Large-span steel truss arch bridge construction time-space correlation mechanism construction and linear control method

Abstract

The invention discloses a large-span steel truss arch bridge construction space-time correlation mechanism construction and linear control method, and relates to the technical field of bridge construction control. The environment temperature, the component mass and the actual elevation after beam falling of the first n hoisting stages are collected; calculating theoretical elevations before beam falling, theoretical elevations after beam falling and error adjustment variables in the first n hoisting stages; defining an interval, and performing normalization processing on the data to form a data set; dividing the data set to obtain a training set and a test set; constructing a space-time correlation mechanism model by using the training set; using the test set to test the model validity; predicting an error adjustment variable of the l control point in the (n + 1) th hoisting stage by using the model; and correcting the theoretical elevation before beam falling in the (n + 1) th hoisting stage to realize linear control. According to the method, the theoretical elevation under the condition of considering different influence weights of various space-time factors can be calculated, and the defect that a traditional construction prediction method cannot consider adverse effects of environment temperature, component quality and construction accumulative errors on the weights is overcome.

Description

Technical Field

[0001] The invention relates to the technical field of bridge construction control, and in particular to a construction method of a time-space correlation mechanism and a linear control method for a large-span steel truss arch bridge. Background Art

[0002] The steel truss arch bridge is composed of two structural systems, truss and arch, and has the stress characteristics of both truss and arch. The truss part bears the axial force, while the horizontal thrust of the arch reduces the mid-span bending moment, so that the mid-span solid web section mainly bears the axial pressure under the action of the dead load. It combines the advantages of truss and arch, mainly bearing axial force, which is conducive to saving materials. Compared with other structures of the same span beam bridge, the truss arch bridge saves more steel and has good adaptability.

[0003] The existing prediction methods for construction control of steel truss arch bridges are mainly based on the grey system theory method and the Kalman filter method. The grey system theory method can gray out various factors that affect the construction status of the bridge, collect these influencing factors and feed them back to the target vector in a centralized manner, and finally complete the entire prediction process through the target vector; the Kalman filter method extracts the real signal from the signal that has been polluted by noise and estimates the real state of the system. Both methods can be well applied in the construction of cable-stayed bridges, but for steel truss arch bridges, the influence of ambient temperature and component quality on construction errors needs to be considered. Due to the limitations of their own methods, these two methods cannot consider the influence weights of various factors.

[0004] In view of the above reasons, the present invention establishes a spatiotemporal correlation mechanism for the hoisting construction of a steel truss arch bridge on the basis of the traditional long short-term memory network. It can take into account the influence of ambient temperature, component quality and cumulative construction errors while considering the influence of historical hoisting errors, avoiding the gradient explosion and gradient vanishing linearity of other neural network models, and combining the Dropout method and self-attention mechanism to prevent model overfitting, increase the influence weight of hoisting errors in nearby times, and improve the accuracy of model prediction. Summary of the invention

[0005] In order to address the shortcomings of the background technology and aiming at the problem of theoretical elevation calculation for construction control of large-span steel truss arch bridges, the present invention provides a method for constructing a spatiotemporal correlation mechanism and a linear control method for the construction of large-span steel truss arch bridges. The method can calculate the theoretical elevation under different influence weights of various spatiotemporal factors, thereby making up for the fact that traditional construction prediction methods cannot consider the adverse effects of ambient temperature, component quality and cumulative construction errors on weights.

[0006] To achieve the above object, the present invention adopts the following technical scheme: a method for constructing a time-space correlation mechanism and linear control for the construction of a large-span steel truss arch bridge, comprising the following steps:

[0007] Step 1: Collect measured data

[0008] Define the hoisting stage i=1,2,...,n-1,n, each hoisting stage has a control point j=1,2,...,l-1,l, collect the ambient temperature T, component mass M and the actual elevation f after the beam is dropped at n×l control points in the first n hoisting stages 3 , the data formats are as follows:

[0009] T=(t 1 ,t 2 …,t n-1 ,t n ) T ,M=(m 1 ,m 2 …,m n-1 ,m n ) Τ ,

[0010] Where, t i and m i denote the ambient temperature and component mass of the i-th lifting stage, respectively. It represents the actual elevation of the jth control point after the beam is dropped in the i-th lifting stage;

[0011] Step 2: Calculate theoretical data

[0012] Calculate the theoretical elevation f of n×l control points before beam drop in the first n hoisting stages 1 , Theoretical elevation after beam drop f 2 And the error adjustment variable δ, the data forms are as follows:

[0013]

[0014] In the formula, and δ i,j They represent the theoretical elevation before beam drop, theoretical elevation after beam drop and error adjustment variable of the jth control point in the i-th lifting stage respectively;

[0015] Step 3: Data processing to form a data set

[0016] Define the interval [x min ,x max ],in:

[0017] x max =Q 3 +2×IQR,x min =Q 1 -2×IQR

[0018] In the formula, IQR represents the interquartile range, IQR = Q 3-Q 1 , Q 3 is the third quantile, Q 1 is the first quantile;

[0019] The data beyond the range are marked as outliers and deleted, processed by the mean filling method, and then the data is normalized. The normalized data set is as follows:

[0020]

[0021] In the formula, and are the normalized theoretical elevation before beam drop, ambient temperature, component mass and error adjustment variables;

[0022] Step 4: Divide the dataset

[0023] The normalized data set X sd Divide the training set X into time series T and the test set X E , define the lifting stage i in the training set 1 =1,2,...,k-1,k, then:

[0024]

[0025] In the formula, and They are the normalized theoretical elevation before beam drop, ambient temperature, component mass and error adjustment variables in the training set;

[0026] Define the hoisting stage i in the test set 2 =k+1,k+2,...,n-1,n, then:

[0027]

[0028] In the formula, and They are the normalized theoretical elevation before beam drop, ambient temperature, component mass and error adjustment variables in the test set;

[0029] Step 5: Use training set X T Constructing a spatiotemporal correlation mechanism model

[0030] Specify the training set X T In is the input data, training set X T In To output data, a Dropout layer is established to regularize the input data:

[0031]

[0032] Where ο is the Hadamard product, and D is a random vector drawn from the Bernoulli distribution;

[0033] Establish a long short-term memory layer and define the input weight matrices of the forget gate, input gate, unit candidate and output gate as W f , W i , W c , W o , the hidden weight matrices are U f , U i , U c , U o , the bias terms are b f 、b i 、b c 、b o , X i' As the input of the long short-term memory layer i' time step, we have:

[0034] Forget Gate:

[0035] F i' =σ(W f X i' +U f H i'-1 +b f )

[0036] Input Gate:

[0037] I i' =σ(W i X i' +U i H i'-1 +b i )

[0038] Unit Candidates:

[0039]

[0040] Unit Status:

[0041]

[0042] Output Gate:

[0043] O i' =σ(W o X i' +U o H i'-1 +b o )

[0044] Hidden state:

[0045] H i' =Oi' tanh(C i' )

[0046] Where σ(·) is the sigmoid function, tanh(·) is the hyperbolic tangent function, and H i'-1 is the hidden state of the long short-term memory layer i'-1 time step, C i'-1 is the cell state of the long short-term memory layer i'-1 time step;

[0047] Establish a self-attention mechanism layer and calculate the attention weight:

[0048]

[0049] Applying it to the hidden state of the long short-term memory layer, we get the hidden state after the self-attention mechanism is processed:

[0050]

[0051] In the formula, softmax(·) is the activation function, is the weight vector, w a is the weight matrix, b a As the bias term, use and Calculate the gradient of the loss function and update W through the back-propagation algorithm f , W i , W c , W o and U f , U i , U c , U o and b f , b i , b c , b o , iterate until the gradient is 0, completing the construction and training of the spatiotemporal correlation mechanism model;

[0052] Step 6: Use the test set X E Verify model validity

[0053] The test set X E Input the spatiotemporal correlation mechanism model and output the predicted value of the normalized error adjustment variable Calculate its If the root mean square error meets the required accuracy requirements, proceed to step seven, otherwise return to step five to update the spatiotemporal correlation mechanism model parameters until the root mean square error meets the required accuracy requirements;

[0054] Step 7: Use the model to predict the error adjustment variable of the l control point in the n+1th lifting stage

[0055] The theoretical elevation, ambient temperature and component mass before the beam is dropped at the l control point in the n+1th hoisting stage are normalized as follows:

[0056]

[0057] In the formula, is the normalized theoretical elevation of the l control point before beam drop in the n+1th hoisting stage, and are the normalized ambient temperature and component mass of the n+1th lifting stage, respectively;

[0058] Will Input the spatiotemporal correlation mechanism model, output the predicted value of the normalized error adjustment variable of the l control points in the n+1th lifting stage, and perform inverse normalization to obtain the predicted error adjustment variable δ n+1,j ;

[0059] Step 8: Correct the theoretical elevation before beam drop in the n+1th hoisting stage to achieve linear control

[0060] The theoretical elevation of the l control point before the beam is dropped in the n+1th hoisting stage is corrected and calculated according to the following formula:

[0061]

[0062] In the formula, The theoretical elevation of the jth control point before the beam is dropped in the n+1th lifting stage, using the theoretical elevation correction value before the beam is dropped Replace the theoretical elevation before beam drop

[0063] Furthermore, in step 2 and δ i,j The calculation is as follows:

[0064]

[0065]

[0066]

[0067] In the formula, and They are the designed elevation, designed pre-camber, pre-throw height and theoretical elastic deformation of the j-th control point in the i-th lifting stage respectively.

[0068] Furthermore, in step 4, the normalized data set X sd The time series is divided in a ratio of 4:1, and the first 4 / 5 of the data is used as the training set X T, the last 1 / 5 of the data is used as the test set X E .

[0069] Furthermore, the inverse normalization process in step seven is expressed as follows:

[0070]

[0071] In the formula, δ n+1,j is the error adjustment variable predicted for the jth control point in the n+1th lifting stage, is the predicted value of the normalized error adjustment variable of the jth control point in the n+1th lifting stage, δ max is the maximum value of the error adjustment variable in the data set, δ min Minimum value of the error-adjusted variable in the data set.

[0072] Compared with the prior art, the invention has the following beneficial effects: the invention is applicable to the linear control of large-span steel truss arch bridges during hoisting construction, and can calculate the theoretical elevation under different influence weights of various time and space factors, making up for the fact that the traditional construction prediction method cannot consider the adverse effects of ambient temperature, component quality and construction cumulative error on weights, and can effectively improve the hoisting construction accuracy, and has the following advantages:

[0073] 1. The constructed spatiotemporal correlation mechanism model only needs to collect basic parameters such as theoretical elevation, ambient temperature, and component quality, without the need for complex finite element modeling and nonlinear analysis;

[0074] 2. The constructed spatiotemporal correlation mechanism model can take into account the influence weights of different factors and the impact of historical construction errors, which can effectively improve the accuracy of hoisting construction and make the alignment of the bridge closer to the ideal state;

[0075] 3. The constructed spatiotemporal correlation mechanism model can focus on the influence of ambient temperature, thereby accurately predicting the error adjustment variables under the specified ambient temperature and proposing a more reasonable theoretical elevation. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 is a flow chart of the method of the present invention;

[0077] Figure 2 3 is a comparison chart of the predicted value and the true value of the test set data in the embodiment. DETAILED DESCRIPTION

[0078] The technical solution of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0079] like Figure 1 As shown in the figure, the construction method of time-space correlation mechanism and linear control of long-span steel truss arch bridge construction includes the following steps:

[0080] Step 1: Collect measured data

[0081] Define the hoisting stage i=1,2,...,n-1,n, each hoisting stage has a control point j=1,2,...,l-1,l, collect the ambient temperature T, component mass M and the actual elevation f after the beam is dropped at n×l control points in the first n hoisting stages 3 , the data formats are as follows:

[0082] T=(t 1 ,t 2 …,t n-1 ,t n ) T ,M=(m 1 ,m 2 …,m n-1 ,m n ) Τ ,

[0083] Where, t i represents the ambient temperature of the i-th lifting stage, m i represents the mass of the component in the i-th lifting stage, It represents the actual elevation of the jth control point after the beam is dropped in the i-th lifting stage;

[0084] Step 2: Calculate theoretical data

[0085] Calculate the theoretical elevation f of n×l control points before beam drop in the first n hoisting stages 1 , Theoretical elevation after beam drop f 2 And the error adjustment variable δ, the data forms are as follows:

[0086]

[0087] In the formula, It represents the theoretical elevation of the jth control point before the beam is dropped in the i-th lifting stage. represents the theoretical elevation of the jth control point after the beam is dropped in the i-th lifting stage, δ i,jrepresents the error adjustment variable of the jth control point in the i-th lifting stage and is calculated as follows:

[0088]

[0089]

[0090]

[0091] In the formula, is the design elevation of the jth control point in the i-th lifting stage, is the designed pre-camber of the jth control point in the i-th lifting stage, is the pre-throw height of the jth control point in the i-th lifting stage, is the theoretical elastic deformation of the jth control point in the i-th lifting stage;

[0092] Step 3: Data processing to form a data set

[0093] Define the interval [x min ,x max ],in:

[0094] x max =Q 3 +2×IQR,x min =Q 1 -2×IQR

[0095] In the formula, IQR represents the interquartile range, IQR = Q 3 -Q 1 , Q 3 is the third quantile, Q 1 is the first quantile.

[0096] The data beyond the range are marked as outliers and deleted, processed by the mean filling method, and then the data are normalized as follows:

[0097]

[0098] In the formula, x sd represents normalized data, x represents original data, including the theoretical elevation before beam drop Ambient temperature i , component mass m i and the error adjustment variable δ i,j .

[0099] The normalized data set is as follows:

[0100]

[0101] In the formula, is the normalized theoretical elevation before beam drop, is the normalized ambient temperature, is the normalized component mass, is the normalized error adjustment variable, all passed through x sd Solving calculation formulas;

[0102] Step 4: Divide the dataset

[0103] The normalized data set X sd The time series is divided in a ratio of 4:1, and the first 4 / 5 of the data is used as the training set X T , the last 1 / 5 of the data is used as the test set X E Of course, it is also feasible to use other ratios for division. The division ratio can be selected according to the actual situation. Define the lifting stage i in the training set 1 =1,2,...,k-1,k, then:

[0104]

[0105] In the formula, is the normalized theoretical elevation before beam drop in the training set, is the normalized ambient temperature in the training set, is the normalized component quality in the training set, The error adjustment variable normalized in the training set.

[0106] Define the hoisting stage i in the test set 2 =k+1,k+2,...,n-1,n, then:

[0107]

[0108] In the formula, is the normalized theoretical elevation before beam drop in the test set, is the normalized ambient temperature in the test set, is the normalized component quality in the test set, Error adjustment variable normalized in the test set;

[0109] Step 5: Use training set X T Constructing a spatiotemporal correlation mechanism model

[0110] Specify the training set X T In is the input data, training set X T In To output data, a Dropout layer is established to regularize the input data:

[0111]

[0112] Where ο is the Hadamard product, and D is a random vector drawn from a Bernoulli distribution, whose value is 0 or 1.

[0113] Establish a long short-term memory layer and define W f , W i , W c , W o are the input weight matrices of the forget gate, input gate, unit candidate, and output gate, respectively, U f , U i , U c , U o are the hidden weight matrices of the forget gate, input gate, unit candidate, and output gate, respectively, and b f , b i , b c , b o are the bias items of the forget gate, input gate, unit candidate and output gate respectively, and X i' As the input of the long short-term memory layer i' time step, we have:

[0114] Forget Gate:

[0115] F i' =σ(W f X i' +U f H i'-1 +b f )

[0116] Input Gate:

[0117] I i' =σ(W i X i' +U i H i'-1 +b i )

[0118] Unit Candidates:

[0119]

[0120] Unit Status:

[0121]

[0122] Output Gate:

[0123] O i' =σ(W o X i' +U o H i'-1 +b o )

[0124] Hidden state:

[0125] H i' =O i' tanh(C i' )

[0126] Where σ(·) is the sigmoid function, tanh(·) is the hyperbolic tangent function, and H i'-1 is the hidden state of the LSTM layer at time step i'-1, and its dimension is determined by the number of cells in the LSTM layer. i'-1 is the cell state of the long short-term memory layer i'-1 time step.

[0127] Establish a self-attention mechanism layer and calculate the attention weight:

[0128]

[0129] Applying it to the hidden state of the long short-term memory layer, we get the hidden state after the self-attention mechanism is processed:

[0130]

[0131] In the formula, softmax(·) is the activation function, is the weight vector, w a is the weight matrix, b a is the bias term.

[0132] use and Calculate the gradient of the loss function and update W through the back-propagation algorithm f , W i , W c , W o and U f , U i , U c , U o and b f 、b i 、b c 、b o , iterate until the gradient is 0, completing the construction and training of the spatiotemporal correlation mechanism model;

[0133] Step 6: Use the test set X E Verify model validity

[0134] The test set X E Input the spatiotemporal correlation mechanism model and output the predicted value of the normalized error adjustment variable Use the following formula to calculate its The root mean square error is:

[0135]

[0136] If the root mean square error MSE meets the required accuracy requirements, proceed to step seven, otherwise return to step five to update the spatiotemporal correlation mechanism model parameters until the root mean square error MSE meets the required accuracy requirements;

[0137] Step 7: Use the model to predict the error adjustment variable of the l control point in the n+1th lifting stage

[0138] The theoretical elevation, ambient temperature and component mass before the beam is dropped at the l control point in the n+1th hoisting stage are normalized as follows:

[0139]

[0140] In the formula, is the normalized theoretical elevation of the l control point before beam drop in the n+1th hoisting stage, is the normalized ambient temperature of the n+1th lifting stage, is the normalized mass of the components in the n+1th lifting stage, all of which are expressed by x sd Solve the calculation formula.

[0141] Will Input the spatiotemporal correlation mechanism model and output the predicted value of the normalized error adjustment variable of the l control points in the n+1th lifting stage And it is inverse normalized to obtain the predicted error adjustment variable, which is expressed as follows:

[0142]

[0143] In the formula, δ n+1,j is the error adjustment variable predicted for the jth control point in the n+1th lifting stage, is the predicted value of the normalized error adjustment variable of the jth control point in the n+1th lifting stage, δ max is the maximum value of the error adjustment variable in the data set, δ min The minimum value of the error-adjusted variable in the data set;

[0144] Step 8: Correct the theoretical elevation before beam drop in the n+1th hoisting stage to achieve linear control

[0145] The theoretical elevation of the l control point before the beam is dropped in the n+1th hoisting stage is corrected and calculated according to the following formula:

[0146]

[0147] In the formula, The theoretical elevation of the jth control point before the beam is dropped in the n+1th lifting stage, using the theoretical elevation correction value before the beam is dropped Replace the theoretical elevation before beam drop To achieve linear control.

[0148] Example

[0149] Taking the Chaobai River long-span steel truss arch bridge of Changtong Road in Beijing as an example, the error adjustment variable prediction was carried out, and the corrected theoretical elevation before beam drop was used for hoisting construction.

[0150] ① Collect the ambient temperature T, component mass M and the actual elevation f of the four control points after beam drop in the first 50 hoisting stages 3 , collect measured data as shown in Table 1:

[0151] Table 1 Collected measured data

[0152]

[0153]

[0154] ② Calculate the theoretical elevation f before beam drop in the first 50 lifting stages 1 , Theoretical elevation after beam drop f 2 As well as the error adjustment variable δ, some theoretical calculation data are shown in Table 2:

[0155] Table 2 Calculation theoretical data (partial)

[0156] Hoisting stage 1 2 3 4 5 <![CDATA[f 1 n,1 ]]> 32.403 32.369 32.262 32.343 32.387 <![CDATA[f 1 n,2 ]]> 32.286 32.504 32.376 32.484 32.267 <![CDATA[f 1 n,3 ]]> 32.382 32.351 32.271 32.359 32.389 <![CDATA[f 1 n,4 ]]> 32.271 32.484 32.383 32.455 32.274 <![CDATA[f 2 n,1 ]]> 32.383 32.355 32.252 32.335 32.384 <![CDATA[f 2 n,2 ]]> 32.273 32.488 32.371 32.482 32.259 <![CDATA[f 2 n,3 ]]> 32.376 32.334 32.262 32.354 32.388 <![CDATA[f 2 n,4 ]]> 32.265 32.484 32.370 32.439 32.258 <![CDATA[δ n,1 ]]> -0.085 -0.003 -0.224 32.335 32.384 <![CDATA[δ n,2 ]]> -0.001 0.114 0.090 32.482 32.259 <![CDATA[δ n,3 ]]> -0.110 -0.032 -0.221 32.354 32.388 <![CDATA[δ n,4 ]]> -0.126 0.218 0.002 32.439 32.258

[0157] ③ The data is normalized, and the partially normalized data set obtained is shown in Table 3:

[0158] Table 3 Normalized data set (part)

[0159] phase 1 2 3 4 <![CDATA[f 1 n,1 ]]> 0.115196 0.087418 0 0.066176 <![CDATA[f 1 n,2 ]]> 0.015807 0.197171 0.090682 0.180532 <![CDATA[f 1 n,3 ]]> 0.102024 0.076923 0.012146 0.083401 <![CDATA[f 1 n,4 ]]> 0.204502 0.341479 0.276527 0.32283 T 0.055556 0 0 0 M 1 0.454553 0.984466 0.503854 <![CDATA[f 2 n,1 ]]> 0.11156 0.088925 0.005659 0.072757 <![CDATA[f 2 n,2 ]]> 0.011647 0.190516 0.093178 0.185524 <![CDATA[f 2 n,3 ]]> 0.100324 0.066343 0.008091 0.082524 <![CDATA[f 2 n,4 ]]> 0.203215 0.344051 0.27074 0.315113 <![CDATA[δ n,1 ]]> 0.522727 0.363636 0.386364 0.204545 <![CDATA[δ n,2 ]]> 0.378378 0.702703 0.486486 1 <![CDATA[δ n,3 ]]> 0.651163 0.069767 0.906977 0.906977 <![CDATA[δ n,4 ]]> 0.447368 0.921053 0.578947 0

[0160] ④ The normalized data set X sd Divide the time series, and use the first 4 / 5 data as the training set X T , the last 1 / 5 of the data is used as the test set X E , then the lifting stage i in the training set 1 =1,2,...,40, hoisting stage i in the test set 2 =41,42,...,50, each lifting stage has control points j=1,2,3,4, training set X T and the test set X E They are respectively expressed as follows:

[0161]

[0162]

[0163] ⑤Call the tensorflow open source library to build a spatiotemporal correlation mechanism model and select the optimizer and loss function:

[0164] Create an input layer and specify the model’s input shape as [X.shape[1], X.shape[2]], where X.shape[1] is the number of time steps and X.shape[2] is the number of features per time step.

[0165] Create a Dropout layer and specify the loss parameters of this layer;

[0166] Create a long short-term memory layer, specify the number of neurons, and use the input of the Dropout layer as the output of this layer;

[0167] Add a self-attention mechanism layer, use the input of the long short-term memory layer as the output of this layer, and convert the output into two-dimensional data;

[0168] Create an output layer, which will input two-dimensional data and output data for four predicted targets;

[0169] Specify the time step as 1, the number of cycles as 40, and the number of training samples as 2 to train this model.

[0170] ⑥Call the trained spatiotemporal correlation mechanism model to predict the test set data, and combine the results Figure 2 shown.

[0171] ⑦ The time-space correlation mechanism model is used to predict the error adjustment variables of the four control points in the 51st lifting stage. The results are shown in Table 4:

[0172] Table 4. Forecast error adjustment variables

[0173] phase <![CDATA[δ 51,1 ]]> <![CDATA[δ 51,2 ]]> <![CDATA[δ 51,3 ]]> <![CDATA[δ 51,4 ]]> 51 0.008 -0.002 -0.020 -0.008

[0174] ⑧ Calculate the theoretical elevation correction value before beam drop, and use the correction value to replace the original theoretical elevation before beam drop to achieve linear control. The correction value results are shown in Table 5:

[0175] Table 5 Theoretical elevation correction value before beam drop

[0176]

[0177] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other forms of assembly without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered in all respects as exemplary and non-restrictive, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations within the meaning and range of equivalents of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

[0178] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. Construction of time-space correlation mechanism and linear control method for long-span steel truss arch bridge construction, characterized by: The following steps are involved: Step 1: Collect measured data Define the hoisting stage i=1,2,...,n-1,n, each hoisting stage has a control point j=1,2,...,l-1,l, collect the ambient temperature T, component mass M and the actual elevation f after the beam is dropped at n×l control points in the first n hoisting stages 3 , the data formats are as follows: T=(t1,t2…,t n-1 ,t n ) T ,M=(m1,m2…,m n-1 ,m n ) Τ , Where, t i and m i denote the ambient temperature and component mass of the i-th lifting stage, respectively. It represents the actual elevation of the jth control point after the beam is dropped in the i-th lifting stage; Step 2: Calculate theoretical data Calculate the theoretical elevation f of n×l control points before beam drop in the first n hoisting stages 1 , Theoretical elevation after beam drop f 2 And the error adjustment variable δ, the data forms are as follows: In the formula, and δ i,j They represent the theoretical elevation before beam drop, theoretical elevation after beam drop and error adjustment variable of the jth control point in the i-th lifting stage respectively; Step 3: Data processing to form a data set Define the interval [x min ,x max ],in: x max =Q3+2×IQR,x min =Q1-2×IQR In the formula, IQR represents the interquartile range, IQR = Q3-Q1, Q3 is the third quartile, Q1 is the first quartile; The data beyond the range are marked as outliers and deleted, processed by the mean filling method, and then the data is normalized. The normalized data set is as follows: In the formula, and are the normalized theoretical elevation before beam drop, ambient temperature, component mass and error adjustment variables; Step 4: Divide the dataset The normalized data set X sd Divide the training set X into time series T and the test set X E , define the lifting stage i1=1,2,...,k-1,k in the training set, then: In the formula, and They are the normalized theoretical elevation before beam drop, ambient temperature, component mass and error adjustment variables in the training set; Define the hoisting stage i2 in the test set as k+1, k+2, ..., n-1, n, then: In the formula, and They are the normalized theoretical elevation before beam drop, ambient temperature, component mass and error adjustment variables in the test set; Step 5: Use training set X T Constructing a spatiotemporal correlation mechanism model Specify the training set X T In is the input data, training set X T In To output data, a Dropout layer is established to regularize the input data: Where ο is the Hadamard product, and D is a random vector drawn from the Bernoulli distribution; Establish a long short-term memory layer and define the input weight matrices of the forget gate, input gate, unit candidate and output gate as W f , W i , W c , W o , the hidden weight matrices are U f , U i , U c , U o , the bias terms are b f , b i , b c , b o , X i' As the input of the long short-term memory layer i' time step, we have: Forget Gate: F i' =σ(W f X i' +U f H i'-1 +b f ) Input Gate: I i' =σ(W i X i' +U i H i'-1 +b i ) Unit Candidates: Unit Status: Output Gate: The i' =σ(W o X i' +U o H i'-1 +b o ) Hidden state: H i' =O i' ·tanh(C i' ) Where σ(·) is the sigmoid function, tanh(·) is the hyperbolic tangent function, and H i'-1 is the hidden state of the long short-term memory layer i'-1 time step, C i'-1 is the cell state of the long short-term memory layer i'-1 time step; Establish a self-attention mechanism layer and calculate the attention weight: Applying it to the hidden state of the long short-term memory layer, we get the hidden state after the self-attention mechanism is processed: In the formula, softmax(·) is the activation function, is the weight vector, w a is the weight matrix, b a As the bias term, use and Calculate the gradient of the loss function and update W through the back-propagation algorithm f , W i , W c , W o and U f , U i , U c , U o and b f , b i , b c , b o , iterate until the gradient is 0, completing the construction and training of the spatiotemporal correlation mechanism model; Step 6: Use the test set X E Verify model validity The test set X E Input the spatiotemporal correlation mechanism model and output the predicted value of the normalized error adjustment variable Calculate its If the root mean square error meets the required accuracy requirements, proceed to step seven, otherwise return to step five to update the spatiotemporal correlation mechanism model parameters until the root mean square error meets the required accuracy requirements; Step 7: Use the model to predict the error adjustment variable of the l control point in the n+1th lifting stage The theoretical elevation, ambient temperature and component mass before the beam is dropped at the l control point in the n+1th hoisting stage are normalized as follows: In the formula, is the normalized theoretical elevation of the l control point before beam drop in the n+1th hoisting stage, and are the normalized ambient temperature and component mass of the n+1th lifting stage, respectively; Will Input the spatiotemporal correlation mechanism model, output the predicted value of the normalized error adjustment variable of the l control points in the n+1th lifting stage, and perform inverse normalization to obtain the predicted error adjustment variable δ n+1,j ; Step 8: Correct the theoretical elevation before beam drop in the n+1th hoisting stage to achieve linear control The theoretical elevation of the l control point before the beam is dropped in the n+1th hoisting stage is corrected and calculated according to the following formula: In the formula, The theoretical elevation of the jth control point before the beam is dropped in the n+1th lifting stage, using the theoretical elevation correction value before the beam is dropped Replace the theoretical elevation before beam drop 2. The method for constructing a time-space correlation mechanism and controlling the linear shape of a long-span steel truss arch bridge according to claim 1 is characterized by: In step 2 and δ i,j The calculation is as follows: In the formula, and They are the designed elevation, designed pre-camber, pre-throw height and theoretical elastic deformation of the j-th control point in the i-th lifting stage respectively.

3. The method for constructing a time-space correlation mechanism and controlling the linear shape of a long-span steel truss arch bridge according to claim 1 is characterized by: In step 4, the normalized data set X sd The time series is divided in a ratio of 4:1, and the first 4 / 5 of the data is used as the training set X T , the last 1 / 5 of the data is used as the test set X E .

4. The method for constructing a time-space correlation mechanism and controlling the linear shape of a long-span steel truss arch bridge according to claim 1 is characterized by: The inverse normalization process in step 7 is expressed as follows: In the formula, δ n+1,j is the error adjustment variable predicted for the jth control point in the n+1th lifting stage, is the predicted value of the normalized error adjustment variable of the jth control point in the n+1th lifting stage, δ max is the maximum value of the error adjustment variable in the data set, δ min Minimum value of the error-adjusted variable in the data set.

Images