Anti-seismic device and anti-seismic method for hollow thin-wall high-pier bridge

Abstract

The invention provides an anti-seismic device and an anti-seismic method for a hollow thin-wall high-pier bridge, and can be applied to the technical field of bridge seismic resistance. The device comprises a diaphragm plate mounted in the hollow thin-wall high-pier bridge; the damper is mounted on the transverse partition plate; the vibration detection sensor comprises a stress sensor, a displacement sensor and / or an acceleration sensor and is mounted on the hollow thin-wall high-pier bridge; the data processor is electrically connected with the vibration detection sensor; and the damping parameter controller is electrically connected with the data processor and the damper respectively. According to the anti-seismic device for the hollow thin-wall high-pier bridge and the anti-seismic method, real-time monitoring and numerical calculation can be conducted on deformation, displacement and / or vibration of the bridge, parameters of the damper are adjusted, intelligent active control over the damper is achieved, the anti-seismic capacity of the bridge is improved, and the service life of the bridge is prolonged. And the device has the advantages of simple structure, easiness in installation and maintenance and the like, and has good economic and social benefits.

Description

Technical Field

[0001] This application relates to the field of bridge seismic resistance technology, specifically to seismic resistance devices and methods for hollow thin-walled high-pier bridges. Background Technology

[0002] Hollow thin-walled high-pier bridges are widely used in engineering because thin-walled high piers can better adapt to the requirements of terrain, geology, navigation and route indicators, and the stress is reasonable.

[0003] To ensure the stability of hollow thin-walled high piers, bridge engineers typically employ the method of installing diaphragms inside the piers. Current standards designate diaphragms as structural components, which generally guarantee structural safety when not subjected to seismic or wind vibration loads. However, their safety under seismic loads is questionable, and the seismic resistance of thin-walled high-pier bridges is particularly important in earthquake-prone areas.

[0004] Currently, the technology for adding dampers to thin-walled high pier diaphragms as intelligent seismic resisting devices has the following shortcomings:

[0005] Existing technologies rely heavily on passive control methods, which cannot achieve intelligent and systematic control and are not suitable for the needs of bridge vibration control in the new environment.

[0006] The manufacturing, replacement and maintenance costs of seismic isolation bearings are relatively high, and it is difficult to quantify and adjust the seismic isolation effect under different seismic loads. Summary of the Invention

[0007] In view of the problems in the prior art, the embodiments of this application provide a seismic resisting device and method for hollow thin-walled high-pier bridges, which can at least partially solve the problems existing in the prior art.

[0008] On the one hand, this application proposes a seismic resisting device for hollow thin-walled high-pier bridges, comprising:

[0009] Transverse diaphragms are installed inside hollow thin-walled high-pier bridges;

[0010] A damper is mounted on the diaphragm;

[0011] Vibration detection sensors, including stress sensors, displacement sensors and / or acceleration sensors, are installed on the hollow thin-walled high pier bridge;

[0012] The data processor is electrically connected to the vibration detection sensor;

[0013] The damping parameter controller is electrically connected to both the data processor and the damper.

[0014] In some embodiments, the vibration detection sensor is installed in the middle of the bridge span, on the upper part of the pier, in the middle of the pier, at the lower part of the pier, on both sides of the expansion joint and / or on the diaphragm.

[0015] In some embodiments, the stress sensor collects stress data of the bridge piers, diaphragms, and main beams under seismic loads and sends the stress data to the data processor; and / or

[0016] The displacement sensor collects displacement data of the bridge piers, diaphragms, and main beams in the horizontal and vertical directions under seismic loads, and sends the displacement data to the data processor; and / or

[0017] The accelerometer collects vibration data of the bridge piers, diaphragms, and main beams under seismic loads and sends the vibration data to the data processor.

[0018] In some embodiments, the damper is a magnetically controlled liquid damper.

[0019] On the other hand, this application provides a seismic resistance method for hollow thin-walled high-pier bridges, based on the seismic resistance device for hollow thin-walled high-pier bridges described in any of the above embodiments, the method comprising:

[0020] The actual response data of the bridge structure under seismic loading is acquired by the vibration detection sensor at the first moment. The actual response data includes stress data, displacement data and / or acceleration data.

[0021] The actual response data is input into the pre-trained prediction model to obtain the response data of the bridge structure under seismic action at the second time, output by the prediction model.

[0022] If at least one of the response data at the second time exceeds the safety threshold of that response data, then the response data at the second time is input into a pre-trained neural network model to obtain the damping parameter control data output by the neural network model.

[0023] The damping parameter control data is sent to the damping parameter controller, which is used to adjust the damping force of the damper according to the damping parameter control data.

[0024] In some embodiments, the training process of the prediction model is as follows:

[0025] The actual response data of the bridge structure under earthquake action collected by the vibration detection sensor at the first historical time is input into the preset prediction model to obtain the response data of the bridge structure under earthquake action at the second historical time predicted by the prediction model.

[0026] Based on the response data predicted by the prediction model at the second historical time and the actual response data collected by the vibration detection sensor at the second historical time, the first prediction loss of the prediction model is calculated.

[0027] The model parameters of the prediction model are adjusted based on the first prediction loss to obtain the prediction model after the first parameter adjustment.

[0028] The actual response data of the bridge structure under earthquake action collected by the vibration detection sensor at the second historical time is input into the prediction model after the first parameter adjustment to obtain the response data of the bridge structure under earthquake action predicted by the prediction model at the third historical time.

[0029] Based on the response data predicted by the prediction model at the third historical time and the actual response data collected by the vibration detection sensor at the third historical time, the second prediction loss of the prediction model after the first parameter adjustment is calculated, wherein the time difference between the third historical time and the second historical time, the time difference between the second historical time and the first historical time, and the time difference between the second time and the first time are all equal.

[0030] The model parameters of the prediction model after the first parameter adjustment are adjusted according to the second prediction loss to obtain the prediction model after the second parameter adjustment.

[0031] This process continues until the prediction loss of the prediction model is less than the first target value or the number of model parameter adjustments reaches the second target value, thus obtaining a trained prediction model.

[0032] In some embodiments, the training process of the neural network model is as follows:

[0033] The actual response data of the bridge structure under seismic loading collected by the vibration detection sensor at the fourth historical time is processed by time delay to obtain the first time delay data;

[0034] The first time delay data is input into a preset neural network model to obtain the damping parameter control data of the fourth historical time output by the neural network model.

[0035] The first prediction loss of the neural network model is calculated based on the damping parameter control data of the fourth historical time and the ideal damping parameter control data of the fourth historical time.

[0036] The model parameters of the neural network model are adjusted based on the first prediction loss to obtain the neural network model after the first parameter adjustment.

[0037] The actual response data of the bridge structure under seismic loading collected by the vibration detection sensor at the fifth historical time is processed by time delay to obtain the second time delay data;

[0038] The second time delay data is input into the neural network model after the first parameter adjustment to obtain the damping parameter control data of the fifth historical time output by the neural network model;

[0039] Based on the damping parameter control data of the fifth historical time and the ideal damping parameter control data of the fifth historical time, calculate the second prediction loss of the neural network model after the first parameter adjustment;

[0040] The model parameters of the neural network model after the first parameter adjustment are adjusted according to the second prediction loss to obtain the neural network model after the second parameter adjustment.

[0041] This process continues until the prediction loss of the neural network model is less than the first target value or the number of model parameter adjustments reaches the second target value, thus obtaining a trained neural network model.

[0042] In some embodiments, when performing time-delay processing on the actual response data, the time delay length is 1 / 4 to 1 / 5 of the vibration period of the first-order frequency mode of the bridge.

[0043] In some embodiments, the method further comprises:

[0044] If none of the response data in the second time interval exceeds the safety threshold of that response data, then the actual response data of the bridge structure under seismic action collected by the vibration detection sensor in the second time interval will continue to be acquired.

[0045] Furthermore, this application provides a data processor, comprising:

[0046] The acquisition unit is used to acquire the actual response data of the bridge structure under seismic loading collected by the vibration detection sensor at the first moment. The actual response data includes stress data, displacement data and / or acceleration data.

[0047] The first prediction unit is used to input the actual response data into a pre-trained prediction model to obtain the response data of the bridge structure under seismic action at the second time, output by the prediction model.

[0048] The second prediction unit is used to input the response data of the second time into a pre-trained neural network model if at least one of the response data in the second time exceeds the safety threshold of the response data, so as to obtain the damping parameter control data output by the neural network model.

[0049] A transmitting unit is used to transmit the damping parameter control data to a damping parameter controller, which is used to adjust the damping force of the damper according to the damping parameter control data.

[0050] This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps of the method described in any of the above embodiments.

[0051] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the method described in any of the above embodiments.

[0052] The seismic resisting device for hollow thin-walled high-pier bridges provided in this application adds dampers to the thin-walled diaphragms, which has little impact on the structural stiffness of the hollow thin-walled high-pier bridge. In addition, it can monitor and numerically calculate the bridge deformation, displacement and / or vibration in real time, and adjust the damper parameters, realizing intelligent active control of the damper, improving the seismic resistance of the bridge, and has the advantages of simple structure, easy installation and maintenance, and good economic and social benefits. Attached Figure Description

[0053] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings:

[0054] Figure 1 This is a partial structural schematic diagram of an anti-seismic device for hollow thin-walled high-pier bridges provided in an embodiment of this application.

[0055] Figure 2 This is a schematic diagram showing the arrangement of a portion of the structure of the anti-seismic device for hollow thin-walled high-pier bridges provided in an embodiment of this application on a bridge.

[0056] Figure 3 This is a schematic diagram showing the arrangement of a portion of the structure of the anti-seismic device for hollow thin-walled high-pier bridges provided in an embodiment of this application on a bridge.

[0057] Figure 4 This is a schematic diagram showing the arrangement of the damper and damping parameter controller on the diaphragm according to an embodiment of this application.

[0058] Figure 5 This is a schematic flowchart of a seismic resistance method for hollow thin-walled high-pier bridges provided in an embodiment of this application.

[0059] Figure 6 This is a schematic diagram of the training process of a prediction model provided in an embodiment of this application.

[0060] Figure 7 This is a schematic diagram of the training process of a neural network model provided in an embodiment of this application.

[0061] Figure 8 This is a schematic diagram of the overall process of a seismic resistance method for hollow thin-walled high-pier bridges provided in an embodiment of this application.

[0062] Figure 9 This is a schematic diagram of the structure of a processor provided in one embodiment of this application. Detailed Implementation

[0063] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the embodiments of this application will be further described in detail below with reference to the accompanying drawings. Here, the illustrative embodiments and their descriptions are used to explain this application, but are not intended to limit this application. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of this application can be arbitrarily arranged.

[0064] The terms “first,” “second,” etc., used in this document are not intended to specifically refer to order or sequence, nor are they used to limit this application; they are merely used to distinguish elements or operations described using the same technical terms.

[0065] The words “include,” “including,” “have,” “contain,” etc. used in this document are open-ended terms, meaning including but not limited to.

[0066] The term "and / or" as used in this document includes any or all of the items mentioned.

[0067] The seismic resistance method for hollow thin-walled high-pier bridges provided in this application includes, but is not limited to, computers.

[0068] like Figure 1 , Figure 2 , Figure 3 as well as Figure 4 As shown in the embodiment of this application, the seismic resisting device for hollow thin-walled high-pier bridges includes:

[0069] 8 diaphragms are installed inside hollow thin-walled high-pier bridges;

[0070] Damper 6 is installed on the diaphragm 8;

[0071] Vibration detection sensors, including stress sensor 1, displacement sensor 2 and / or acceleration sensor 3, are installed on the hollow thin-walled high pier bridge;

[0072] Data processor 4 is electrically connected to the vibration detection sensor;

[0073] The damping parameter controller 5 is electrically connected to the data processor 4 and the damper 6, respectively.

[0074] Specifically, the thin-walled high pier is a thin-walled structure. Adding dampers to the thin-walled diaphragms has little impact on the structural stiffness and enables it to have sufficient stiffness in both the lateral and longitudinal directions.

[0075] The data processor collects sensor data, analyzes and processes the data uploaded by the sensors, and issues adjustment commands to the damping parameter controller based on the analysis results, further performing real-time monitoring and numerical calculation of bridge deformation, displacement and / or vibration.

[0076] The damping parameter controller adjusts the resistance value of the external resistor of the magnetically controlled fluid damper according to the adjustment instructions of the data processor, thereby actively adjusting the damper parameters.

[0077] The damper can be a magnetically controlled liquid damper, which, after parameter adjustment, can better dissipate the energy generated by the earthquake and maximize the control of the vibration and displacement of thin-walled high piers, diaphragms and / or main beams under seismic action.

[0078] The seismic resisting device for hollow thin-walled high-pier bridges provided in this application adds dampers to the thin-walled diaphragms, which has little impact on the structural stiffness of the hollow thin-walled high-pier bridge. In addition, it can monitor and numerically calculate the bridge deformation, displacement and / or vibration in real time, and adjust the damper parameters, realizing intelligent active control of the damper, improving the seismic resistance of the bridge, and has the advantages of simple structure, easy installation and maintenance, and good economic and social benefits.

[0079] like Figure 2 As shown, in some embodiments, the vibration detection sensor is installed in the middle of the bridge span 10, on the upper part of the bridge pier 7, in the middle of the bridge pier 7, at the lower part of the bridge pier 7, on both sides of the expansion joint 9 and / or on the diaphragm 8.

[0080] In some embodiments, the stress sensor collects stress data of the bridge piers, diaphragms, and main beams under seismic loads and sends the stress data to the data processor; and / or

[0081] The displacement sensor collects displacement data of the bridge piers, diaphragms, and main beams in the horizontal and vertical directions under seismic loads, and sends the displacement data to the data processor; and / or

[0082] The accelerometer collects vibration data of the bridge piers, diaphragms, and main beams under seismic loads and sends the vibration data to the data processor.

[0083] Figure 5 This is a schematic flowchart illustrating a seismic resistance method for hollow thin-walled high-pier bridges provided in an embodiment of this application. Figure 5 As shown in the embodiments of this application, the seismic resistance method for hollow thin-walled high-pier bridges includes:

[0084] S101. Obtain the actual response data of the bridge structure under seismic action collected by the vibration detection sensor at the first moment. The actual response data includes stress data, displacement data and / or acceleration data.

[0085] In step S101, the first time can be a point in time or a period of time. When the first time is a period of time, the actual response data collected by the vibration detection sensor at the first time is a set of time-series data.

[0086] S102. Input the actual response data into the pre-trained prediction model to obtain the response data of the bridge structure under seismic action at the second time, output by the prediction model.

[0087] In step S102, when the first time is a point in time, the second time is also a point in time; when the first time is a time period, the second time is also a time period, and the duration of the second time is the same as the duration of the first time. The prediction model can be trained based on actual response data from historical periods.

[0088] S103. If at least one of the response data in the second time exceeds the safety threshold of the response data, the response data in the second time is input into the pre-trained neural network model to obtain the damping parameter control data output by the neural network model.

[0089] In step S103, a safety threshold is set in advance for each type of response data. If one or more physical quantities in each response data exceed the safety threshold, the active control process is initiated, which means that the trained neural network model is used to calculate the precise control data.

[0090] S104. The damping parameter control data is sent to the damping parameter controller, which is used to adjust the damping force of the damper according to the damping parameter control data.

[0091] In step S104, the damping parameter control data is transmitted to the damping parameter controller to achieve active vibration control of the bridge. By continuously iterating and adjusting the control commands, and combining the vibration response data obtained from real-time monitoring, the control effect is further optimized, and the accuracy and stability of vibration control are improved.

[0092] The seismic resistance method for hollow thin-walled high-pier bridges provided in this application acquires actual response data of the bridge structure under seismic loading, collected by vibration detection sensors at a first-time moment. This actual response data includes stress data, displacement data, and / or acceleration data. The actual response data is input into a pre-trained prediction model to obtain the bridge structure's response data under seismic loading at a second-time moment, output by the prediction model. If at least one response data point in the second-time moment exceeds a safety threshold, the second-time moment response data is input into a pre-trained neural network model to obtain damping parameter control data output by the neural network model. This damping parameter control data is then sent to a damping parameter controller, which adjusts the damping force of the damper based on the control data. This method enables real-time monitoring and numerical calculation of bridge deformation, displacement, and / or vibration, and allows for adjustment of damper parameters, achieving intelligent and active control of the damper and improving the bridge's seismic resistance.

[0093] like Figure 6 As shown, in some embodiments, the training process of the prediction model is as follows:

[0094] S201. Input the actual response data of the bridge structure under earthquake action collected by the vibration detection sensor at the first historical time into the preset prediction model to obtain the response data of the bridge structure under earthquake action predicted by the prediction model at the second historical time.

[0095] S202. Calculate the first prediction loss of the prediction model based on the response data predicted by the prediction model at the second historical time and the actual response data collected by the vibration detection sensor at the second historical time.

[0096] S203. Adjust the model parameters of the prediction model according to the first prediction loss to obtain the prediction model after the first parameter adjustment;

[0097] S204. Input the actual response data of the bridge structure under earthquake action collected by the vibration detection sensor at the second historical time into the prediction model after the first parameter adjustment to obtain the response data of the bridge structure under earthquake action predicted by the prediction model at the third historical time.

[0098] S205. Based on the response data predicted by the prediction model at the third historical time and the actual response data collected by the vibration detection sensor at the third historical time, calculate the second prediction loss of the prediction model after the first parameter adjustment, wherein the time difference between the third historical time and the second historical time, the time difference between the second historical time and the first historical time, and the time difference between the second time and the first time are all equal.

[0099] S206. Adjust the model parameters of the prediction model after the first parameter adjustment according to the second prediction loss to obtain the prediction model after the second parameter adjustment.

[0100] S207. This process continues until the prediction loss of the prediction model is less than the first target value or the number of model parameter adjustments of the prediction model reaches the second target value, thus obtaining a trained prediction model.

[0101] Specifically, during training, training data samples are input into the model to obtain prediction results, and the loss function is calculated. Then, based on the value of the loss function and the parameter update rules, the model parameters are updated. This process is repeated multiple times until a stopping condition is met. By iteratively optimizing the model parameters, they can better fit the seismic response data of the bridge structure. After training, a prediction model is obtained, which can predict the values ​​of physical quantities such as stress, displacement, and acceleration based on the input seismic response data of the bridge structure.

[0102] like Figure 7 As shown, in some embodiments, the training process of the neural network model is as follows:

[0103] S301. Perform time-delay processing on the actual response data of the bridge structure under earthquake action collected by the vibration detection sensor at the fourth historical time to obtain the first time-delay data;

[0104] S302. Input the first time delay data into a preset neural network model to obtain the damping parameter control data of the fourth historical time output by the neural network model.

[0105] S303. Calculate the first prediction loss of the neural network model based on the damping parameter control data of the fourth historical time and the ideal damping parameter control data of the fourth historical time.

[0106] S304. Adjust the model parameters of the neural network model according to the first prediction loss to obtain the neural network model after the first parameter adjustment;

[0107] S305. Perform time-delay processing on the actual response data of the bridge structure under earthquake action collected by the vibration detection sensor at the fifth historical time to obtain the second time-delay data;

[0108] S306. Input the second time delay data into the neural network model after the first parameter adjustment to obtain the damping parameter control data of the fifth historical time output by the neural network model;

[0109] S307. Based on the damping parameter control data of the fifth historical time and the ideal damping parameter control data of the fifth historical time, calculate the second prediction loss of the neural network model after the first parameter adjustment.

[0110] S308. Adjust the model parameters of the neural network model after the first parameter adjustment according to the second prediction loss to obtain the neural network model after the second parameter adjustment.

[0111] S309. This process continues until the prediction loss of the neural network model is less than the first target value or the number of model parameter adjustments of the neural network model reaches the second target value, thus obtaining a trained neural network model.

[0112] Specifically, to control the damper and thus control bridge vibration, time delay effects need to be considered, requiring the training of a neural network model that takes time delay into account. Assuming the vibration period of the bridge's first-order frequency mode is T, the time delay length τ can be set to 1 / 4 to 1 / 5 of T. The response data is then processed to account for the time delay, constructing time-delayed data. The displacement X after time delay processing is then... i (t-τ), velocity and / or acceleration The vector is used as input, and the control command is calculated through a neural network model.

[0113] In some embodiments, the method further includes: if each type of response data in the response data at the second time does not exceed the safety threshold of the response data, then continue to acquire the actual response data of the bridge structure under seismic action collected by the vibration detection sensor at the second time.

[0114] To better understand the seismic resistance method for hollow thin-walled high-pier bridges provided in this application, the following detailed description of the seismic resistance method for hollow thin-walled high-pier bridges provided in this application is provided through a specific embodiment.

[0115] This embodiment aims to achieve precise control of bridge seismic vibration by comprehensively utilizing data acquisition, machine learning, and neural network control technologies. First, seismic response data such as stress, displacement, and acceleration collected from the bridge structure are calibrated and converted into actual physical quantities. A preprocessing procedure is then used to extract and represent features from the data to adapt it to the input requirements of the neural network.

[0116] Secondly, the preprocessed data is used to train a machine learning model using training, validation, and test sets. Through multiple iterations, the model parameters are optimized to better fit the bridge's seismic response data, thereby enabling prediction of unknown vibration conditions. When control is required, time-delay data processing needs to be introduced during model training to account for the dynamics of the actual vibration response and the delay effects in the process.

[0117] Then, based on the displacement, velocity, and acceleration vectors processed with time delay, a trained neural network model is used to calculate precise control commands. These commands are transmitted to the control device to achieve active vibration control of the bridge. By continuously iterating and adjusting the control commands, and combining them with vibration response data obtained from real-time monitoring, the control effect is further optimized, improving the accuracy and stability of vibration control.

[0118] Finally, based on the dynamic model of bridge vibration and real-time monitoring data, it is determined whether the bridge's vibration response meets the preset threshold. If it does not meet the requirements, the corresponding control strategy is adjusted, and the control commands are recalculated to achieve more precise vibration control. The seismic resistance method for hollow thin-walled high-pier bridges provided in this embodiment can effectively reduce the vibration response of bridges under seismic loading, improve their structural safety and seismic resistance, and provide a new technical means for the design and practice of bridge engineering.

[0119] like Figure 8 As shown in the figure, the specific implementation process of the seismic resistance method for hollow thin-walled high-pier bridges provided in this embodiment is as follows:

[0120] Step S1: Collect response data of the bridge structure under seismic loading, including stress S. i (t), displacement X i (t), acceleration Data calibration and standardization are performed to convert the raw signals output by the sensors into actual physical quantities. Preprocessing is then used to add velocity data. The raw data is transformed into feature representations suitable for processing by neural network models. This preprocessing improves data quality and reliability, ensuring the accuracy of subsequent machine learning model training.

[0121] Step S2: Divide the dataset after preprocessing in Step S1 into training set, validation set and test set for machine learning training. 70% to 80% of the data is used for training set, 10% to 15% of the data is used for validation set, and the remaining proportion is divided into test set.

[0122] During training, training data samples are input into the model to obtain prediction results, and the loss function is calculated. Then, the model parameters are updated according to the value of the loss function and the parameter update rules. This process is repeated multiple times until a stopping condition is met. By iteratively optimizing the model parameters, the model can better fit the seismic response data of the bridge structure. After training is completed, a machine-trained model is obtained, which can predict the values ​​of physical quantities such as stress, displacement, and acceleration based on the input seismic response data of the bridge structure.

[0123] Step S3: Set thresholds based on specifications and machine learning models. If one or more physical quantities—stress, displacement, acceleration, or velocity—exceed the threshold, the active control process begins; otherwise, active control is skipped.

[0124] Step S4: To control the damper and thus control bridge vibration, time delay effects need to be considered, and a neural network considering time delay effects needs to be trained. Assuming the vibration period of the bridge's first-order frequency mode is T, the time delay length τ is set to 1 / 4 or 1 / 5 of T. The preprocessed data is then subjected to time delay processing to construct time-delay data.

[0125] Define loss function Among them, U i It is the control command vector, U 目标，i It is the expected control command vector.

[0126] The gradient descent optimization method is used to adjust the parameters θ of the neural network by minimizing the loss function.

[0127] Gradient formula

[0128] Where θ-t is the parameter value of the t-th iteration;

[0129] θ-{t+1} is the parameter value after iteration t+1;

[0130] α is the learning rate, which controls the magnitude of parameter updates;

[0131] It is the gradient of the loss function J with respect to the parameter θ at the point θ-t, expressed as a vector.

[0132] The specific steps are: initialize the parameter θ, calculate... Update θ-{t+1}=θ-t-α▽J(θ-t), and repeat the iteration until the change in the loss parameter is no longer significant, thus finding the optimal parameter configuration of the model.

[0133] Step S5: Based on step S4, the neural network controller model is represented as follows:

[0134] U i (t)=fθ (X i (t-τ), f θ This represents a neural network model with parameter θ.

[0135] The displacement X after time delay processing i (t-τ), velocity acceleration The vector is used as input, and the control command is calculated through a neural network model.

[0136] Step S6: Control command U i (t) Output;

[0137] Step S7: The damper executes the control command. According to command U... i (t), calculate the damper adjustment value, such as the current command I. i (t), I i (t)=K i ×U i (t)+I0(t). Where K i I0 is the proportional constant of the current command. By adjusting it, the amplitude of the current can be controlled, which in turn can adjust the damping force and damping effect of the damper; I0(t) is the base current.

[0138] Step S8: Calculate the bridge's displacement, velocity, and acceleration responses based on the dynamic equations of motion, and determine if the vibration response is within a preset threshold range. If the vibration response meets the threshold requirements, the loop control ends; if the vibration response fails to meet the threshold requirements, the iteration and adjustment process continues until the requirements are met. In each iteration, based on the current response results and threshold requirements, the controller updates the control commands and transmits the updated commands to the dampers to adjust their damping effect. Through continuous iteration and adjustment, the control algorithm and model parameters are gradually optimized to achieve more precise vibration control, ensuring that the bridge's vibration under seismic loading remains within acceptable limits, thereby improving the structure's safety and stability.

[0139] Figure 9 This is a schematic diagram of the structure of a data processor provided in one embodiment of this application. Figure 9 As shown, the data processor provided in this application embodiment includes:

[0140] Acquisition unit 41 is used to acquire the actual response data of the bridge structure under seismic action collected by the vibration detection sensor at the first moment. The actual response data includes stress data, displacement data and / or acceleration data.

[0141] The first prediction unit 42 is used to input the actual response data into a pre-trained prediction model to obtain the response data of the bridge structure under earthquake action at the second time, output by the prediction model.

[0142] The second prediction unit 43 is used to input the response data of the second time into a pre-trained neural network model if at least one of the response data in the second time exceeds the safety threshold of the response data, so as to obtain the damping parameter control data output by the neural network model.

[0143] The transmitting unit 44 is used to send the damping parameter control data to the damping parameter controller, which is used to adjust the damping force of the damper according to the damping parameter control data.

[0144] The data processor provided in this application acquires the actual response data of the bridge structure under seismic loading, collected by vibration detection sensors in the first instance. This actual response data includes stress data, displacement data, and / or acceleration data. The actual response data is input into a pre-trained prediction model to obtain the bridge structure's response data under seismic loading in the second instance, output by the prediction model. If at least one response data point in the second instance exceeds a safety threshold, the second instance response data is input into a pre-trained neural network model to obtain damping parameter control data output by the neural network model. This damping parameter control data is then sent to a damping parameter controller, which adjusts the damping force of the damper based on the control data. This enables real-time monitoring and numerical calculation of bridge deformation, displacement, and / or vibration, and allows for adjustment of damper parameters, achieving intelligent and active control of the damper and improving the bridge's seismic resistance.

[0145] This embodiment discloses a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions, and when the program instructions are executed by a computer, the computer can perform the methods provided in the above-described method embodiments.

[0146] This embodiment provides a computer-readable storage medium storing a computer program that causes the computer to execute the methods provided in the above-described method embodiments.

[0147] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0148] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0149] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0150] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0151] In the description of this specification, the references to terms such as "an embodiment," "a specific embodiment," "some embodiments," "for example," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0152] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this application. It should be understood that the above descriptions are merely specific embodiments of this application and are not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A seismic device for hollow thin-walled high pier bridges, characterized in that: include: Diaphragms are installed inside hollow thin-walled high pier bridges; A damper mounted on the diaphragm; Vibration detection sensors, including stress sensors, displacement sensors and / or acceleration sensors, are installed on the hollow thin-walled high pier bridge; A data processor, electrically connected to the vibration detection sensor; The damping parameter controller is electrically connected to the data processor and the damper respectively.

2. The device according to claim 1, characterized in that The vibration detection sensor is installed in the middle of the bridge span, the upper part of the bridge pier, the middle part of the bridge pier, the lower part of the bridge pier, both sides of the expansion joint and / or the diaphragm.

3. The device according to claim 1 or 2, characterized in that: The stress sensor collects stress data of the bridge piers, diaphragms and main beams under the action of earthquake loads, and sends the stress data to the data processor; and / or The displacement sensor collects displacement data of the bridge piers, diaphragms and main beams in horizontal and vertical directions under the action of earthquake loads, and sends the displacement data to the data processor; and / or The acceleration sensor collects vibration data of the bridge piers, diaphragms and main beams under the action of earthquake loads, and sends the vibration data to the data processor.

4. The device according to claim 1, characterized in that The damper is a magnetically controlled fluid damper.

5. A method for earthquake resistance of hollow thin-walled high pier bridges, characterized in that: Based on the earthquake-resistant device for hollow thin-walled high pier bridges according to any one of claims 1 to 4, the method comprises: Acquire actual response data of the bridge structure under the action of an earthquake collected by the vibration detection sensor at the first time, wherein the actual response data includes stress data, displacement data and / or acceleration data; Inputting the actual response data into a pre-trained prediction model to obtain response data of the bridge structure under the earthquake action at a second time output by the prediction model; If at least one of the response data at the second time exceeds the safety threshold of the response data, inputting the response data at the second time into a pre-trained neural network model to obtain damping parameter control data output by the neural network model; The damping parameter control data is sent to a damping parameter controller, and the damping parameter controller is used to adjust the damping force of the damper according to the damping parameter control data.

6. The method according to claim 5, characterized in that The training process of the prediction model is as follows: Inputting the actual response data of the bridge structure under the action of earthquake collected by the vibration detection sensor at the first historical time into a preset prediction model to obtain the response data of the bridge structure under the action of earthquake at the second historical time predicted by the prediction model; Calculating a first predicted loss of the prediction model according to the response data at the second historical time predicted by the prediction model and the actual response data collected by the vibration detection sensor at the second historical time; Adjusting the model parameters of the prediction model according to the first prediction loss to obtain the prediction model after the first parameter adjustment; Inputting the actual response data of the bridge structure under the earthquake action collected by the vibration detection sensor at the second historical time into the prediction model after the first parameter adjustment, to obtain the response data of the bridge structure under the earthquake action at the third historical time predicted by the prediction model; Calculate the second prediction loss of the prediction model after the first parameter adjustment according to the response data of the third historical time predicted by the prediction model and the actual response data collected by the vibration detection sensor at the third historical time, wherein the time difference between the third historical time and the second historical time, the time difference between the second historical time and the first historical time, and the time difference between the second time and the first time are all equal; Adjusting the model parameters of the prediction model after the first parameter adjustment according to the second prediction loss to obtain the prediction model after the second parameter adjustment; And so on, until the prediction loss of the prediction model is less than the first target value or the number of model parameter adjustments of the prediction model reaches the second target value, a trained prediction model is obtained.

7. The method according to claim 5, characterized in that The training process of the neural network model is as follows: Performing time-delay processing on the actual response data of the bridge structure under the earthquake action collected by the vibration detection sensor at the fourth historical time to obtain first time-delay data; Inputting the first time-lag data into a preset neural network model to obtain damping parameter control data of a fourth historical time output by the neural network model; Calculating a first prediction loss of the neural network model according to the damping parameter control data of the fourth historical time and the ideal damping parameter control data of the fourth historical time; Adjusting the model parameters of the neural network model according to the first prediction loss to obtain the neural network model after the first parameter adjustment; Performing time-delay processing on the actual response data of the bridge structure under the earthquake action collected by the vibration detection sensor at the fifth historical time to obtain second time-delay data; Inputting the second time-lag data into the neural network model after the first parameter adjustment to obtain the damping parameter control data of the fifth historical time output by the neural network model; Calculating a second prediction loss of the neural network model after the first parameter adjustment according to the damping parameter control data of the fifth historical time and the ideal damping parameter control data of the fifth historical time; Adjusting the model parameters of the neural network model after the first parameter adjustment according to the second prediction loss to obtain the neural network model after the second parameter adjustment; And so on, until the prediction loss of the neural network model is less than the first target value or the number of model parameter adjustments of the neural network model reaches the second target value, a trained neural network model is obtained.

8. The method according to claim 7, characterized in that When time-delay processing is performed on actual response data, the time-delay length is 1 / 4 to 1 / 5 of the vibration period of the first-order frequency mode of the bridge.

9. The method according to claim 5, characterized in that The method further comprises: If each type of response data in the response data at the second time does not exceed the safety threshold of the response data, then continue to obtain actual response data of the bridge structure under the earthquake action collected by the vibration detection sensor at the second time.

10. A data processor, characterized in that: include: An acquisition unit, used to acquire actual response data of the bridge structure under the action of an earthquake collected by the vibration detection sensor at the first time, wherein the actual response data includes stress data, displacement data and / or acceleration data; A first prediction unit is used to input the actual response data into a pre-trained prediction model to obtain response data of the bridge structure under the earthquake at a second time output by the prediction model; A second prediction unit is configured to input the response data at the second time into a pre-trained neural network model to obtain damping parameter control data output by the neural network model if at least one response data among the response data at the second time exceeds a safety threshold of the response data; The sending unit is used to send the damping parameter control data to the damping parameter controller, and the damping parameter controller is used to adjust the damping force of the damper according to the damping parameter control data.

Images