Safety early warning method for underneath pass of existing important building based on digital twinning construction

Abstract

The invention provides a digital twinning-based safety early warning method for underneath pass of an existing important building during construction, which comprises the following steps: collecting historical engineering case data, and constructing an engineering case database, the historical engineering case data comprising environmental data, construction parameters and construction monitoring data; predicting the environmental data and the construction parameters in the engineering case database through a risk prediction model to obtain a prediction result; determining a relationship between risk levels according to the environmental data and the construction monitoring data in the engineering case database, and constructing a risk evaluation system based on a neural network; inputting the prediction result into the risk evaluation system for safety early warning; and if the condition corresponding to the risk level is triggered, issuing early warning information. According to the method, the accuracy of prediction and risk grade evaluation is improved.

Description

Technical Field

[0001] The present invention relates to the field of civil engineering monitoring technology, and in particular to a safety early warning method for construction underpasses of existing important buildings based on digital twins. Background Art

[0002] As the integrated three-dimensional transportation network continues to improve, more and more tunnels are forced to pass under existing buildings (structures) due to environmental factors. In the process of tunnels passing under existing buildings (structures), soil excavation inevitably affects the ground surface and existing buildings (structures). Since the settlement standards required for important existing buildings (structures) are relatively strict, any carelessness in tunnel construction will affect the normal use functions of important existing buildings (structures). Tunnel construction faces many challenges and limitations. Therefore, it is necessary to make accurate predictions during the construction process when the tunnel face is not close to the building (structure), conduct sufficient risk identification and evaluation of the predictions, and timely issue early warning information and reinforcement measures plans to ensure the safety of important existing buildings (structures) and tunnel construction.

[0003] While existing simulation models for mountain tunnels underpassing existing structures reflect the tunnel excavation status to some extent, they are static and lack real-time performance. Regarding risk assessment, due to the complexity of stratum lithology and groundwater development, as well as the impact of tunnel excavation parameters on existing structures, conventional risk assessment methods often rely on expert engineering experience, resulting in qualitative, empirical results and potentially inaccurate environmental risk assessments. In the absence of extensive historical data, ensuring accurate prediction and assessment of risk levels remains a major challenge. Summary of the Invention

[0004] In view of this, the present invention provides a safety warning method for construction under existing important buildings based on digital twins to solve the above problems.

[0005] The present invention provides a safety early warning method for construction under existing important buildings based on digital twins, including: collecting historical engineering case data and constructing an engineering case database, wherein the historical engineering case data includes environmental data, construction parameters and construction monitoring data; predicting the environmental data and construction parameters in the engineering case database through a risk prediction model to obtain prediction results; determining the relationship between risk levels based on the environmental data and construction monitoring data in the engineering case database, and constructing a risk assessment system based on a neural network; inputting the prediction results into the risk assessment system for safety early warning; and issuing early warning information if the conditions corresponding to the risk level are triggered.

[0006] In another implementation of the present invention, the environmental data includes the spatial position relationship and geometric information of mountains, buildings and tunnels; the construction parameters include various parameters proposed in the engineering design drawings; and the construction monitoring data includes horizontal convergence, vault deformation and surface settlement data of mountain tunnels.

[0007] In another implementation of the present invention, the environmental data and construction parameters in the engineering case database are predicted by a risk prediction model to obtain a prediction result, including: using a risk prediction model based on a finite element numerical simulation method to model a complex virtual scene for the environmental information and construction parameters in the engineering case database; performing finite element analysis on the construction process in the complex virtual scene where the face is not close to the bottom of the building to obtain a prediction result.

[0008] In another implementation of the present invention, the risk prediction model structure includes an input layer, a hidden layer 1, a hidden layer 2, and an output layer; the hidden layer 1 contains H1 neurons, and a ReLU activation function is used to introduce nonlinearity. The ReLU function formula is:

[0009] a1=ReLU(W1×X+b1)=max(θ, W1×X+b1)

[0010] Where W1 is the weight matrix from the input layer to hidden layer 1, b1 is the bias vector of hidden layer 1, X is the input vector, and a1 is the output vector of hidden layer 1.

[0011] The hidden layer 2 contains H2 neurons and uses the ReLU series activation function:

[0012] a2=ReLU(W2×a1+b2)

[0013] Where W2 is the weight matrix from hidden layer 1 to hidden layer 2, b2 is the bias vector of hidden layer 2, and a2 is the output vector of hidden layer 2.

[0014] The output layer contains C nodes, where C is the number of risk level categories. The Softmax activation function is used to convert the output into a distribution representing the probability of each risk level. The Softmax function formula is:

[0015] y pred =Softmax(W3×a2+b3)

[0016]

[0017] Among them, W3 is the weight matrix from hidden layer 2 to output layer, b3 is the bias vector of output layer, z = W3 × a2 + b3 is the original score of output layer, y pred is the predicted probability vector, y pred-jIt represents the predicted probability that the sample belongs to the jth class.

[0018] In another implementation of the present invention, the loss function of the risk prediction model is expressed as:

[0019] J(θ)=-(1 / N)×∑[∑(y true-i ×log(y pred-j ))]

[0020] Where N is the total number of training samples, y true-i is the one-hot vector of the true label of the sample, then y true =[0,1,0]), y pred-i is the probability of the corresponding category predicted by the model.

[0021] In another implementation of the present invention, the relationship between risk levels is determined based on the environmental data and construction monitoring data in the engineering case database, and a risk assessment system based on a neural network is constructed, including: determining the relationship between risk levels based on the environmental data and construction monitoring data in the engineering case database; determining the importance of each evaluation factor based on the relationship between the risk levels, and establishing a scale table for comparing the importance; constructing a judgment matrix based on the scale table, and constructing a risk assessment system based on a neural network.

[0022] In another implementation of the present invention, constructing a judgment matrix according to the scaling table and constructing a risk assessment system based on a neural network includes: calculating the maximum eigenvalue of the judgment matrix, and performing a consistency check on the judgment matrix based on the maximum eigenvalue; if the consistency check fails, adjusting the model parameters until the consistency check passes, thereby constructing a risk assessment system based on a neural network model.

[0023] On the other hand, the present invention provides a safety warning system for construction under existing important buildings based on digital twins, including: a physical layer, a logical layer, a functional layer, and an application layer; the physical layer is used to collect historical engineering case data and construct an engineering case database, and the historical engineering case data includes environmental data, construction parameters and construction monitoring data; the logical layer is used to predict the environmental data and construction parameters in the engineering case database through a risk prediction model to obtain prediction results; the functional layer is used to determine the relationship between risk levels based on the environmental data and construction monitoring data in the engineering case database, and construct a risk assessment system based on a neural network; the prediction results are input into the risk assessment system for safety warning; the application layer is used to issue warning information if the conditions corresponding to the risk level are triggered.

[0024] Another aspect of the present invention provides an electronic device, comprising: a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the steps of a safety warning method for construction under an existing important building based on digital twins as described in any one of the above items are implemented.

[0025] Another aspect of the present invention provides a computer storage medium, characterized in that a computer program is stored on the computer storage medium, and when the computer program is executed by a processor, the steps of the safety warning method for construction under existing important buildings based on digital twins as described in any of the above items are implemented.

[0026] The digital twin-based safety warning method for construction under existing important buildings of the present invention adopts digital twin technology and neural network model and other technologies to realize dynamic safety warning of mountain tunnel construction under existing important buildings; the system can intuitively display the digital three-dimensional model of mountain tunnel construction under existing important buildings corresponding to the physical world, and combine the monitoring data to deduce the deformation state of the mountain tunnel, the surface settlement state and the deformation state of existing important buildings in real time, with interoperability, scalability, real-time, fidelity and closed-loop properties; vividly simulate and predict the various risks brought about by the construction process of the face close to the bottom of the building, and provide a reference for subsequent management personnel's decision-making; the constructed warning system is conducive to management personnel to identify the severity of the warning situation of mountain tunnel construction under important buildings (structures) in advance, so as to efficiently respond to measures and ensure the safe progress of construction. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the embodiments or the description of the prior art. By reading the detailed description of the embodiments below, the advantages and benefits of the solutions will become clear to those skilled in the art. The drawings are only for the purpose of illustrating preferred embodiments and are not to be considered as limiting the present invention. In the drawings:

[0028] Figure 1 This is a flow chart of a safety warning method for construction under existing important buildings based on digital twins according to an embodiment of the present invention.

[0029] Figure 2 The figure is a plan view of the measurement point arrangement according to an embodiment of the present invention.

[0030] Figure 3 The diagram is a schematic diagram of measuring point arrangement taking a cross section of a road surface as an example according to an embodiment of the present invention.

[0031] Figure 4This is a schematic diagram of measuring point arrangement taking the cross section of a ground slope platform as an example according to an embodiment of the present invention.

[0032] Figure 5 This is a schematic diagram of risk assessment indicators for mountain tunnel construction underpasses of existing important buildings (structures) according to one embodiment of the present invention.

[0033] Figure 6 This is a cloud map of surface settlement of a tunnel under a highway according to one embodiment of the present invention.

[0034] Figure 7 This is a cloud diagram of the vertical deformation of soil in a tunnel under a highway according to an embodiment of the present invention.

[0035] Figure 8 This is a vertical displacement cloud map of a highway tunnel according to an embodiment of the present invention.

[0036] Figure 9 This is a cloud map of horizontal displacement of a highway tunnel according to an embodiment of the present invention.

[0037] Figure 10 This is a schematic diagram of the framework of a safety warning system for construction under existing important buildings based on digital twins, according to an embodiment of the present invention. DETAILED DESCRIPTION

[0038] In order to enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention, the technical solutions in the embodiments of the present invention will be clearly and detailedly described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by those skilled in the art based on the embodiments in the embodiments of the present invention should fall within the scope of protection of the embodiments of the present invention.

[0039] Figure 1 A flowchart of a safety warning method for construction under existing important buildings based on digital twins is provided in an embodiment of the present invention, as shown in FIG. Figure 1 As shown, this embodiment mainly includes:

[0040] S101. Collect historical engineering case data and build an engineering case database. The historical engineering case data includes environmental data, construction parameters, and construction monitoring data.

[0041] S102: Predict the environmental data and construction parameters in the engineering case database using a risk prediction model to obtain a prediction result.

[0042] S103: Determine the relationship between risk levels based on the environmental data and construction monitoring data in the engineering case database, and construct a risk assessment system based on a neural network.

[0043] S104: Input the prediction result into the risk assessment system for safety warning.

[0044] S105. If the conditions corresponding to the risk level are triggered, a warning message is issued.

[0045] The digital twin-based safety warning method for construction under existing important buildings of the present invention adopts digital twin technology and neural network model and other technologies to realize dynamic safety warning of mountain tunnel construction under existing important buildings; the system can intuitively display the digital three-dimensional model of mountain tunnel construction under existing important buildings corresponding to the physical world, and combine the monitoring data to deduce the deformation state of the mountain tunnel, the surface settlement state and the deformation state of existing important buildings in real time, with interoperability, scalability, real-time, fidelity and closed-loop properties; vividly simulate and predict the various risks brought about by the construction process of the face close to the bottom of the building, and provide a reference for subsequent management personnel's decision-making; the constructed warning system is conducive to management personnel to identify the severity of the warning situation of mountain tunnel construction under important buildings (structures) in advance, so as to efficiently respond to measures and ensure the safe progress of construction.

[0046] In another implementation of the present invention, the environmental data includes the spatial position relationship and geometric information of mountains, buildings and tunnels; the construction parameters include various parameters proposed in the engineering design drawings; and the construction monitoring data includes horizontal convergence, vault deformation and surface settlement data of mountain tunnels.

[0047] For example, spatial sensing equipment is used to focus on collecting the geometric information of the mountain and buildings along the proposed mountain tunnel. Based on the engineering design drawings, the spatial positions of the proposed mountain tunnel and buildings and the construction parameters of the tunnel excavation are determined.

[0048] During the construction of mountain tunnels, arch deformation observation points and horizontal convergence observation points are set up, and monitoring devices such as total stations are used to observe the arch deformation and horizontal convergence data.

[0049] According to the Technical Specifications for Monitoring and Measurement of Railway Tunnels (Q / CR 9218-2015), if the tunnel is shallowly buried and passes under buildings, surface settlement observation points should be set up before tunnel excavation, and the surface settlement points should be arranged at the same cross-sectional mileage as the measurement points inside the tunnel. Figure 2 、 3 As shown in Figures 4 and 5, when the construction face is close to the building (structure), a total station is used to monitor its settlement.

[0050] Specifically, data preparation and division,collecting a database of historical engineering cases, each case contains:

[0051] Input features (X):

[0052] X1: Geological conditions. This can be expressed using numerical codes, for example: 1 = stable rock mass, 2 = broken rock mass, 3 = soft soil, 4 = sand and gravel, etc., or using more refined rock mass quality indicators (such as RMR, Q value).

[0053] X2: Groundwater situation. It can be expressed by numerical coding, for example: 0 = no groundwater, 1 = a small amount of seepage, 2 = moderate water inflow, 3 = a large amount of water inflow, or using water inflow volume (m 3 / day) and other quantitative indicators.

[0054] X3: Tunnel cross-sectional area (m 2 ).

[0055] X4: Tunnel burial depth (m).

[0056] Output label (Y): Risk level, determined based on a comprehensive evaluation of the three monitoring indicators of vault deformation, horizontal convergence, and surface settlement (e.g., through expert scoring, multi-indicator fusion algorithms, or standardized thresholds). For example, it can be defined as: 1 = low risk (Level I), 2 = medium risk (Level II), and 3 = high risk (Level III).

[0057] Standardize or normalize the input features X1, X2, X3, and X4 to eliminate the dimensionality effect and accelerate model convergence. Example of standardization formula:

[0058] X n =(X-μ) / σ

[0059] Where: μ is the mean of the feature, σ is the standard deviation of the feature.

[0060] The processed dataset is randomly shuffled and then split into a training set and an independent test set in a ratio of approximately 8:2. The training set is used for model learning and parameter updates, while the test set is only used to evaluate the generalization performance of the model and does not participate in any training process.

[0061] In another implementation of the present invention, the environmental data and construction parameters in the engineering case database are predicted by a risk prediction model to obtain a prediction result, including: using a risk prediction model based on a finite element numerical simulation method to model a complex virtual scene for the environmental information and construction parameters in the engineering case database; performing finite element analysis on the construction process in the complex virtual scene where the face is not close to the bottom of the building to obtain a prediction result.

[0062] For example, the relevant data of historical engineering cases are obtained and applied to the neural network model, combining the four environmental information of different geological conditions, groundwater conditions, tunnel cross-sectional area, and tunnel burial depth, engineering information, and the relationship between the risk levels determined according to the three monitoring indicators of arch deformation, horizontal convergence, and surface settlement. The data in the engineering case database are divided into a training set and a test set. The training set data is used to train the neural network model, and the test set data is used to monitor the accuracy of the neural network model. If the training accuracy is poor, the neural network model parameters are optimized and retrained, and the test is repeated in this way.

[0063] In another implementation of the present invention, a multilayer feedforward neural network (Multilayer Perceptron, MLP) is constructed as a risk prediction model.

[0064] Example model structure:

[0065] Input layer: 4 nodes, corresponding to the four features X1, X2, X3, and X4.

[0066] Hidden layer 1: Contains H1 neurons (e.g. H1 = 64). Uses ReLU activation function to introduce nonlinearity. ReLU function formula:

[0067] a1=ReLU(W1×X+b1)=max(θ, W1×X+b1)

[0068] Where: W1 is the weight matrix from the input layer to hidden layer 1 (dimension [4×H1]), b1 is the bias vector of hidden layer 1 (dimension [H1×1]), X is the input vector (dimension [4×1]), and a1 is the output vector of hidden layer 1 (dimension [H1×1]).

[0069] Hidden layer 2: Contains H2 neurons (e.g., H2 = 64). Also uses the ReLU activation function.

[0070] a2=ReLU(W2×a1+b2)

[0071] Where: W2 is the weight matrix from hidden layer 1 to hidden layer 2 (dimension [H1×H2]), b2 is the bias vector of hidden layer 2 (dimension [H2×1]), and a2 is the output vector of hidden layer 2 (dimension [H2×1]).

[0072] Output layer: Contains C nodes (C = the number of risk level categories, for example, C = 3). Use the Softmax activation function to convert the output into a distribution representing the probability of each risk level. Softmax function formula:

[0073] y pred =Softmax(W3×a2+b3)

[0074]

[0075] Where: W3 is the weight matrix from hidden layer 2 to the output layer (dimension [H2×C]), b3 is the bias vector of the output layer (dimension [C×1]), z=W3×a2+b3 is the raw score of the output layer, y pred is the predicted probability vector (dimension [C×1]), y pred-j It represents the predicted probability that the sample belongs to the jth class.

[0076] Model parameters: The set of all learnable parameters of the model is denoted as θ = {W1, b1, W2, b2, W3, b3}.

[0077] In another implementation of the present invention, the loss function of the model is defined as follows: the classification cross entropy loss is used. This loss measures the predicted probability distribution y pred One-hot encoding y of the true label true The difference between. Loss function formula:

[0078] J(θ)=-(1 / N)×∑[∑(y true-i ×log(y pred-j ))]

[0079] Where: N is the total number of training samples, the outer layer summation traverses all samples, and the inner layer summation traverses all categories (C). true-i is the one-hot vector of the true label of the sample (for example, if the true label is "medium risk (level II)", then y true =[0,1,0]), y pred-i is the probability of the corresponding category predicted by the model.

[0080] Select the optimization algorithm: Use the adaptive moment estimation optimizer to update the model parameters θ. Basic update steps:

[0081] (1) Calculate the gradient g = ΔθJ(θ) of the loss J(θ) on the current mini-batch (m samples) with respect to the parameter θ.

[0082] (2) Update the first-order moment estimate m t =β1×m t-1 +(1-β1)×g

[0083] (3) Update the second-order moment estimate v t =β2×v t-1 +(1-β2)×g 2

[0084] (4) Correction of first-order moment deviation

[0085] (5) Correction of second-order moment deviation

[0086] (6) Update parameters

[0087] Where: t is the number of iterations, α is the initial learning rate, β1, β2 are the decay rates of moment estimates, ε is the numerical stability term, m t , v t Initialized to 0.

[0088] Training loop: Set the number of training rounds, such as 100 or 200. Set the batch size, such as 32 or 64. For each round of training, divide the training into multiple small batches, perform forward propagation, and calculate the input X batch Get the prediction y through the model pred-batch .

[0089] Calculate the loss J(θ): Perform backpropagation and calculate the gradient g of the loss J(θ) with respect to all parameters θ.

[0090] The adaptive moment estimation optimizer is used to update the parameter θ according to the gradient g. After each training, the evaluation indicators (such as accuracy) are calculated on the validation set or the training set to monitor the training process and prevent overfitting.

[0091] Evaluate and monitor the model. After training, use an independent test set to evaluate model performance.

[0092] Key evaluation metrics: Accuracy: (TP+TN) / (TP+TN+FP+FN), where TP, TN, FP, and FN represent true positives, true negatives, false positives, and false negatives, respectively (macro-average or micro-average calculation is required for multiple classifications); Precision, Recall, and F1 value: calculated for each risk level, followed by macro-average or micro-average calculation; Confusion Matrix: intuitively displays the model's predictions for each category.

[0093] Determine whether the training accuracy is "poor": Set an acceptable performance threshold (for example, test set accuracy ≥ 85%, or F1 macro average ≥ 0.80). If the evaluation metric is lower than the threshold, the training accuracy is considered poor and the model needs to be optimized.

[0094] Optimize and retrain the model. If the model performance on the test set is not up to standard (poor training accuracy), try one or more of the following optimizations:

[0095] (1) Adjust the model structure: increase or decrease the number of hidden layers (for example, try 1 or 3 layers); increase or decrease the number of neurons in each hidden layer (for example, try H1=32, H2=32 or H1=128, H2=128); try different activation functions (for example, try tanh or LeakyReLU in the hidden layer).

[0096] (2) Adjust training parameters: change the learning rate α (for example, try 0.0005, 0.01); change the batch size; increase the number of training rounds.

[0097] Preferably, in order to improve the calculation efficiency of the model, reasonable assumptions are made in terms of boundary conditions, material parameters, contact settings, processes, load conditions, and mesh division during the modeling process, in an effort to restore the real state of the physical world.

[0098] Based on the risk prediction model, the construction process of the tunnel face close to the building (structure) is simulated to predict the deformation of the mountain tunnel vault, horizontal convergence, surface settlement and deformation of the building (structure). Taking a tunnel under a highway tunnel as an example, the surface settlement and vertical deformation of the tunnel are as follows: Figure 6 、 7 As shown in the figure, the vertical displacement cloud diagram and horizontal displacement of the high-speed tunnel are as follows: Figure 8 、 9 shown.

[0099] The prediction results include the crown deformation, horizontal convergence and surface settlement results of the mountain tunnel; the prediction results are checked with the on-site monitoring data to determine the accuracy of the risk prediction model.

[0100] It should be understood that the digital twin model is interoperable, scalable, real-time, fidelity and closed-loop. It has the ability to map the physical world into a digital model and reflect the state of the physical world. The digital twin is dynamic and can evolve in real time with the physical world.

[0101] In another implementation of the present invention, the relationship between risk levels is determined based on the environmental data and construction monitoring data in the engineering case database, and a risk assessment system based on a neural network is constructed, including: determining the relationship between risk levels based on the environmental data and construction monitoring data in the engineering case database; determining the importance of each evaluation factor based on the relationship between the risk levels, and establishing a scale table for comparing the importance; constructing a judgment matrix based on the scale table, and constructing a risk assessment system based on a neural network. For example, Figure 5As shown, in a risk assessment system, the key issue that the neural network model needs to address is determining the importance of each pair of evaluation factors based on the contribution of the model input parameters to the output parameters. A 1-9 scale is used to determine the relative importance of each pair of indicators. A weight scale of 1 indicates that two factors are equally important. Increasing weight scales indicate that the former is increasingly more important than the latter, until a weight scale of 9 indicates that the former is extremely more important than the latter. This establishes a scale for comparing importance. In another implementation of the present invention, constructing a judgment matrix based on the scale table and building a neural network-based risk assessment system includes: calculating the maximum eigenvalue of the judgment matrix, performing a consistency check on the judgment matrix based on the maximum eigenvalue; if the consistency check fails, adjusting model parameters until the consistency check passes, thereby constructing a risk assessment system based on the neural network model. For example, if the consistency check fails, adjusting model parameters based on the situation to address poor test data performance until the consistency check passes, thereby constructing a risk assessment system based on the neural network model. Another implementation of the present invention further includes: based on the prediction results and the warning information, applying different reinforcement measures in the digital twin model to obtain prediction results corresponding to the different measures, and issuing a construction plan. For example, the prediction results based on the risk prediction model are input into a risk assessment system for mountain tunnel construction underpasses of existing important buildings (structures), setting four warning levels: "red," "orange," "yellow," and "blue." The warning database and digital twin simulation model are systematically integrated, and warning information is issued when conditions corresponding to the risk level are triggered. Based on the prediction results of the risk prediction model, different reinforcement measures are applied in the digital twin model to obtain different digital twin model prediction results, including arch deformation, horizontal convergence, surface settlement and deformation state of buildings (structures), and a construction plan is issued. Commonly used auxiliary construction methods include shotcrete sealing of excavation face, advance anchor rods or advance small pipe support, advance small pipe peripheral grouting reinforcement of strata, long pipe shed advance support reinforcement of strata, horizontal rotary jet pile advance support and other pre-support measures.

[0102] In another implementation of the present invention, the method further includes: visually displaying all prediction results in the form of animation.

[0103] For example, the predictions of the digital twin model and the prediction results of different reinforcement measures are visualized in the form of animation, which facilitates managers to make decisions based on the prediction results, warning results, and plan results.

[0104] Another aspect of the present invention, as Figure 10 As shown, a safety warning system for construction under existing important buildings based on digital twins is provided, including:

[0105] Physical layer, logical layer, functional layer, application layer.

[0106] The physical layer is used to collect historical engineering case data and build an engineering case database. The historical engineering case data includes environmental data, construction parameters and construction monitoring data.

[0107] For example, the physical layer includes the environmental information and construction parameter collection modules required to build the digital twin model, as well as the project case database required to construct the risk assessment system. The physical layer is a digital representation of the physical world, capturing its key characteristics. Changes to these key characteristics are reflected in the environmental information and construction parameter collection modules within the physical layer in real time.

[0108] The logic layer is used to predict the environmental data and construction parameters in the engineering case database through a risk prediction model to obtain a prediction result.

[0109] Exemplarily, the logic layer includes a risk prediction model and a risk assessment system. The logic layer is the core part of the digital twin model that analyzes the physical world. It is a bridge connecting environmental information and risk levels, and can reflect the correspondence between environmental information and risks.

[0110] The functional layer is used to determine the relationship between risk levels based on the environmental data and construction monitoring data in the engineering case database, and to build a risk assessment system based on a neural network; the prediction results are input into the risk assessment system for safety warning.

[0111] Exemplarily, the functional layer includes a prediction module, an early warning module, and a plan release module. The prediction module is used to simulate the construction process of the face close to the building (structure) directly below the face according to the risk prediction model, and predict the vault deformation, horizontal convergence, surface settlement and deformation state of the mountain tunnel. The early warning module is used to input the prediction results of the risk prediction model into the risk assessment system for the construction of mountain tunnels under existing important buildings (structures), and issue early warning information if the conditions corresponding to the risk level are triggered. The plan release module applies different reinforcement measures in the digital twin model based on the prediction results of the risk prediction model to obtain different digital twin prediction results. The application layer is used to issue early warning information if the conditions corresponding to the risk level are triggered.

[0112] The application layer includes a visualization display module, which visualizes the prediction results of the digital twin model in the form of animation, making it easier for managers to make decisions based on the prediction results, warning results, and plan results.

[0113] The digital twin-based safety warning system for construction under existing important buildings of the present invention adopts digital twin technology and neural network model and other technologies to realize dynamic safety warning of mountain tunnel construction under existing important buildings; the system can intuitively display the digital three-dimensional model of mountain tunnel construction under existing important buildings corresponding to the physical world, and combine monitoring data to deduce the deformation state of mountain tunnel, surface settlement state and deformation state of existing important buildings in real time, with interoperability, scalability, real-time, fidelity and closed-loop properties; vividly simulate and predict various risks brought about by the construction process of the face close to the bottom of the building, and provide a reference for subsequent management personnel's decision-making; the constructed warning system is conducive to management personnel to identify the severity of the alarm situation of mountain tunnel construction under important buildings (structures) in advance, so as to efficiently respond to measures and ensure the safe progress of construction.

[0114] In another aspect of the present invention, an electronic device includes a processor, a memory, a communication bus, and a communication interface.

[0115] in:

[0116] The processor, memory and communication interface communicate with each other through a communication bus.

[0117] Communication interface, used to communicate with other electronic devices or servers.

[0118] The processor is used to execute the program, and specifically can execute the steps of any one of the above-mentioned embodiments of the safety warning method for construction underpasses of existing important buildings based on digital twins.

[0119] Specifically, the program may include program codes including computer operation instructions.

[0120] The processor may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application. The one or more processors included in the smart device may be processors of the same type, such as one or more CPUs; or different types of processors, such as one or more CPUs and one or more ASICs.

[0121] Memory is used to store programs. The memory may include high-speed RAM memory, and may also include non-volatile memory (non-volatile memory), such as at least one disk storage.

[0122] The program can be specifically used to enable the processor to execute the steps to implement any one of the safety warning methods for construction under existing important buildings based on digital twins described in the embodiments. The specific implementation of each step in the program can refer to the corresponding descriptions in the steps and units executed by any one of the safety warning methods for construction under existing important buildings based on digital twins in the above steps, which will not be repeated here. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working process of the equipment and modules described above can refer to the corresponding process description in the aforementioned method embodiment.

[0123] The exemplary embodiments of the present application further provide a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable a computer to execute the methods of the various embodiments of the present application.

[0124] The method according to the embodiment of the present invention described above can be implemented in hardware, firmware, or as software or computer code that can be stored in a recording medium (such as a CD ROM, RAM, floppy disk, hard disk or magneto-optical disk), or as computer code that is originally stored in a remote recording medium or a non-temporary machine-readable medium downloaded via a network and will be stored in a local recording medium, so that the method described herein can be stored in such software processing on a recording medium using a general-purpose computer, a dedicated processor or programmable or dedicated hardware (such as an ASIC or FPGA). It can be understood that a computer, a processor, a microprocessor controller or programmable hardware includes a storage component (e.g., RAM, ROM, flash memory, etc.) that can store or receive software or computer code, and when the software or computer code is accessed and executed by a computer, a processor or hardware, the method described herein is implemented. In addition, when a general-purpose computer accesses the code for implementing the method shown here, the execution of the code converts the general-purpose computer into a dedicated computer for executing the method shown here.

[0125] Thus far, specific embodiments of the present invention have been described. Other embodiments are within the scope of the appended claims. In some cases, the actions recited in the claims can be performed in a different order and still achieve the desired results. Additionally, the processes depicted in the accompanying drawings do not necessarily require the specific order shown, or sequential order, to achieve the desired results.

[0126] It should be noted that all directional indications in the embodiments of the present invention (such as up, down, left, right, back, etc.) are only used to explain the relative position relationship between the components in a certain specific order (as shown in the accompanying drawings). If the specific order changes, the directional indication will also change accordingly.

[0127] In the description of the present invention, the terms "first" and "second" are used solely to facilitate description of different components or names and should not be construed as indicating or implying a sequential relationship, relative importance, or implicitly specifying the quantity of the technical features being described. Therefore, features specified as "first" or "second" may explicitly or implicitly include at least one of such features.

[0128] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0129] It should be noted that although the specific embodiments of the present invention are described in detail in conjunction with the accompanying drawings, this should not be construed as limiting the scope of protection of the present invention. Within the scope described by the claims, various modifications and variations that can be made by those skilled in the art without creative effort still fall within the scope of protection of the present invention.

[0130] The examples of the embodiments of the present invention are intended to briefly illustrate the technical features of the embodiments of the present invention so that those skilled in the art can intuitively understand the technical features of the embodiments of the present invention, and are not intended to improperly limit the embodiments of the present invention.

[0131] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A safety warning method for construction under existing important buildings based on digital twins, characterized by: include: Collect historical engineering case data and build an engineering case database, wherein the historical engineering case data includes environmental data, construction parameters and construction monitoring data; Predicting environmental data and construction parameters in the engineering case database using a risk prediction model to obtain prediction results; Determine the relationship between risk levels based on environmental data and construction monitoring data in the engineering case database, and construct a risk assessment system based on a neural network; Inputting the prediction results into the risk assessment system for safety warning; If the conditions corresponding to the risk level are triggered, an early warning message will be issued.

2. The method according to claim 1, characterized in that The environmental data includes the spatial position relationship and geometric information of mountains, buildings and tunnels; The construction parameters include various parameters proposed in the engineering design drawings; The construction monitoring data includes mountain tunnel horizontal convergence, vault deformation and surface settlement data.

3. The method according to claim 1, characterized in that The risk prediction model is used to predict the environmental data and construction parameters in the engineering case database to obtain prediction results, including: Using a risk prediction model based on finite element numerical simulation method, modeling a complex virtual scene based on the environmental information and construction parameters in the engineering case database; Finite element analysis is performed on the construction process in the complex virtual scene where the tunnel face is not close to the bottom of the building to obtain prediction results, which include the crown deformation, horizontal convergence and surface settlement results of the mountain tunnel.

4. The method according to claim 3, characterized in that The risk prediction model structure includes an input layer, a hidden layer 1, a hidden layer 2 and an output layer; The hidden layer 1 contains H1 neurons and uses the ReLU activation function to introduce nonlinearity. The ReLU function formula is: a1=ReLU(W1×X+b1)=max(θ, W1×X+b1) Where W1 is the weight matrix from the input layer to the hidden layer 1, b1 is the bias vector of the hidden layer 1, X is the input vector, and a1 is the output vector of the hidden layer 1; The hidden layer 2 contains H2 neurons and uses the ReLU series activation function: a2=ReLU(W2×a1+b2) Where W2 is the weight matrix from hidden layer 1 to hidden layer 2, b2 is the bias vector of hidden layer 2, and a2 is the output vector of hidden layer 2; The output layer contains C nodes, where C is the number of risk level categories. The Softmax activation function is used to convert the output into a distribution representing the probability of each risk level. The Softmax function formula is: and pred =Softmax(W3×a2+b3) Among them, W3 is the weight matrix from hidden layer 2 to output layer, b3 is the bias vector of output layer, z = W3 × a2 + b3 is the original score of output layer, y pred is the predicted probability vector, y pred-j It represents the predicted probability that the sample belongs to the jth class.

5. The method according to claim 4, characterized in that The loss function of the risk prediction model is expressed as: J(θ)=-(1 / N)×∑[∑(y true-i ×log(y pred-j ))] Where N is the total number of training samples, y true-i is the one-hot vector of the true label of the sample, then y true =[0,1,0]), y pred-i is the probability of the corresponding category predicted by the model.

6. The method according to claim 1, characterized in that The method of determining the relationship between risk levels based on the environmental data and construction monitoring data in the engineering case database and constructing a risk assessment system based on a neural network includes: Determining relationships between risk levels based on environmental data and construction monitoring data in the engineering case database; Based on the relationship between the risk levels, determine the importance of each evaluation factor and establish a scale for comparing the importance; A judgment matrix is ​​constructed according to the scale table, and a risk assessment system based on a neural network is constructed.

7. The method according to claim 6, characterized in that The method of constructing a judgment matrix according to the scale table and constructing a risk assessment system based on a neural network includes: Calculating the maximum eigenvalue of the judgment matrix, and performing a consistency check on the judgment matrix based on the maximum eigenvalue; If the consistency test fails, the model parameters are adjusted until the consistency test passes, and a risk assessment system based on the neural network model is constructed.

8. A safety warning system for construction under existing important buildings based on digital twins, characterized by: include: Physical layer, logical layer, functional layer, application layer; The physical layer is used to collect historical engineering case data and build an engineering case database. The historical engineering case data includes environmental data, construction parameters and construction monitoring data. The logic layer is used to predict the environmental data and construction parameters in the engineering case database through a risk prediction model to obtain a prediction result; The functional layer is used to determine the relationship between risk levels based on the environmental data and construction monitoring data in the engineering case database, and to build a risk assessment system based on a neural network; Inputting the prediction results into the risk assessment system for safety warning; The application layer is used to issue early warning information if conditions corresponding to the risk level are triggered.

9. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of a safety warning method for construction under an existing important building based on digital twins are implemented as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that The computer storage medium stores a computer program, which, when executed by a processor, implements the steps of a safety early warning method for construction under an existing important building based on digital twins as described in any one of claims 1 to 7.

Images