An intelligent construction method for the global time-varying temperature field of a long-span concrete bridge structure

Abstract

An intelligent construction method for the global time-varying temperature field of a long-span concrete bridge structure, which relates to the technical field of bridge structure temperature field prediction. Discretize the bridge and calculate the position coordinates; consider the solar radiation received by the entire bridge and divide it according to the position coordinates; collect measured environmental data; construct a data set and perform normalization processing; construct a physical constraint loss function and a data constraint loss function; construct a neural network model with the output target being the final predicted value; during the model training process, configure weight coefficients for the constraint terms and update until the loss converges to obtain the temperature prediction result; evaluate the neural network model and complete the training; use the neural network model to predict the temperature of the actual bridge. Establish an association mechanism for bridge temperature prediction, predict the bridge temperature at a future moment while considering historical temperature data, and introduce a physical mechanism to form a constraint loss with double constraints of physical data, thereby improving the prediction accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of predicting the temperature field of bridge structures, and specifically to an intelligent construction method for the global time-varying temperature field of long-span concrete bridge structures. Background Technique

[0002] With the development of bridge construction in China, more and more bridges have been put into use and become the main hubs connecting various places. With the long-term use of most bridges, the main tasks for bridges have gradually shifted from traditional bridge construction to bridge monitoring.

[0003] It is very necessary to consider the long-term operation and maintenance of bridges. However, in the actual use process of bridges, they are affected by various loads, which reduces the service life of bridges. The influence of temperature load on bridges is also an issue that cannot be ignored. The traditional method of monitoring bridge status based on sensor layout requires a large amount of upfront costs and has limited controllability. Therefore, it is very necessary to explore a simple method for obtaining the temperature of bridge structures.

[0004] With the rapid development of artificial intelligence, in data-driven modeling techniques, artificial neural networks can handle complex non-linear relationships, are applicable to various problem domains, can adapt to high-dimensional data, complex inputs, diverse forms, and have a certain tolerance for noise and missing data. They also have significant advantages in image processing. Therefore, proposing a convenient and fast bridge temperature prediction scheme based on neural networks is of great significance for the safety operation monitoring of bridges. Summary of the Invention

[0005] Aiming at the deficiencies of the traditional method of monitoring bridge temperature by arranging sensors, such as high construction and maintenance costs, long time-consuming for equipment layout and data acquisition, and low efficiency in analyzing massive data, the present invention provides an intelligent construction method for the global time-varying temperature field of long-span concrete bridge structures. It establishes an association mechanism for bridge temperature prediction on the basis of the traditional long short-term memory network, can predict the bridge temperature at a certain future moment while considering historical temperature data, and introduces a physical mechanism into the model to form a constraint loss with double constraints of physical data, effectively improving the accuracy of model prediction.

[0006] To achieve the above object, the present invention adopts the following technical solution: An intelligent construction method for the global time-varying temperature field of long-span concrete bridge structures, including the following steps:

[0007] Step 1: Calculate the position coordinates of the bridge

[0008] The bridge is discretized at equal intervals along the length direction L, width direction W, and vertical direction Z by N respectively. The position coordinates of a certain point are expressed as:

[0009]

[0010] Then the position coordinates of all points are expressed as:

[0011] X = (X0 X1…X N-1 X N ) T

[0012] Among them, for a certain element X i , i = 0, 1,..., N - 1, N, its form is:

[0013]

[0014] Step 2: Calculate solar radiation parameters

[0015] Calculate the solar radiation intensity I at time t solar , considering the solar radiation received by the entire bridge area, the solar radiation intensity I solar According to the division of position coordinates, it is expressed as:

[0016] I solar = (I0 I1…I N-1 I N ) T

[0017] Among them, for a certain element I i , its form is:

[0018]

[0019] Step 3: Collect measured environmental data

[0020] Collect the environmental data at different position coordinates of the bridge at time t, including environmental temperature T env , wind speed V mind , relative humidity RH and historical temperature data T history , where:

[0021] Environmental temperature T env is expressed as:

[0022] For a certain element its form is:

[0023]

[0024] Wind speed V mind is expressed as: V mind = (V0 V1…V N-1 V N ) T

[0025] For a certain element V i , its form is expressed as:

[0026]

[0027] The relative humidity RH is expressed as: RH = (R0 R1…R N-1 R N ) T

[0028] For a certain element R i , its form is expressed as:

[0029]

[0030] Historical temperature data T history is expressed as:

[0031] For a certain element its form is expressed as:

[0032]

[0033] Step Four: Construct a dataset

[0034] Construct the position coordinates X and their corresponding characteristic data, including solar radiation intensity I solar , ambient temperature T env , wind speed V mind , relative humidity RH and historical temperature data T history into a dataset, perform normalization processing to limit its range within the bridge dimensions, mark data outside this range as outliers and delete them. The dataset after normalization processing is expressed as follows:

[0035] G = [X I solar T env V mind RH T history ;

[0036] Step Five: Construct a physical constraint loss function

[0037] Define the heat conduction equation to describe the change of the internal temperature of the bridge over time and space considering only concrete, and convert it into a heat conduction constraint loss, expressed as:

[0038]

[0039] In the formula, N' is the number of data samples, ρ is the material density, c is the specific heat capacity, is the temperature prediction value, is the derivative of the temperature with respect to time, κ is the thermal conductivity, is the derivative of temperature in space, where x, y, and z respectively correspond to the length direction L, width direction W, and vertical direction Z of the discretized bridge;

[0040] Define convective heat transfer to describe the heat exchange boundary condition between the bridge surface and the air. The physical process is described by Newton's cooling law, and then construct the convective heat transfer constraint loss, expressed as:

[0041]

[0042] In the formula, h is the convective heat transfer coefficient, q i is the heat flux at the corresponding point, and T surface is the temperature distribution on the bridge surface, that is, the temperature prediction value n is the normal direction of the bridge surface;

[0043] Then, the physical constraint loss function is expressed as:

[0044] L physics = Loss conduct + Loss conv ;

[0045] Step 6: Construct the data constraint loss function

[0046] Construct the data constraint loss to measure the gap between the temperature prediction value and the actual temperature. Use the mean square error to represent the data constraint loss function, expressed as:

[0047]

[0048] Step 7: Construct the neural network model

[0049] Divide the normalized dataset G into a training set G T and a test set G E ;

[0050] Define the input target as the other data in the training set G T except for the historical temperature data T history ;

[0051] Define the output target as the final prediction value while the historical temperature data T history is used as the validation data;

[0052] Use a long short-term memory network to construct the neural network model and output the final prediction value expressed as:

[0053]

[0054] Among them, the temperature prediction value of a certain element Its form is:

[0055]

[0056] Step 8: Constructing physical data dual constraints

[0057] During the training of the neural network model, weight coefficients are configured for the physical constraint items and the data constraint items respectively, and the sum of the weight coefficients is set to 1. The loss function expression of the physical and data dual constraints is as follows:

[0058] L total =λ1Loss data +λ2L physics

[0059] Where λ1 is the weight coefficient of data constraint loss, and λ2 is the weight coefficient of physical constraint loss;

[0060] Use the gradient optimization algorithm to update to determine the optimal weights, and repeat the update on the entire normalized data set G until the loss converges to obtain the temperature prediction result It is expressed as:

[0061]

[0062] Among them, for an element Its form is:

[0063]

[0064] Step 9: Neural Network Model Evaluation

[0065] Using the test set G E To measure the ability of the neural network model, the mean square error is used to evaluate the neural network model. If the neural network model meets the mean square error condition during training, the training is completed. If not, backpropagation is continued until the condition is met to obtain a trained neural network model.

[0066] Step 10: Use the neural network model to make predictions

[0067] The temperature of the entire bridge at time point t′ is predicted based on the actual bridge working condition information, and the position coordinate X′, solar radiation intensity I s ' olar 、Ambient temperature T e ' nv , wind speed V m ' ind , relative humidity RH′ and historical temperature data T h ' istoryAfter normalization, the data set G′ is obtained. The data set G′ is input into the trained neural network model to obtain the final temperature prediction output.

[0068] Furthermore, in step 2, the solar radiation intensity I solar Direct radiation I dir , scattered radiation I diff and surface reflected radiation I ref The sum of , where:

[0069]

[0070] Where, t u is the Link turbidity coefficient, k a is the relative atmospheric pressure, m is the light mass, I0 is the solar constant, φ is the angle between the illuminated surface and the incident light, P is the atmospheric transparency coefficient, β s is the solar altitude angle, β n is the angle between the outer normal of the illuminated surface and the ground, r e is the ground reflectivity.

[0071] Furthermore, in step 7, a long short-term memory network is used to model time series data, and W is defined f , W i , W c and W o are the input weight matrices of the forget gate, input gate, memory unit update, and output gate, respectively, b f , b i , b c and b o are the bias terms of the forget gate, input gate, memory unit update and output gate respectively, then:

[0072] Input Gate:

[0073] i(t)=σ(W i [h(t-1),X(t)]+b i )

[0074] Forget Gate:

[0075] f(t)=σ(W f [h(t-1),X(t)]+b f )

[0076] Output Gate:

[0077] o(t)=σ(W o [h(t-1),X(t)]+b o )

[0078] Memory unit update:

[0079]

[0080] Hide status update:

[0081] h(t)=o(t)*tanh(C(t))

[0082] Output prediction:

[0083] The output h(t) of LSTM can get the final prediction value through the fully connected layer

[0084] Where σ(·) is the sigmoid function, tanh(·) is the hyperbolic tangent function, and W out is the weight of the output layer, b out is the bias term of the output layer.

[0085] Furthermore, in step nine, during the training of the neural network model, the mean square error (MSE) is defined to evaluate the performance of the neural network model every 10 epochs, as shown below:

[0086]

[0087] Where N valid is the number of samples in the test set, is the loss of the i-th sample.

[0088] Furthermore, in step 10, the final temperature prediction output is The data is displayed through a three-dimensional cloud map.

[0089] Compared with the existing technology, the beneficial effects of the present invention are: the present invention targets the significant time-varying characteristics, time history accumulation characteristics and spatial characteristics of bridge temperature, trains the neural network model input characteristics and model training algorithm, and applies them to the prediction of bridge structure temperature. Compared with the traditional sensor deployment method for monitoring bridge temperature, it can help save a lot of time and cost, and can infer the bridge temperature at a certain moment in the future based on existing data. In addition, physical mechanism constraints are introduced into the model to connect theoretical and actual bridges, help reasonably set boundary conditions, ensure the accuracy of model prediction, and achieve a time-saving, labor-saving and high-precision temperature monitoring effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0090] Figure 1 is a flow chart of the method of the present invention;

[0091] Figure 2 It is a three-dimensional cloud map of the final temperature prediction output obtained in the embodiment. DETAILED DESCRIPTION

[0092] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0093] As Figure 1 shown, an intelligent construction method for the full-domain time-varying temperature field of a long-span concrete bridge structure includes the following steps:

[0094] Step 1: Calculate the bridge position coordinates

[0095] The bridge is discretized at equal intervals along the length direction L, width direction W, and vertical direction Z by N. The position coordinates of a certain point can be expressed as:

[0096]

[0097] Then the position coordinates of all points can be expressed as:

[0098] X = (X0 X1…X N-1 X N ) T

[0099] Among them, for a certain element X i , i = 0, 1,..., N - 1, N, its form is:

[0100]

[0101] Step 2: Calculate the solar radiation parameters

[0102] Calculate the solar radiation intensity I solar at time t. Considering that the solar radiation intensity I solar is the sum of direct radiation I dir , scattered radiation I diff and ground-reflected radiation I ref , that is:

[0103] I solar = I dir + I diff + I ref

[0104] Among them,

[0105] For direct radiation I dir , considering the effects of impurities such as air molecules, aerosols, and dust in the atmosphere, its formula is defined as:

[0106]

[0107] Scattered radiation I diff Considering the influence of the top and bottom plates of the box girder, its formula is defined as:

[0108]

[0109] Surface reflected radiation I ref Considering the differences in surface morphology, its formula is defined as:

[0110]

[0111] In the formula, t u is the Linke turbidity coefficient, k a is the relative atmospheric pressure, m is the optical mass, I0 is the solar constant, φ is the angle between the irradiated surface and the incident light, P is the atmospheric transparency coefficient, β s is the solar altitude angle, β n is the angle between the outer normal of the irradiated surface and the ground, r e is the ground reflectivity.

[0112] Considering the solar radiation received by the entire bridge area, the solar radiation intensity I solar According to the division of position coordinates, it is expressed as:

[0113] I solar =(I0 I1…I N-1 I N ) T

[0114] where, for a certain element I i i = 0, 1,..., N - 1, N, its form is:

[0115]

[0116] Step 3: Collect measured environmental data

[0117] Collect the environmental data at different position coordinates of the bridge at time t, including the environmental temperature T env , wind speed V mind , relative humidity RH and historical temperature data T history , where:

[0118] Environmental temperature T env is expressed as:

[0119]

[0120] where, for a certain element its form is:

[0121]

[0122] Wind speed V mind Expressed as:

[0123] V mind =(V0 V1…V N-1 V N ) T

[0124] Where, for a certain element V i i = 0, 1, ..., N - 1, N, its form is:

[0125]

[0126] Relative humidity RH is expressed as:

[0127] RH=(R0 R1…R N-1 R N ) T

[0128] Where, for a certain element R i i = 0, 1, ..., N - 1, N, its form is:

[0129]

[0130] Historical temperature data T history Expressed as:

[0131]

[0132] Where, for a certain element Its form is:

[0133]

[0134] Step 4: Construct a data set

[0135] Construct the position coordinates X and the corresponding characteristic data, including solar radiation intensity I solar , ambient temperature T env , wind speed V mind , relative humidity RH and historical temperature data T history into a data set, and normalize the data and position coordinates in the data set so that their range is limited between the bridge dimensions, and the data outside this range is marked as outliers and deleted.

[0136] The normalization process is expressed as follows:

[0137]

[0138] In the formula, α normis the normalized data, α is the original data of each feature data, α min and α max are the minimum and maximum values of the original data of each feature data respectively.

[0139] The normalized data set is expressed as follows:

[0140] G=[XI solar T env V mind RH T history ]

[0141] Step 5: Constructing a physical constraint loss function

[0142] The heat conduction equation is defined to describe the time and space variations of the internal temperature of the bridge when only concrete is considered, which can be expressed as:

[0143]

[0144] Where ρ is the material density, c is the specific heat capacity, is the rate of change of temperature with time, κ is the thermal conductivity, is the Laplace operator of the temperature field, x, y, and z correspond to the length direction L, width direction W, and vertical direction Z of the discretized bridge, respectively.

[0145] Converting it into a loss function, the temperature prediction value is The heat conduction constraint loss is calculated by the following formula:

[0146]

[0147] Where N' is the number of data samples, is the time derivative of temperature, is the derivative of temperature in space,

[0148] Convective heat transfer is defined to describe the boundary conditions for heat exchange between the bridge surface and the air. The physical process is described by Newton's law of cooling, which is expressed as:

[0149]

[0150] Where h is the convective heat transfer coefficient, T surface is the temperature distribution on the bridge surface, that is, the temperature prediction value q i is the heat flow at the corresponding point, and n is the normal direction of the bridge surface.

[0151] Then, the convection heat transfer constraint loss is constructed so that the temperature prediction value satisfies the convection heat transfer law, which is expressed as:

[0152]

[0153] Then, the physical constraint loss function is expressed as:

[0154] L physics = Loss conduct + Loss conv

[0155] Step 6: Construct the data constraint loss function

[0156] Construct the data constraint loss to measure the gap between the predicted temperature value and the actual temperature, and use the mean square error to represent the data constraint loss function, which is expressed as:

[0157]

[0158] Step 7: Construct the neural network model

[0159] Divide the normalized dataset G into two subsets in a time series ratio of 4:1, and use 80% of the data as the training set G T , and 20% of the data as the test set G E ;

[0160] Define the input target as the data in the training set G T except for the historical temperature data T history , that is, the solar radiation intensity I solar , the ambient temperature T env , the wind speed V mind , the relative humidity RH, and the position coordinate X;

[0161] Define the output target as the final predicted value while the historical temperature data T history is used as the validation data;

[0162] Use the long short-term memory network to model the time series data, and define W f , W i , W c and W o as the input weight matrices of the forget gate, input gate, memory cell update, and output gate respectively, and b f , b i , b c and b o as the bias terms of the forget gate, input gate, memory cell update, and output gate respectively, then there is:

[0163] Input gate:

[0164] i(t) = σ(W i[h(t - 1), X(t)] + b i )

[0165] Forgotten gate:

[0166] f(t) = σ(W f [h(t - 1), X(t)] + b f )

[0167] Output gate:

[0168] o(t) = σ(W o [h(t - 1), X(t)] + b o )

[0169] Memory cell update:

[0170]

[0171] Hidden state update:

[0172] h(t) = o(t) * tanh(C(t))

[0173] Output prediction:

[0174] The output h(t) of the LSTM can obtain the final predicted value through the fully connected layer

[0175] In the formula, σ(·) is the sigmoid function, tanh(·) is the hyperbolic tangent function, W out is the weight of the output layer, b out is the bias term of the output layer, which is expressed as:

[0176]

[0177] Among them, for the temperature predicted value of a certain element its form is:

[0178]

[0179] Step 8: Construct the double constraints of physical data

[0180] The temperature predicted value is initially obtained After that, the physical constraint loss function in Step 5 and the data constraint loss function in Step 6 are called for further fitting.

[0181] During the training process of the neural network model, weight coefficients are respectively configured for the physical constraint term and the data constraint term, and the sum of each weight coefficient is set to 1. Then the loss function expression of the double constraints of physical data is as follows:

[0182] L total= λ1Loss data + λ2L physics

[0183] Where λ1 is the weight coefficient of the data constraint loss, and λ2 is the weight coefficient of the physical constraint loss.

[0184] Use the gradient optimization algorithm to update to determine the optimal weights, and repeat the update on the entire normalized dataset G until the loss converges to obtain the temperature prediction result. It is expressed as:

[0185]

[0186] Among them, for a certain element Its form is:

[0187]

[0188] Step Nine: Neural Network Model Evaluation

[0189] Use the test set G E To measure the ability of the neural network model. During the training process, every 10 epochs, define the mean squared error MSE to evaluate the performance of the neural network model, which is expressed as follows:

[0190]

[0191] Where N valid Is the number of samples in the test set, Is the loss of the i-th sample.

[0192] If the neural network model satisfies the mean squared error condition during training, the training is completed. If not, continue backpropagation until the condition is met to obtain the trained neural network model.

[0193] Step Ten: Use the Neural Network Model for Prediction

[0194] Based on the trained neural network model, perform full-bridge temperature prediction for a certain time point t′ for the actual bridge working conditions information.

[0195] Normalize the position coordinates X′, solar radiation intensity I s ′ olar , ambient temperature T e ′ nv , wind speed V m ′ ind , relative humidity RH′ and historical temperature data T h ′ istory After processing, the obtained dataset G′ is expressed as follows:

[0196] G′ = [X′I s ′ olar T e ′ nv V m ′ ind RH′T h ′ istory

[0197] Input the dataset G′ into the trained neural network model to obtain the final temperature prediction output It is expressed as:

[0198]

[0199] Finally, a three-dimensional cloud map at this time point is obtained through data visualization for display.

[0200] Embodiment

[0201] Taking a concrete bridge in Hebei Province as an example, when it is summer and the time is 12:00 noon, predict the overall temperature of the bridge at 12:00 noon the next day.

[0202] (1) Calculate the bridge location coordinates

[0203] It is known through inquiry that the latitude of this bridge location is 39°N, the area is in a temperate monsoon climate, the solar declination angle is 23.44°, the bridge runs from northwest to southeast, the included angle between the longitudinal axis and the north-south direction is 30°, the length direction L of the bridge is 100m, the width direction W is 10m, and the vertical Z is 10m. Discretize them at equal intervals of N = 10. The position coordinates of all points can be expressed as:

[0204] X = (X0 X1…X9 X 10 ) T

[0205] For a certain element X i , its form is:

[0206]

[0207] Taking X1 as an example, it is expressed as

[0208] And so on, the position coordinates of the entire bridge can be obtained.

[0209] (2) Calculate the solar radiation parameters

[0210] By inputting the atmospheric parameters and the position coordinates of the entire bridge, the solar radiation intensity I solar According to the division of the position coordinates, it is expressed as:

[0211] ​

[0212] (3) Collect measured environmental data

[0213] When the time is 12 noon and the environmental temperature is 33.4 degrees Celsius, the environmental temperature T env The element T in e1 is expressed as:

[0214]

[0215] The wind speed V mind The element V1 in is expressed as:

[0216]

[0217] The element R1 in the relative humidity RH is expressed as:

[0218]

[0219] Collect the temperature data of the bridge in the past week and construct the historical temperature data T history , where the element T h1 is expressed as:

[0220]

[0221] (4) Construct a data set

[0222] Normalize the data, and the partial results are shown in Table 1 below:

[0223] Table 1 Normalized data set (partial)

[0224] Data 1 2 3 4 <![CDATA[I solar ]]> 0.19925 0.06082 0.05385 0.10750 <![CDATA[V mind ]]> 0.38322 0.19717 0.00228 0.15428 RH 0.18340 0.076923 0.22215 0.08157 <![CDATA[T history > 0.15266 0.063122 0.129991 0.02112

[0225] The normalized data set G can be obtained: G = [X I solar T env V mind RH T history

[0226] (5) Construct a physical constraint loss function

[0227] This bridge is made of C40 concrete, with a density of 2400 (kg / m 3 ), a specific heat capacity of 900 (J / kg·k), a thermal conductivity of 1.8 (W / m·k), and a convective heat transfer coefficient of 34 (W / m 2 ·K) under this working condition. Introduce them into the physical formula.

[0228] Then the heat conduction equation: Substitute the parameters to obtain the loss function:

[0229] ​

[0230] For the convective heat transfer equation, substituting the parameters gives the loss function:

[0231]

[0232] (6) Construct the data constraint loss function

[0233]

[0234] (7) Construct the neural network model

[0235] Divide the normalized dataset G into two subsets according to a time series ratio of 4:1. Take 80% of the data as the training set G T , and 20% of the data as the test set G E .

[0236] Define the input target as the data in the training set G T except for the historical temperature data T history in it.

[0237] Define the output target as the final predicted value while the historical temperature data T history is used as the validation data.

[0238] (8) Construct the double constraint of physical data

[0239] Call the physical constraint loss function and the data constraint loss function, and obtain the temperature prediction value through preliminary steps for weight allocation. During the training process of the neural network model, configure weight coefficients for the physical constraint term and the data constraint term respectively. Initially, take 0.5 and 0.5 respectively, limit the sum to 1, specify the time step as 1, and the number of loops as 100 times. Train this model, and the training results are shown in Table 2 below:

[0240] Table 2 Weight allocation results (partial)

[0241]

[0242]

[0243] When the weight allocation is λ1 = 0.502392344 and λ2 = 0.497607656, the loss is minimized, the weight allocation is completed, and the final data training is performed. The final prediction results are shown in Table 3 below:

[0244] Table 3 Final predicted temperature results (partial)

[0245]

[0246] (9) Neural network model evaluation

[0247] Use the test set G E to measure the ability of the neural network model. During the training process, every 10 epochs, the mean square error (MSE) is defined to evaluate the performance of the neural network model. The training results are shown in Table 4 below:

[0248] Table 4 Error verification results (partial)

[0249]

[0250] When training reaches 91 times, the loss and mean square error remain unchanged, that is, the mean square error condition is satisfied, and the optimal solution of the model is obtained, completing the training of the neural network model.

[0251] (10) Use the neural network model for prediction

[0252] Based on the trained neural network model, predict the full-bridge temperature at 12:00 on the second day. The local ambient temperature is known to be 34.5 °C.

[0253] Normalize the position coordinates X′, solar radiation intensity I s ′ olar , ambient temperature T e ′ nv , wind speed V m ′ ind , relative humidity RH′ and historical temperature data T h ′ istory at this time point, and the obtained data is shown in Table 5 below:

[0254] Table 5 Data parameters (partial)

[0255]

[0256]

[0257] Input the processed data set into the trained neural network model to obtain the final temperature prediction output, and finally visualize it with a three-dimensional cloud map as shown in Figure 2 shown.

[0258] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other forms without departing from the spirit or basic characteristics of the present invention. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent conditions of the claims are intended to be encompassed by the present invention. Any reference signs in the claims should not be construed as limiting the claimed rights.

[0259] In addition, it should be understood that although this specification is described according to embodiments, not every embodiment only contains an independent technical solution. This narrative manner of the specification is only for clarity. Those skilled in the art should regard the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. An intelligent construction method for the global time-varying temperature field of a long-span concrete bridge structure, characterized in that: It includes the following steps: Step 1: Calculate the bridge position coordinates The bridge is discretized at equal intervals along the length direction L, width direction W, and vertical direction Z by N. The position coordinates of a certain point are expressed as: Then the position coordinates of all points are expressed as: X = (X0 X1…X N-1 X N ) T Among them, for a certain element X i , where i = 0, 1, ..., N - 1, N, it is expressed in the form of: Step 2: Calculate the solar radiation parameters Calculate the solar radiation intensity I at time t solar , considering the solar radiation received by the entire bridge area, obtain the solar radiation intensity I solar According to the division of position coordinates, it is expressed as: I solar =(I0 I1…I N-1 I N ) T Among them, for a certain element I i , its form is as follows: Step 3: Collect measured environmental data Collect environmental data at different position coordinates of the bridge at time t, including environmental temperature T env , wind speed V mind , relative humidity RH and historical temperature data T history , where: Ambient temperature T env Expressed as: For a certain element Its form is as follows: Wind speed V mind It is expressed as: V mind =(V0 V1…V N-1 V N ) T For a certain element V i , it is presented in the form of: Relative humidity RH is expressed as: RH=(R0 R1…R N-1 R N ) T For a certain element R i , it is manifested in the form of: Historical temperature data T history Expressed as: For an element Its form is: Step 4: Construct a dataset The position coordinate X and its corresponding characteristic data, including the solar radiation intensity I solar 、Ambient temperature T env , wind speed V mind , relative humidity RH and historical temperature data T history A dataset was constructed and normalized to limit its range to the bridge size. Data outside this range were marked as outliers and deleted. The normalized dataset is shown below: G = [X I solar T env V mind RH T history ; Step 5: Construct a physical constraint loss function Define the heat conduction equation to describe the change of the internal temperature of the bridge with time and space considering only concrete, and convert it into a heat conduction constraint loss, which is expressed as: where N' is the number of data samples, ρ is the material density, c is the specific heat capacity, is the predicted temperature value, is the derivative of temperature with respect to time, κ is the thermal conductivity, is the derivative of temperature with respect to space, x, y, and z correspond to the length direction L, width direction W, and vertical direction Z of the discretized bridge, respectively; Define the convective heat transfer to describe the heat exchange boundary condition between the bridge surface and the air. The physical process is described by Newton's cooling law, and then construct the convective heat transfer constraint loss, which is expressed as: where h is the convective heat transfer coefficient, and q i is the heat flux at the corresponding point, and T surface is the temperature distribution on the bridge surface, i.e., the temperature prediction value n is the normal direction of the bridge surface; Then, the physical constraint loss function is expressed as: L physics = Loss conduct + Loss conv ; Step 6: Construct a data constraint loss function Construct a data constraint loss to measure the gap between the predicted temperature and the actual temperature. Use the mean square error to represent the data constraint loss function, which is expressed as: Step 7: Construct a neural network model Divide the normalized dataset G into a training set G T and a test set G E ; Define the input target as the training set G T except for the historical temperature data T history in the others; Define the output target as the final predicted value while the historical temperature data T history is used as the validation data; Build a neural network model using a long short-term memory network and output the final predicted value It is expressed as: Among them, the predicted temperature value for a certain element It is expressed in the form of: Step 8: Construct a physical-data double constraint During the training process of the neural network model, weight coefficients are configured for the physical constraint term and the data constraint term respectively, and the sum of all weight coefficients is set to 1. Then the loss function expression of the physical-data double constraint is as follows: L total =λ1Loss data +λ2L physics In the formula, λ1 is the weight coefficient of the data constraint loss, and λ2 is the weight coefficient of the physical constraint loss; Use the gradient optimization algorithm to update to determine the optimal weights, and repeat the update on the entire normalized dataset G until the loss converges to obtain the temperature prediction result It is expressed as: Among them, for a certain element Its form is as follows: Step 9: Evaluate the neural network model Using the test set G E To measure the ability of the neural network model, the mean square error is used to evaluate the neural network model. If the neural network model meets the mean square error condition during training, the training is completed. If not, backpropagation is continued until the condition is met to obtain a trained neural network model. Step 10: Use the neural network model for prediction Perform full-bridge temperature prediction at time point t′ for the actual bridge working condition information, and normalize the position coordinate X′, solar radiation intensity I s ′ olar , ambient temperature T e ′ nv , wind speed V m ′ ind , relative humidity RH′ and historical temperature data T h ′ istory After normalization, the dataset G′ is obtained. The dataset G′ is input into the trained neural network model to obtain the final temperature prediction output 2. The intelligent construction method for the global time-varying temperature field of a long-span concrete bridge structure according to claim 1, characterized in that: In the second step, the solar radiation intensity I solar is the direct radiation I dir , the diffuse radiation I diff and the ground reflected radiation I ref sum, where: where t u is the Linke turbidity coefficient, k a is the relative atmospheric pressure, m is the light quality, I0 is the solar constant, φ is the angle between the irradiated surface and the incident light, P is the atmospheric transparency coefficient, β s is the solar altitude angle, β n is the angle between the outer normal of the irradiated surface and the ground, r e is the ground reflectivity.

3. An intelligent construction method for the global time-varying temperature field of a long-span concrete bridge structure according to claim 1, characterized in that: In step seven, a long short-term memory network is used to model time series data, and W f , W i , W c and W o are the input weight matrices of the forget gate, input gate, memory cell update, and output gate respectively, and b f , b i , b c and b o are the bias terms of the forget gate, input gate, memory cell update, and output gate respectively. Then we have: Input gate: i(t) = σ(W i [h(t - 1), X(t)] + b i ) Forget gate: f(t) = σ(W f [h(t - 1), X(t)] + b f ) Output gate: o(t) = σ(W o [h(t - 1), X(t)] + b o ) Memory cell update: Hidden state update: h(t) = o(t) * tanh(C(t)) Output prediction: The output h(t) of the LSTM can obtain the final predicted value through the fully connected layer where, σ(·) is the sigmoid function, tanh(·) is the hyperbolic tangent function, W out is the weight of the output layer, b out is the bias term of the output layer.

4. An intelligent construction method for the global time-varying temperature field of a long-span concrete bridge structure according to claim 1, characterized in that: In step 9, during the training process of the neural network model, every 10 epochs, define the mean square error MSE to evaluate the performance of the neural network model, which is expressed as follows: Where N valid is the number of samples in the test set, and is the loss of the i-th sample.

5. The method for intelligently constructing the global time-varying temperature field of a long-span concrete bridge structure according to claim 1 is characterized by: In step 10, the final temperature prediction output is The data is displayed through a three-dimensional cloud map.

Images