Bridge damper performance detection method and system based on deep learning

Abstract

The application belongs to the technical field of damper performance detection, and discloses a bridge damper performance detection method and system based on deep learning. The method comprises the following steps: constructing a bridge damper performance detection model and a risk item performance detection verification guidance model; setting corresponding real-time performance detection working condition parameters and real-time performance detection environment parameters for real-time bridge damper attribute parameters of the bridge damper; using the bridge damper performance detection model to perform bridge damper performance detection prediction; using the risk item performance detection verification guidance model to perform risk item performance detection verification guidance; performing risk item performance detection verification on the bridge damper; and combining real-time performance detection prediction results and real-time risk item performance detection verification results to obtain real-time performance detection final results of the bridge damper. The application solves the problems of poor performance detection effect, efficiency and cost limitations, and low detection precision in the prior art.

Description

Technical Field

[0001] The present invention belongs to the technical field of damper performance detection, and specifically relates to a method and system for detecting the performance of a bridge damper based on deep learning. Background Art

[0002] Dampers are critical components of bridge structures, reducing vibrations caused by external forces such as wind and earthquakes. Performance testing ensures proper damper operation, thereby safeguarding the structural safety of bridges. Regular damper performance testing can identify potential problems, such as wear and aging, and prevent accidents caused by damper failure. Performance testing also allows for assessment of damper condition, enabling timely maintenance and replacement, thereby extending the lifespan of bridges.

[0003] The existing bridge damper performance testing technology has the following defects:

[0004] 1) Poor performance testing results: Many bridge damper performance tests still rely on traditional manual testing methods, making it difficult to comprehensively and accurately evaluate the actual performance of bridge dampers. Furthermore, testing can often only be performed on specific dampers or specific working conditions, failing to fully cover various complex environments and working conditions, resulting in poor performance testing results.

[0005] 2) Efficiency and cost limitations: Existing bridge damper performance testing methods are time-consuming, labor-intensive, and inefficient. They are difficult to meet the needs of large-scale, high-frequency testing and are costly. Furthermore, as the number of bridges increases, maintenance costs increase exponentially.

[0006] 3) Low detection accuracy: Existing performance detection methods for bridge dampers mainly rely on manual detection and simple instruments and equipment. The performance detection accuracy is not high and the results are not accurate enough. With the increasing complexity of the structure, material and characteristics of bridge dampers, traditional performance detection techniques can no longer meet actual needs. Some hidden and deep defects cannot be discovered in time, posing a safety hazard. Summary of the Invention

[0007] In order to solve the problems of poor performance detection effect, efficiency and cost limitations, and low detection accuracy in the existing technology, the purpose of the present invention is to provide a performance detection method and system for bridge dampers based on deep learning.

[0008] The technical solution adopted in the present invention is:

[0009] A method for detecting performance of a bridge damper based on deep learning comprises the following steps:

[0010] Using deep learning algorithms and reinforcement learning algorithms, a performance detection model for bridge dampers and a risk item performance detection and verification guidance model were constructed;

[0011] According to the performance detection requirements, corresponding real-time performance detection working condition parameters and real-time performance detection environment parameters are set for the real-time bridge damper attribute parameters of the bridge damper to obtain real-time input data;

[0012] According to the real-time input data, the bridge damper performance detection model is used to perform bridge damper performance detection prediction and obtain real-time performance detection prediction results;

[0013] Based on the real-time performance test prediction results, the risk item performance test verification guidance model is used to provide risk item performance test verification guidance and obtain the real-time risk item performance test verification strategy;

[0014] According to the real-time risk item performance test and verification strategy, the risk item performance test and verification of the bridge damper is carried out to obtain the real-time risk item performance test and verification results;

[0015] The final results of the real-time performance test of the bridge damper are obtained by combining the real-time performance test prediction results and the real-time risk item performance test verification results.

[0016] Furthermore, using deep learning algorithms and reinforcement learning algorithms, a bridge damper performance detection model and a risk item performance detection and verification guidance model are constructed, including the following steps:

[0017] Using CAE simulation software, a performance test simulation environment was built. Based on the performance test simulation environment, different damper attribute parameters for historical bridges, historical performance test operating condition parameters, and historical performance test environment parameters were set to perform performance test simulations and obtain several historical performance test simulation results.

[0018] Combining historical bridge damper parameters, historical performance test operating condition parameters, historical performance test environment parameters, and historical performance test simulation results of the same performance test process, a number of model training samples are obtained, and dimensionality reduction processing is performed on the number of model training samples to obtain a number of reduced-dimensionality model training samples;

[0019] Based on several dimensionality reduction model training samples, a deep learning algorithm is used to construct a performance detection model for bridge dampers, and several historical performance detection prediction results are obtained. Based on several historical performance detection prediction results, a reinforcement learning algorithm is used to construct a risk item performance detection verification guidance model.

[0020] Furthermore, the historical bridge damper attribute parameters include historical elastic modulus, historical Poisson's ratio, historical density, historical damping coefficient, historical stiffness, historical maximum damping force, and historical stroke of the bridge damper;

[0021] The real-time bridge damper property parameters include the real-time elastic modulus, real-time Poisson's ratio, real-time density, real-time damping coefficient, real-time stiffness, real-time maximum damping force and real-time stroke of the bridge damper;

[0022] The historical performance test condition parameters include the historical static load, historical dynamic load, historical impact load, historical vehicle load and historical bridge deadweight of the bridge damper;

[0023] Real-time performance testing parameters include real-time static load, real-time dynamic load, real-time impact load, real-time vehicle load and real-time bridge deadweight of bridge dampers;

[0024] The historical performance test environment parameters include the historical ambient temperature, historical ambient humidity, historical ambient wind speed, historical ambient rainfall and historical ambient sunshine of the simulated environment;

[0025] The real-time performance testing environment parameters include the real-time ambient temperature, real-time ambient humidity, real-time ambient wind speed, real-time ambient rainfall and real-time ambient sunshine of the simulated environment.

[0026] Furthermore, combining historical bridge damper parameters, historical performance test operating condition parameters, historical performance test environment parameters, and historical performance test simulation results of the same performance test process, a number of model training samples are obtained, and dimensionality reduction processing is performed on the number of model training samples to obtain a number of reduced-dimensionality model training samples, including the following steps:

[0027] Combining historical bridge damper parameters, historical performance test working condition parameters, and historical performance test environment parameters of the same performance test process to obtain a number of historical input data;

[0028] The historical performance test simulation results are used as the true labels of the corresponding historical input data to obtain several model training samples;

[0029] Use the RF algorithm to obtain the importance scores of several features of the historical input data in the model training sample for the true label, and use the features with the highest importance scores as key features;

[0030] According to several key features, the historical input data of each model training sample is subjected to dimensionality reduction processing to obtain several model training samples after dimensionality reduction.

[0031] Furthermore, the performance detection model of the bridge damper is constructed based on the LSTM-MLP algorithm, and the performance detection model of the bridge damper includes a data feature extraction module constructed based on the LSTM algorithm and a performance detection prediction module constructed based on the MLP algorithm, which are connected in sequence.

[0032] Further, the risk item performance detection verification guidance model is constructed based on the MPO-MOGRPO algorithm, and the verification strategy generation model comprises a meta-strategy optimization module constructed based on the MPO algorithm and a verification strategy generation module constructed based on the MOGRPO algorithm, and the verification strategy generation module comprises a target function set, a strategy network, an experience replay pool and an agent, the agent is connected with the target function set and the strategy network respectively, and the meta-strategy optimization module is connected with the strategy network.

[0033] Further, according to a plurality of dimension-reduced model training samples, a bridge damper performance detection model is constructed using a deep learning algorithm, a plurality of historical performance detection prediction results are obtained, and a risk item performance detection verification guidance model is constructed using a reinforcement learning algorithm according to the plurality of historical performance detection prediction results, comprising the following steps:

[0034] An initial bridge damper performance detection model is constructed using an LSTM-MLP algorithm, and an initial risk item performance detection verification guidance model is constructed using an MPO-MOGRPO algorithm.

[0035] The model parameters of the initial bridge damper performance detection model and the initial risk item performance detection verification guidance model are optimized using an ISGA algorithm to obtain an optimized bridge damper performance detection model and an optimized risk item performance detection verification guidance model.

[0036] A plurality of dimension-reduced model training samples are input to train the optimized bridge damper performance detection model to obtain a final bridge damper performance detection model and a plurality of historical performance detection prediction results.

[0037] According to the plurality of historical performance detection prediction results, the optimized risk item performance detection verification guidance model is trained to obtain a final risk item performance detection verification guidance model and a plurality of historical verification strategy generation experiences, and the plurality of historical verification strategy generation experiences are stored in an experience replay pool.

[0038] Further, according to real-time input data, a bridge damper performance detection prediction is performed using the bridge damper performance detection model to obtain a real-time performance detection prediction result, comprising the following steps:

[0039] The real-time input data is preprocessed to obtain preprocessed real-time input data, and the preprocessed real-time input data is input into the bridge damper performance detection model.

[0040] The real-time data features of the preprocessed real-time input data are extracted using a data feature extraction module of the bridge damper performance detection model.

[0041] The performance detection prediction module of the bridge damper performance detection model is used to perform bridge damper performance detection prediction according to real-time data characteristics to obtain real-time performance detection prediction results.

[0042] Furthermore, based on the real-time performance test prediction results, the risk item performance test verification guidance model is used to perform risk item performance test verification guidance, and a real-time risk item performance test verification strategy is obtained, which includes the following steps:

[0043] Based on the real-time performance test prediction results, the meta-strategy optimization module of the risk item performance test verification guidance model is used to update the policy network of the verification strategy generation module to obtain an updated policy network;

[0044] Randomly extract several historical verification strategy generation experiences from the experience replay pool, and update the intelligent agent of the verification strategy generation module based on the several historical verification strategy generation experiences and real-time performance test prediction results to obtain an updated intelligent agent;

[0045] A real-time objective function is selected from the objective function set, and based on the real-time objective function, a verification strategy generation module provided with an updated strategy network and an updated intelligent agent is used to generate a verification strategy to obtain a real-time risk item performance detection verification strategy.

[0046] A deep learning-based bridge damper performance detection system is used to implement a bridge damper performance detection method. The system includes a model construction unit, an input data setting unit, a performance detection prediction unit, a verification guidance unit, a verification result acquisition unit, and a final result generation unit.

[0047] The beneficial effects of the present invention are:

[0048] The present invention discloses a performance detection method and system for bridge dampers based on deep learning. By adopting deep learning and reinforcement learning algorithms, the actual performance of bridge dampers can be evaluated more comprehensively and accurately. It not only covers various types of dampers, but also can adapt to a variety of complex environments and working conditions, significantly improving the performance detection effect; it realizes automated and intelligent performance detection, greatly reduces manual intervention, improves detection efficiency, and reduces labor costs. In addition, after large-scale application, the growth of maintenance costs will be effectively controlled, achieving cost optimization; using advanced deep learning models, it can accurately identify subtle performance changes and hidden defects of dampers, including deep defects, thereby improving detection accuracy and reducing safety hazards; according to the real-time risk item performance detection and verification strategy, risk item performance detection and verification are carried out on bridge dampers, further improving the accuracy of performance detection.

[0049] Other beneficial effects of the present invention will be further described in the specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 This is a flowchart of the bridge damper performance detection method based on deep learning in the present invention.

[0051] Figure 2 This is a structural block diagram of the bridge damper performance detection system based on deep learning in the present invention. DETAILED DESCRIPTION

[0052] The present invention will be further explained below with reference to the accompanying drawings and specific embodiments.

[0053] Example 1:

[0054] like Figure 1 As shown, this embodiment provides a bridge damper performance detection method based on deep learning, comprising the following steps:

[0055] S1: Using deep learning algorithms and reinforcement learning algorithms, a bridge damper performance detection model and a risk item performance detection and verification guidance model are constructed, including the following steps:

[0056] S1-1: Use Computer-Aided Engineering (CAE) simulation software to build a performance test simulation environment. Based on this performance test simulation environment, set different damper property parameters for historical bridges, historical performance test operating condition parameters, and historical performance test environment parameters. Perform performance test simulations and obtain several historical performance test simulation results.

[0057] Perform performance testing simulations in a performance testing simulation environment using CAE simulation software to obtain a large amount of performance testing-related data, providing data support for subsequent model training;

[0058] The historical bridge damper property parameters include the historical elastic modulus, historical Poisson's ratio, historical density, historical damping coefficient, historical stiffness, historical maximum damping force and historical stroke of the bridge damper;

[0059] The historical performance test condition parameters include the historical static load, historical dynamic load, historical impact load, historical vehicle load and historical bridge deadweight of the bridge damper;

[0060] The historical performance test environment parameters include the historical ambient temperature, historical ambient humidity, historical ambient wind speed, historical ambient rainfall and historical ambient sunshine of the simulated environment;

[0061] S1-2: Combining historical bridge damper parameters, historical performance test operating condition parameters, historical performance test environment parameters, and historical performance test simulation results of the same performance test process, a number of model training samples are obtained, and dimensionality reduction processing is performed on the number of model training samples to obtain a number of model training samples after dimensionality reduction, including the following steps:

[0062] S1-2-1: Combine historical bridge damper parameters, historical performance test operating condition parameters, and historical performance test environment parameters of the same performance test process to obtain several historical input data;

[0063] S1-2-2: Use the historical performance test simulation results as the true labels of the corresponding historical input data to obtain several model training samples;

[0064] S1-2-3: Use the Random Forest (RF) algorithm to obtain the importance scores of several features of the historical input data in the model training sample for the true label, and use the features with the highest importance scores as key features;

[0065] S1-2-4: Based on several key features, the historical input data of each model training sample is subjected to dimensionality reduction processing to obtain several model training samples after dimensionality reduction;

[0066] S1-3: Based on several dimensionality-reduced model training samples, a deep learning algorithm is used to construct a bridge damper performance detection model, and several historical performance detection prediction results are obtained. Based on these historical performance detection prediction results, a reinforcement learning algorithm is used to construct a risk item performance detection verification guidance model, including the following steps:

[0067] S1-3-1: Use the Long Short-Term Memory (LSTM)-Multilayer Perceptron (MLP) algorithm to build an initial bridge damper performance detection model, and use the Meta-Policy Optimization (MPO)-Multi-Objective Group Relative Policy Optimization (MOGRPO) algorithm to build an initial risk item performance detection and verification guidance model;

[0068] The performance detection model of the bridge damper is constructed based on the LSTM-MLP algorithm, and the performance detection model of the bridge damper includes a data feature extraction module constructed based on the LSTM algorithm and a performance detection prediction module constructed based on the MLP algorithm, which are connected in sequence;

[0069] The LSTM can capture the complex relationship between the data and the performance detection result, improve the understanding ability of the model to the data, provide more accurate input features for the performance detection prediction module, and improve the accuracy of performance detection prediction; the performance detection prediction module MLP further processes and analyzes the extracted data features, learns the complex mapping relationship between the input features and the output labels, and improves the accuracy of performance detection prediction by using the powerful nonlinear modeling capability of the MLP;

[0070] The risk item performance detection verification guidance model is constructed based on the MPO-MOGRPO algorithm, and the verification strategy generation model includes a meta-strategy optimization module constructed based on the MPO algorithm and a verification strategy generation module constructed based on the MOGRPO algorithm. The verification strategy generation module includes a target function set, a strategy network, an experience replay pool and an agent. The agent is connected with the target function set and the strategy network respectively. The meta-strategy optimization module is connected with the strategy network.

[0071] The meta-strategy optimization module is used for verifying the network parameters of the strategy network in the verification strategy generation module, so that these parameters can quickly adapt to new, unseen performance detection prediction results, improve the generalization ability of the model, and even under unseen performance detection prediction results, the strategy network can be updated based on previous learning experience, improving the adaptability of the verification strategy generation model. The target function set of the verification strategy generation module can handle multiple conflicting goals, such as minimizing performance detection cost, maximizing performance detection accuracy, maximizing performance detection efficiency, etc., to generate verification strategies that balance these goals. The agent learns historical verification strategies through the experience replay pool, and continuously optimizes its strategy generation ability. The agent controls the strategy network according to the learned experience to generate more effective verification strategies. The design of the experience replay pool and the agent enables the model to continuously learn and optimize, improving the quality of strategy generation. The verification strategy generation module adopts a group exploration method to a certain extent, which can avoid falling into a local optimal solution. The strategy network outputs the distribution probability of actions under a given state. The verification strategy generation module directly updates the strategy network through gradient, which saves the Critic model in traditional reinforcement learning, making the algorithm structure more concise.

[0072] S1-3-2: Use the improved snow goose algorithm (Improved Snow Geese Algorithm, ISGA) algorithm to optimize the model parameters of the initial bridge damper performance detection model and the initial risk item performance detection verification guidance model, and obtain the optimized bridge damper performance detection model and the optimized risk item performance detection verification guidance model, including the following steps:

[0073] S1-3-2-1: Integrate the first initial model parameters of the initial bridge damper performance detection model and the second initial model parameters of the initial risk item performance detection and verification guidance model to obtain comprehensive initial model parameters;

[0074] S1-3-2-2: Encode the comprehensive initial model parameters into the individual vectors of the ISGA individuals in the ISGA algorithm, and set the ISGA population parameters and the maximum number of iterations;

[0075] S1-3-2-3: Take minimizing the error value as the optimization goal, and set the fitness function of the swarm intelligence optimization algorithm according to the optimization goal;

[0076] The formula is:

[0077] Fit(P)=minMSN(P)

[0078] Where Fit(P) is the fitness function; MSN(P) is the mean square error function; P is the ISGA individual;

[0079] S1-3-2-4: Based on the individual vectors of the ISGA individuals and the ISGA population parameters, the Circle chaotic mapping sequence is used to generate initial solutions and obtain several initial solutions; each initial solution corresponds to an initial ISGA individual, and several initial ISGA individuals constitute the initial ISGA individual population;

[0080] The formula is:

[0081]

[0082] Where, P i The initial ISGA individual generated by the Circle chaotic map sequence, that is, the initial solution; P o is a randomly generated ISGA individual; i is the ISGA individual indicator; mod(*) is the remainder function;

[0083] Compared with the randomly distributed population, the initial position distribution of the improved ISGA population generated by the Circle chaotic mapping sequence is more uniform, which expands the search range of the ISGA population in space and increases the diversity of group positions. To a certain extent, it improves the defect of the algorithm that it is easy to fall into local extreme values, thereby improving the optimization efficiency of the algorithm.

[0084] S1-3-2-5: Use the fitness function to obtain the initial fitness value of each initial ISGA individual in the initial ISGA population, and take the initial ISGA individual with the lowest fitness value as the leader goose;

[0085] S1-3-2-6: entering the exploration stage, introducing the leader switchover mechanism, the calling guiding mechanism and the dynamic reverse mechanism, iteratively updating the initial ISGA population to obtain an updated ISGA population, and retaining the optimal individual;

[0086] The leader switchover mechanism, in each iteration, competing according to the fitness values of the ISGA individuals to select a new leader, which can avoid the leader falling into local optimum too early and enhance the global search ability of the algorithm;

[0087] The formula is:

[0088]

[0089] In the formula, P i t+1 is the updated leader; is the initial ISGA individual with the third lowest fitness value in the initial ISGA population at the tth and (t+1)th iteration; is the initial ISGA individual with the fifth lowest fitness value in the initial ISGA population at the tth iteration; t is the current iteration number; is the optimal individual; a is the first weight factor; rand is a random number; P i t is the initial leader;

[0090] The calling guiding mechanism, according to the distance between the ISGA individual and the leader, using the sound wave propagation attenuation model to update the individual position, the ISGA individual with a closer distance is greatly affected by the leader and can quickly approach the optimal solution, and the ISGA individual with a farther distance is less affected by the leader and can maintain a certain exploration ability, which can avoid over-concentration or dispersion of the group and improve the local search precision of the algorithm;

[0091] The formula is:

[0092]

[0093] In the formula, is the updated ISGA individual; is the initial ISGA individual at the tth iteration; is the initial ISGA individual the received sound intensity; is the initial ISGA individual at the tth iteration the corresponding sound intensity parameter; L WA is the initial sound intensity; L low is the minimum receivable sound intensity; a" is the convergence factor; is the initial ISGA individual with the farthest distance; r' is a random parameter; B(d) is the Brownian motion function; d is the Brownian motion parameter; is the exclusive OR processing symbol;

[0094]

[0095] wherein a" is a convergence factor; tanh(.) is the hyperbolic tangent function; t is the current iteration number; t max is the maximum iteration number; a max , a min are the maximum and minimum values of the convergence factor respectively; λ is a decreasing rate parameter, k' is a decreasing period parameter, λ = -2π, k' = π;

[0096] a dynamic reverse mechanism is used to dynamically reverse the initial ISGA individual, to improve the diversity of the exploration direction and avoid falling into a local optimum;

[0097] the formula is:

[0098]

[0099] wherein, is the once-updated reverse ISGA individual; γ is a decreasing inertia coefficient; L max , L min are the maximum and minimum values of the vector space respectively;

[0100] the once-updated leader goose, the once-updated ISGA individuals and the once-updated reverse ISGA individuals are integrated to obtain the once-updated ISGA population, and the ISGA individual with the lowest fitness value is reserved as the optimal individual;

[0101] S1-3-2-7: entering the development stage, introducing an abnormal boundary strategy and a Gaussian variation mechanism, performing secondary updating on the once-updated ISGA population to obtain the secondary-updated ISGA population, and reserving the optimal individual;

[0102] the abnormal boundary strategy is used to calculate the difference between the fitness value of each once-updated ISGA individual and the average fitness value of the population, and the position updating mode of the ISGA individual with a fitness value much higher than the average value is adjusted, for example, using a Gaussian variation mechanism, a larger step or a smaller step, which can help the individual to avoid falling into a local optimum and improve the convergence speed and accuracy of the algorithm;

[0103] the formula is:

[0104]

[0105] wherein, is the secondary-updated ISGA individual; is an updated ISGA individual; Fit(*) is the fitness function; Fit avg is the average fitness value of the group; is the ISGA individual with the highest fitness value; a' and e are the second and third weight factors; G(1,1) is the Gaussian mutation mechanism parameter;

[0106] S1-3-2-8: If the number of iterations is greater than or equal to the iteration threshold or the fitness value of the optimal individual is less than the fitness threshold, the optimal individual is output as the optimal solution;

[0107] S1-3-2-9: Decode the individual vector of the optimal individual to obtain the optimal comprehensive initial model parameters, and optimize the model parameters of the initial bridge damper performance detection model and the initial risk item performance detection and verification guidance model based on the optimal comprehensive initial model parameters to obtain the optimized bridge damper performance detection model and the optimized risk item performance detection and verification guidance model;

[0108] S1-3-3: Input several model training samples after dimensionality reduction to train the optimized bridge damper performance detection model, and obtain the final bridge damper performance detection model and several historical performance detection prediction results;

[0109] S1-3-4: Based on several historical performance test prediction results, the optimized risk item performance test verification guidance model is trained to obtain the final risk item performance test verification guidance model and several historical verification strategy generation experiences. The several historical verification strategy generation experiences are stored in the experience replay pool, including the following steps:

[0110] S1-3-4-1: Use the verification strategy to generate the optimization target as the scenario for meta-strategy optimization, and train the optimized meta-strategy optimization module of the optimized risk item performance test verification guidance model based on several historical performance test prediction results to obtain the final meta-strategy optimization module;

[0111] S1-3-4-2: Use the final meta-strategy optimization module to initialize the policy network of the initial verification strategy generation module under different scenarios to obtain the initialized policy network;

[0112] S1-3-4-3: According to the different scenarios of the meta-strategy optimization module, that is, the optimization goal of the verification strategy, the optimized verification strategy generation module of the initialized policy network is set up for the optimized risk item performance detection and verification guidance model, the experience replay pool is set up, and the action space and state space are set up for the intelligent agent of the optimized verification strategy generation module;

[0113] S1-3-4-4: Use the verification strategy generation problem as a simulation environment, and obtain an updated verification strategy generation module based on the initialized strategy network and the intelligent agent with action space and state space;

[0114] S1-3-4-5: Traverse all objective functions in the objective function set, train the updated verification strategy generation module based on several historical performance test prediction results, obtain the final verification strategy generation module, and generate several historical verification strategy generation experiences;

[0115] S1-3-4-5: Integrate the final meta-strategy optimization module and the final verification strategy generation module to obtain the final verification strategy generation model, and store several historical verification strategy generation experiences in the experience replay pool;

[0116] S2: according to the performance test requirements, corresponding real-time performance test condition parameters and real-time performance test environment parameters are set for the real-time bridge damper attribute parameters of the bridge damper to obtain real-time input data;

[0117] The real-time bridge damper property parameters include the real-time elastic modulus, real-time Poisson's ratio, real-time density, real-time damping coefficient, real-time stiffness, real-time maximum damping force and real-time stroke of the bridge damper;

[0118] Real-time performance testing parameters include real-time static load, real-time dynamic load, real-time impact load, real-time vehicle load and real-time bridge deadweight of bridge dampers;

[0119] Real-time performance testing environmental parameters include the real-time ambient temperature, real-time ambient humidity, real-time ambient wind speed, real-time ambient rainfall and real-time ambient sunshine of the simulated environment;

[0120] S3: Based on the real-time input data, a bridge damper performance detection model is used to perform a bridge damper performance detection prediction to obtain a real-time performance detection prediction result, including the following steps:

[0121] S3-1: preprocessing the real-time input data to obtain preprocessed real-time input data, and inputting the preprocessed real-time input data into the bridge damper performance detection model;

[0122] S3-2: Use the data feature extraction module of the bridge damper performance detection model to extract real-time data features of the pre-processed real-time input data;

[0123] S3-3: Using the performance detection prediction module of the bridge damper performance detection model, the bridge damper performance detection prediction is performed according to the real-time data characteristics to obtain the real-time performance detection prediction result;

[0124] Real-time performance test prediction results include real-time pressure resistance performance prediction results, real-time slow and speed-related performance prediction results, real-time frequency-related performance prediction results, real-time temperature-related performance prediction results, real-time seismic performance prediction results, real-time wind vibration load performance prediction results, and real-time fatigue and wear resistance performance prediction results;

[0125] S4: Based on the real-time performance test prediction results, the risk item performance test verification guidance model is used to provide risk item performance test verification guidance, and a real-time risk item performance test verification strategy is obtained, including the following steps:

[0126] S4-1: Based on the real-time performance test prediction results, the meta-strategy optimization module of the risk item performance test verification guidance model is used to update the policy network of the verification strategy generation module to obtain an updated policy network;

[0127] S4-2: Randomly extract several historical verification strategy generation experiences from the experience replay pool, and update the intelligent agent of the verification strategy generation module based on the several historical verification strategy generation experiences and the real-time performance test prediction results to obtain an updated intelligent agent, including the following steps:

[0128] S4-2-1: Randomly extract several historical verification strategy generation experiences from the experience replay pool, generate several possible verification actions based on these historical verification strategy generation experiences, and update the action space of the intelligent agent of the verification strategy generation module based on these possible verification actions to obtain an updated action space;

[0129] S4-2-2: Analyze the real-time performance test prediction results to obtain several real-time performance test prediction states, and update the state space of the intelligent agent of the verification strategy generation module according to the real-time performance test prediction states to obtain an updated state space;

[0130] S4-2-3: Integrate the updated state space and updated action space of the agent to obtain an updated agent;

[0131] S4-3: Selecting a real-time objective function from the objective function set, and generating a verification strategy based on the real-time objective function using a verification strategy generation module provided with an updated strategy network and an updated agent, to obtain a real-time risk item performance detection verification strategy, including the following steps:

[0132] S4-3-1: Select a real-time objective function from the objective function set of the verification strategy generation module, and based on the real-time objective function, use the updated agent of the verification strategy generation module to control the updated policy network to generate a probability distribution of all possible verification actions in the updated action space corresponding to each real-time performance test prediction state in the updated state space;

[0133] S4-3-2: The possible verification action with the highest probability distribution in the updated action space is used as the execution verification action of the corresponding real-time performance detection prediction state;

[0134] S4-3-3: Integrate the execution verification actions of all real-time performance detection prediction states in the updated state space to obtain the real-time risk item performance detection verification strategy;

[0135] The real-time risk item performance test and verification strategy includes real-time risk item selection decision, real-time risk item performance test and verification decision, real-time performance test working condition parameter adjustment decision, and real-time performance test environment parameter adjustment decision;

[0136] S5: According to the real-time risk item performance detection and verification strategy, the risk item performance detection and verification of the bridge damper is performed to obtain the real-time risk item performance detection and verification result;

[0137] S6: The final results of the real-time performance test of the bridge damper are obtained by combining the real-time performance test prediction results and the real-time risk item performance test verification results.

[0138] Example 2:

[0139] like Figure 2 As shown, this embodiment provides a bridge damper performance detection system based on deep learning, which is used to implement a bridge damper performance detection method. The system includes a model building unit, an input data setting unit, a performance detection prediction unit, a verification guidance unit, a verification result collection unit, and a final result generation unit;

[0140] A model building unit, used to build a bridge damper performance detection model and a risk item performance detection and verification guidance model using deep learning algorithms and reinforcement learning algorithms;

[0141] An input data setting unit is used to set corresponding real-time performance detection working condition parameters and real-time performance detection environment parameters for the real-time bridge damper attribute parameters of the bridge damper according to performance detection requirements, so as to obtain real-time input data;

[0142] A performance detection and prediction unit is used to perform a performance detection and prediction of the bridge damper based on real-time input data using a bridge damper performance detection model to obtain a real-time performance detection prediction result;

[0143] A verification guidance unit is used to provide risk item performance test verification guidance based on the real-time performance test prediction result and use the risk item performance test verification guidance model to obtain a real-time risk item performance test verification strategy;

[0144] A verification result acquisition unit is used to acquire real-time risk item performance detection and verification results obtained by performing risk item performance detection and verification on the bridge damper according to the real-time risk item performance detection and verification strategy;

[0145] The final result generating unit is used to combine the real-time performance test prediction result and the real-time risk item performance test verification result to obtain the final result of the real-time performance test of the bridge damper.

[0146] The present invention discloses a performance detection method and system for bridge dampers based on deep learning. By adopting deep learning and reinforcement learning algorithms, the actual performance of bridge dampers can be evaluated more comprehensively and accurately. It not only covers various types of dampers, but also can adapt to a variety of complex environments and working conditions, significantly improving the performance detection effect; it realizes automated and intelligent performance detection, greatly reduces manual intervention, improves detection efficiency, and reduces labor costs. In addition, after large-scale application, the growth of maintenance costs will be effectively controlled, achieving cost optimization; using advanced deep learning models, it can accurately identify subtle performance changes and hidden defects of dampers, including deep defects, thereby improving detection accuracy and reducing safety hazards; according to the real-time risk item performance detection and verification strategy, risk item performance detection and verification are carried out on bridge dampers, further improving the accuracy of performance detection.

[0147] The present invention is not limited to the above optional embodiments. Anyone can derive various other forms of products based on the teachings of the present invention. The above specific embodiments should not be construed as limiting the scope of protection of the present invention. The scope of protection of the present invention shall be based on the scope defined in the claims, and the description can be used to interpret the claims.

Claims

1. A bridge damper performance detection method based on deep learning, characterized by: The steps include: Using deep learning algorithms and reinforcement learning algorithms, a performance detection model for bridge dampers and a risk item performance detection and verification guidance model were constructed; The risk item performance detection and verification guidance model is constructed based on the MPO-MOGRPO algorithm, and the verification strategy generation model includes a meta-strategy optimization module constructed based on the MPO algorithm and a verification strategy generation module constructed based on the MOGRPO algorithm. The verification strategy generation module includes an objective function set, a strategy network, an experience replay pool, and an intelligent agent. The intelligent agent is connected to the objective function set and the strategy network respectively, and the meta-strategy optimization module is connected to the strategy network. The steps include: Use the LSTM-MLP algorithm to build an initial bridge damper performance detection model, and use the MPO-MOGRPO algorithm to build an initial risk item performance detection and verification guidance model; The ISGA algorithm is used to optimize the model parameters of the initial bridge damper performance detection model and the initial risk item performance detection verification guidance model, thereby obtaining the optimized bridge damper performance detection model and the optimized risk item performance detection verification guidance model. Input a number of model training samples after dimensionality reduction to train the optimized bridge damper performance detection model, and obtain the final bridge damper performance detection model and a number of historical performance detection prediction results; Based on several historical performance test prediction results, the optimized risk item performance test verification guidance model is trained to obtain the final risk item performance test verification guidance model and several historical verification strategy generation experiences, and the several historical verification strategy generation experiences are stored in the experience replay pool; According to the performance detection requirements, corresponding real-time performance detection working condition parameters and real-time performance detection environment parameters are set for the real-time bridge damper attribute parameters of the bridge damper to obtain real-time input data; According to the real-time input data, the bridge damper performance detection model is used to perform bridge damper performance detection prediction and obtain real-time performance detection prediction results; Based on the real-time performance test prediction results, the risk item performance test verification guidance model is used to provide risk item performance test verification guidance and obtain the real-time risk item performance test verification strategy; According to the real-time risk item performance test and verification strategy, the risk item performance test and verification of the bridge damper is carried out to obtain the real-time risk item performance test and verification results; The final results of the real-time performance test of the bridge damper are obtained by combining the real-time performance test prediction results and the real-time risk item performance test verification results.

2. The method for detecting performance of a bridge damper based on deep learning according to claim 1, characterized in that: Using deep learning algorithms and reinforcement learning algorithms, a bridge damper performance detection model and a risk item performance detection and verification guidance model are constructed, including the following steps: Using CAE simulation software, a performance test simulation environment was built. Based on the performance test simulation environment, different damper attribute parameters for historical bridges, historical performance test operating condition parameters, and historical performance test environment parameters were set to perform performance test simulations and obtain several historical performance test simulation results. Combining historical bridge damper parameters, historical performance test operating condition parameters, historical performance test environment parameters, and historical performance test simulation results of the same performance test process, a number of model training samples are obtained, and dimensionality reduction processing is performed on the number of model training samples to obtain a number of reduced-dimensionality model training samples; Based on several dimensionality reduction model training samples, a deep learning algorithm is used to construct a performance detection model for bridge dampers, and several historical performance detection prediction results are obtained. Based on several historical performance detection prediction results, a reinforcement learning algorithm is used to construct a risk item performance detection verification guidance model.

3. The method for detecting performance of a bridge damper based on deep learning according to claim 2, characterized in that: The historical bridge damper attribute parameters include the historical elastic modulus, historical Poisson's ratio, historical density, historical damping coefficient, historical stiffness, historical maximum damping force and historical travel of the bridge damper; The real-time bridge damper attribute parameters include the real-time elastic modulus, real-time Poisson's ratio, real-time density, real-time damping coefficient, real-time stiffness, real-time maximum damping force and real-time stroke of the bridge damper; The historical performance test condition parameters include historical static loads, historical dynamic loads, historical impact loads, historical vehicle loads and historical bridge deadweights of bridge dampers; The real-time performance detection working condition parameters include the real-time static load, real-time dynamic load, real-time impact load, real-time vehicle load and real-time bridge deadweight of the bridge damper; The historical performance test environment parameters include the historical environment temperature, historical environment humidity, historical environment wind speed, historical environment rainfall and historical environment sunshine of the simulated environment; The real-time performance detection environmental parameters include the real-time ambient temperature, real-time ambient humidity, real-time ambient wind speed, real-time ambient rainfall and real-time ambient sunshine of the simulated environment.

4. The method for detecting performance of a bridge damper based on deep learning according to claim 3, characterized in that: Combining historical bridge damper parameters, historical performance test operating condition parameters, historical performance test environment parameters, and historical performance test simulation results of the same performance test process, a number of model training samples are obtained, and dimensionality reduction processing is performed on the number of model training samples to obtain a number of reduced-dimensionality model training samples, including the following steps: Combining historical bridge damper parameters, historical performance test working condition parameters, and historical performance test environment parameters of the same performance test process to obtain a number of historical input data; The historical performance test simulation results are used as the true labels of the corresponding historical input data to obtain several model training samples; Use the RF algorithm to obtain the importance scores of several features of the historical input data in the model training sample for the true label, and use the features with the highest importance scores as key features; According to several key features, the historical input data of each model training sample is subjected to dimensionality reduction processing to obtain several model training samples after dimensionality reduction.

5. The method for detecting performance of a bridge damper based on deep learning according to claim 4, characterized in that: The bridge damper performance detection model is constructed based on the LSTM-MLP algorithm, and the bridge damper performance detection model includes a data feature extraction module constructed based on the LSTM algorithm and a performance detection prediction module constructed based on the MLP algorithm, which are connected in sequence.

6. The method for detecting performance of a bridge damper based on deep learning according to claim 5, characterized in that: Based on real-time input data, a bridge damper performance detection model is used to perform bridge damper performance detection prediction to obtain real-time performance detection prediction results, including the following steps: Preprocessing the real-time input data to obtain preprocessed real-time input data, and inputting the preprocessed real-time input data into a performance detection model for a bridge damper; Use the data feature extraction module of the bridge damper performance detection model to extract real-time data features of the pre-processed real-time input data; The performance detection prediction module of the bridge damper performance detection model is used to perform bridge damper performance detection prediction according to real-time data characteristics to obtain real-time performance detection prediction results.

7. The method for detecting performance of a bridge damper based on deep learning according to claim 6, characterized in that: Based on the real-time performance test prediction results, the risk item performance test verification guidance model is used to provide risk item performance test verification guidance, and a real-time risk item performance test verification strategy is obtained, which includes the following steps: Based on the real-time performance test prediction results, the meta-strategy optimization module of the risk item performance test verification guidance model is used to update the policy network of the verification strategy generation module to obtain an updated policy network; Randomly extract several historical verification strategy generation experiences from the experience replay pool, and update the intelligent agent of the verification strategy generation module based on the several historical verification strategy generation experiences and real-time performance test prediction results to obtain an updated intelligent agent; A real-time objective function is selected from the objective function set, and based on the real-time objective function, a verification strategy generation module provided with an updated strategy network and an updated intelligent agent is used to generate a verification strategy to obtain a real-time risk item performance detection verification strategy.

8. A deep learning-based bridge damper performance detection system, used to implement the bridge damper performance detection method according to any one of claims 1 to 7, characterized in that: The system includes a model building unit, an input data setting unit, a performance detection prediction unit, a verification guidance unit, a verification result collection unit and a final result generation unit.

Images