w new The adjusted weights are dimensionless and range from 0 to 1; w old η1 represents the original weight, which is dimensionless and ranges from 0 to 1; η2, η3 are adjustment step size coefficients, also dimensionless, with typical values ​​of η1 = 0.1, η2 = 0.2, and η3 = 0.3, respectively. The weight adjustment range is adaptively adjusted according to different intervals of G(t) value to ensure the smoothness and stability of weight changes.

[0100] The specific implementation of step S07 is to establish a closed-loop feedback control system. The control algorithm uses a proportional-integral-derivative controller, and the control output is expressed as:

[0101]

[0102] In the formula, u(t) is the control output, dimensionless, with a value range of -1 to 1; e(t) is the control deviation, dimensionless, with a value range of -1 to 1; K pK is a proportionality constant, dimensionless, with a typical value of 1.2; i K is the integral coefficient, in units of 1 / s, with a typical value of 2.0; d τ is the differential coefficient, in units of seconds (s), with a typical value of 0.1; τ is the integral variable, in units of seconds. The formula for calculating control deviation is:

[0103] e(t) = I target -I performance (t);

[0104] In the formula, I target The target performance index is dimensionless and has a value of 0.3. The control deviation e(t) represents the difference between the actual system performance and the target performance. When e(t) > 0, it means that the system performance is lower than the target requirement and the control output needs to be increased; when e(t) < 0, it means that the system performance exceeds the target requirement and the control output needs to be reduced.

[0105] It should be explained that the principle of the system dynamics model equations is based on Lagrange dynamics theory and the principle of gas hydrostatic bearings. By establishing mathematical relationships between mass, damping, and stiffness, it describes the dynamic response characteristics of the air-bearing system. Compared with traditional static analysis methods, this model can accurately reflect the transient response of the system under dynamic loads, providing a time-domain analysis basis for parameter optimization and significantly improving the accuracy of vibration reduction parameter design. The application principle of the Reynolds equation is based on the thin-film lubrication theory in fluid mechanics. By describing the spatial variation law of pressure distribution within the air film, this equation, compared with the simplified uniform pressure assumption, can accurately calculate the spatial distribution of air-bearing capacity, providing a theoretical basis for load-bearing capacity optimization and effectively improving the accuracy and reliability of load-bearing capacity calculation. The principle of the load-bearing capacity balance equation is based on mechanical equilibrium theory and pressure integration method. By introducing an optimized air pressure distribution function, this equation, compared with the traditional uniform air pressure scheme, can achieve the maximum configuration of load-bearing capacity, significantly improving the load-bearing performance and load distribution uniformity under heavy load conditions. The principle of the vibration reduction characteristic equation is based on linear system theory and frequency domain analysis. By establishing the mathematical relationship between vibration transmissibility and system parameters, this equation, compared to empirical adjustment methods, can quantitatively analyze the relationship between vibration reduction effect and parameter changes, providing theoretical guidance for precise adjustment of vibration reduction parameters and achieving controllable optimization of vibration reduction performance. The principle of the parameter coupling equation is based on multivariable coupling analysis theory. By introducing partial derivative terms to describe the mutual influence between parameters, this equation, compared to independent parameter adjustment methods, can comprehensively consider the mutual constraints between load-bearing capacity and vibration reduction performance, avoiding performance imbalance caused by single-objective optimization and achieving coordinated and unified system performance. The principle of the comprehensive performance evaluation index is based on multi-objective evaluation theory and weighted fusion methods. By introducing an index correction term, this index can more accurately reflect the comprehensive operating state of the system compared to the simple weighted average method, providing a quantitative basis for the selection of parameter adjustment strategies and improving the scientificity and operability of system performance evaluation. The principle of the gated weight function is based on adaptive control theory and piecewise function design. By selecting different weight adjustment strategies according to the system state, this function can dynamically adjust the importance of parameters according to the system operating state, compared with the fixed weight method, realizing intelligent adjustment of weight allocation and significantly improving the adaptability and accuracy of parameter adjustment.

[0106] To better understand and implement this invention, a specific application scenario, Example 2, is provided below: The technical team first established a system dynamic model for the heavy-duty air-floored pavement. Based on the principles of Lagrange dynamics, a set of motion equations for the air-floored support system in the wind tunnel test section was constructed. Analysis of the air-floored load-bearing mechanism shows that the flow characteristics of compressed gas within the gap follow the Reynolds equation, and the air film pressure distribution function is P(x,y)=P0·f(x,y,h,μ), where P0 is the supply pressure, h is the air-floored gap height, and μ is the gas dynamic viscosity. Combining load balance conditions and boundary constraints, a load-bearing capacity calculation model F was established. c =∫∫P(x,y)dx dy. Vibration characteristic analysis employs linearization theory and modal analysis methods, by introducing a damping coefficient C. d =2.4×10 4 N·s / m and stiffness coefficient K s =3.8×10 7 N / m, a vibration transfer function model was established. Where r is the frequency ratio and ζ is the damping ratio. Multivariate coupling analysis shows that the influence of air pressure distribution on the system stiffness and damping characteristics can be described by the coupling coefficient matrix [K], ultimately forming a comprehensive mathematical model covering key factors such as air pressure distribution, load weight, vibration transmission, and system damping.

[0107] The technical team installed a multi-sensor fusion data acquisition system in the wind tunnel testing section. The pressure sensor array adopted a grid layout, with 480 pressure sensors installed on a 4m × 6m air-bearing surface. The sensor spacing was set to 100mm, the sampling frequency was 1000Hz, the measurement range was 0–1.0MPa, and the accuracy was ±0.1%. The vibration sensor network consisted of 24 triaxial accelerometers, distributed around the perimeter and center of the surface, with a sampling frequency of 2000Hz, a measurement range of ±50g, and a frequency response range of 1–3000Hz. The load sensor adopted a strain gauge structure with a range of 18 tons (1.2 times the maximum load) and an accuracy class of 0.1. The damping testing device used the free vibration decay method, with an excitation frequency range of 1–500Hz, generating a standard excitation signal through an electromagnetic vibrator. After 30 days of continuous monitoring, a multi-dimensional feature dataset was established, including air pressure distribution data, load weight data, vibration frequency data, vibration amplitude data, and system damping data, totaling 52,000 data samples collected.

[0108] The adaptive parameter tuning model employs a multilayer perceptron structure based on a hierarchical attention network architecture. The model consists of an input layer (48 nodes), three hidden layers (256 neurons per layer), and an output layer (12 nodes). The hidden layers use ReLU activation and batch normalization, and the weight coefficients of the hierarchical fusion weight mechanism have a dynamic range of 0.1–0.9. During the training dataset setup, 52,000 samples were used for training, validation, and testing at 70%, 20%, and 10% ratios, respectively. The data underwent Z-score normalization and Gaussian filtering for noise reduction. Model training employed the Adam optimization algorithm with a learning rate of 0.001, a batch size of 128, and a training epoch of 200 epochs. An early stopping mechanism was implemented, stopping training if the validation error did not decrease after 15 consecutive epochs. Cross-validation results showed that the model's mean absolute error on the test set was 0.0283, and the coefficient of determination R0 was [missing data]. 2 =0.9756, which meets the accuracy requirements.

[0109] The technical team established a set of performance optimization equations for the air flotation system for joint optimization. The load-bearing capacity balance equation is based on hydrostatic bearing theory, and the input parameters include a load weight of 15,000 kg and an air flotation area of ​​24 m². 2 The air flotation gap height is 0.05–0.15 mm, and the friction coefficient is 0.002. The maximum bearing capacity parameter, calculated using numerical integration, is 186,000 N. Optimized air pressure distribution parameters show a central region air pressure of 0.62 MPa and an edge region air pressure of 0.48 MPa. The vibration reduction characteristic equation is based on vibration transmission theory, with input parameters including the main vibration frequency of 25–180 Hz, vibration amplitude of 0.1–2.5 μm, system damping ratio of 0.08, and elastic coefficient of 1.2 × 10⁻⁶. 8 N / m 2 Frequency domain analysis yielded a vibration transmissibility parameter of 3.8% and an optimal damping parameter C. opt =2.76×10 4 N·s / m. In the parameter coupling equation, the air pressure coupling coefficient is 0.73, the load change rate parameter is ±12%, the vibration response coefficient is 1.34, the system stability factor is 0.89, and the output coupling configuration parameters and system comprehensive performance index parameters are shown in Table 1.

[0110] Table 1 Optimization results of main performance parameters of the air flotation system

[0111] Parameter type Values ​​before optimization Optimized values Improvement range Maximum load capacity (kN) 156.2 186.0 19.1％ Vibration transmissibility (%) 4.7 3.8 19.1％ air pressure stability 0.81 0.94 16.0％ Response time (s) 1.25 1.08 13.6％ Energy efficiency 0.73 0.82 12.3％ Control accuracy 0.89 0.96 7.9％

[0112] The comprehensive performance evaluation index of the air flotation system was calculated based on the optimization results. The load-bearing capacity performance factor was calculated as the ratio of actual load-bearing capacity to theoretical load-bearing capacity, which is 0.94. The vibration reduction performance factor was calculated as the normalized value of the inverse of the vibration transmissibility, which is 0.92. The system stability factor was calculated based on the response convergence index, which is 0.89. The comprehensive performance evaluation index was calculated using a weighted average method, with the weights allocated as follows: load-bearing capacity performance factor 0.4, vibration reduction performance factor 0.4, and system stability factor 0.2, resulting in a comprehensive evaluation index of 0.92. Since the index is in the range of 0 to 0.4, the system performance is excellent and no parameter adjustment is required. However, to maintain the optimal state, dynamic monitoring is still necessary.

[0113] During the dynamic adjustment of the gating weight function, the input parameters included a comprehensive performance evaluation index of 0.92, a load change rate of 12%, a vibration amplitude of 1.2 μm, and a pressure stability of 0.94. Weight coefficients were set to 0.4, 0.2, 0.2, and 0.2, resulting in a calculated equilibrium value of 0.28. Since the equilibrium value falls within the range of 0 to 0.3, a linear weight adjustment function was used, with an adjustment range of ±10% of the original weights. Real-time monitoring results showed that the pressure distribution ratio was adjusted to 102% in the central region and 96% in the edge region; the system damping parameter was adjusted to 105% of the standard value; and the control gain parameter was adjusted to 98% of the standard value. The adjusted hierarchical fusion weights are shown in Table 2.

[0114] Table 2 Results of Dynamic Adjustment of Hierarchical Fusion Weights

[0115] Network layer Initial weights Adjusted weights Weight changes Input layer 0.25 0.27 +8.0％ Hidden layer 1 0.30 0.28 -6.7％ Hidden layer 2 0.28 0.31 +10.7％ Hidden layer 3 0.17 0.14 -17.6％

[0116] In the establishment of the closed-loop feedback control system, a proportional-integral-derivative (PID) controller was used as the main controller, with a proportional gain of 1.2, an integral time constant of 0.5 s, and a derivative time constant of 0.1 s. The performance index threshold was set to 0.3; when the actual performance index exceeded the threshold, a parameter adjustment program was initiated. The adaptive control algorithm was based on the model reference adaptive control principle, with a parameter update step size of 0.01, and the convergence criterion was that the parameter change rate was less than 1%. The performance monitoring mechanism calculated the comprehensive performance evaluation index every 10 seconds; the system was considered to have reached its optimal state when five consecutive calculations met the performance requirements. Monitoring data from 150 days of actual operation showed that the system's average comprehensive performance evaluation index remained stable within the range of 0.91 to 0.94. When the air-floating platform supported a 15-ton aero-engine model, the vibration transmissibility was consistently controlled below 3.8%, meeting the technical requirements of wind tunnel testing.

[0117] It should be noted that the variables involved in this invention are explained in detail in Table 3.

[0118] Table 3. Variable Explanation Table

[0119]

[0120]

[0121] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for optimizing vibration reduction parameters of heavy-duty air-floor flooring, characterized in that, A system dynamic model of heavy-duty air-floor pavement was established, and a mathematical model describing the coupling relationship between the bearing capacity and vibration reduction performance of the air-floor system was constructed. Operational data of heavy-duty air-floor pavement under different working conditions were collected through pressure sensor array and vibration sensor network, and a multi-dimensional feature dataset containing air pressure distribution data, load weight data, vibration frequency data, vibration amplitude data and system damping data was established. An adaptive parameter adjustment model is used to identify and optimize the nonlinear dynamic characteristics of the air-float system. The influence of different physical parameters on system performance is balanced by adjusting the hierarchical fusion weights. A performance optimization equation set for the air-float system is established to jointly optimize the system's load-bearing capacity and vibration reduction effect, outputting the maximum load-bearing capacity parameter, optimal damping parameter, and coupling configuration parameter. Based on the maximum load-bearing capacity parameter, optimal damping parameter, and coupling configuration parameter, the comprehensive performance evaluation index of the air-float system is calculated. The hierarchical fusion weight allocation of the adaptive parameter adjustment model is dynamically adjusted through a gated weight function. A closed-loop feedback control system for the vibration reduction parameters of the air-float floor is established to adjust the operating parameters of the heavy-duty air-float floor in real time, achieving precise optimization of the vibration reduction parameters.

2. The method for optimizing vibration reduction parameters of heavy-duty air-floor pavement according to claim 1, characterized in that, Heavy-duty air-floored pavement refers to a ground support system with a load-bearing capacity exceeding 10 tons and employing the principle of gas static pressure bearings to achieve load support and vibration isolation. The system dynamics model refers to a mathematical model that describes the dynamic response characteristics and internal parameter variation law of the air-floored system under external excitation. The mathematical model considers the comprehensive influence of air pressure distribution, load weight, vibration transmission, and system damping.

3. The method for optimizing vibration reduction parameters of heavy-duty air-floor pavement according to claim 2, characterized in that, The steps for establishing the system dynamics model are as follows: Specifically, it involves analyzing the interaction between the air-bearing mechanism and the vibration reduction characteristics. The air-bearing mechanism refers to the working principle of supporting the load by forming a static pressure field in the air-bearing gap with compressed gas. The vibration reduction characteristics refer to the air-bearing system's ability to isolate and attenuate external vibration excitation. The coupling relationship refers to the physical relationship between the air-bearing capacity and the vibration reduction performance of the air-bearing system and their mutual influence and constraints.

4. The method for optimizing vibration reduction parameters of heavy-duty air-floor pavement according to claim 3, characterized in that, A pressure sensor array refers to a combination of multiple pressure measuring devices arranged according to spatial distribution patterns, used to monitor the internal air pressure distribution of an air flotation system in real time. The pressure sensor array measures the air pressure distribution data. A vibration sensor network refers to a distributed monitoring system composed of multiple vibration measuring devices, used to acquire vibration response information of heavy-duty air flotation pavement at different locations and directions. The vibration sensor network measures the vibration frequency data and vibration amplitude data.

5. The method for optimizing vibration reduction parameters of heavy-duty air-floor pavement according to claim 4, characterized in that, The steps for establishing a multidimensional feature dataset are as follows: load sensors measure load weight data, damping testing devices measure system damping data, and the multidimensional feature dataset refers to a data set containing multiple key physical parameters during the operation of the air flotation system.

6. The method for optimizing vibration reduction parameters of heavy-duty air-floor pavement according to claim 5, characterized in that, The adaptive parameter adjustment model takes a multidimensional feature dataset as input and outputs an optimized system parameter configuration to achieve intelligent adjustment of the air flotation system parameters. The hierarchical fusion weight refers to the weight coefficients used in the adaptive parameter adjustment model to balance the importance of features at different abstract levels.

7. The method for optimizing vibration reduction parameters of heavy-duty air-floor pavement according to claim 6, characterized in that, The air flotation system performance optimization equation set receives air pressure distribution data, load weight data, vibration frequency data, vibration amplitude data, and system damping data as inputs. By solving multiple interrelated mathematical equations, it determines the optimal combination of system operating parameters to achieve a coordinated balance between load-bearing capacity and vibration reduction performance.

8. The method for optimizing vibration reduction parameters of heavy-duty air-floor pavement according to claim 7, characterized in that, The performance optimization equations for the air flotation system include the load-bearing capacity balance equation, the vibration reduction characteristic equation, and the parameter coupling equation. The load-bearing capacity balance equation is used to calculate the optimal air pressure distribution and load-bearing capacity of the air flotation system under given load conditions. The inputs include load weight data, air flotation area parameters, air pressure distribution data, air flotation gap height parameters, and friction coefficient parameters. The outputs are the maximum load-bearing capacity parameters and the optimized air pressure distribution parameters.

9. The method for optimizing vibration reduction parameters of heavy-duty air-floor pavement according to claim 8, characterized in that, The vibration reduction characteristic equation is used to analyze the vibration transmission characteristics and optimize the vibration reduction effect of the air flotation system. The inputs include vibration frequency data, vibration amplitude data, system damping data, air pressure distribution data, and elastic coefficient parameters. The outputs are vibration transmissibility parameters and optimal damping parameters.

10. The method for optimizing vibration reduction parameters of heavy-duty air-floor pavement according to claim 9, characterized in that, The parametric coupling equation is used to describe the interaction between bearing capacity and vibration reduction performance. The inputs include air pressure coupling coefficient parameters, load change rate parameters, vibration response coefficient parameters, and system stability factor parameters. The outputs are coupling configuration parameters and system comprehensive performance index parameters.

Images