Design method, equipment, medium and program product of prestressed concrete slab

Abstract

One or more embodiments of the present specification provide a method, apparatus, medium and program product for designing a prestressed concrete slab. The method comprises the steps that a cost optimization function of the prestressed concrete slab is obtained, the cost optimization function comprises a cost item and a load constraint item, the cost item takes a design parameter of the prestressed concrete slab as an input variable, and the load constraint item takes a load response of the prestressed concrete slab as an input variable; iteratively executing the following steps until a stop condition is met: determining a current design parameter used in the current iteration; predicting a current load response corresponding to the current design parameter based on a load prediction model; the current design parameters and the current load response are adopted to carry out current iteration on the cost optimization function to optimize the current design parameters, and optimized optimization design parameters are obtained; updating the current design parameter to the optimization design parameter, and executing the next round of iteration; and determining the latest optimization design parameters obtained after iteration is stopped as the design result of the prestressed concrete slab.

Description

Technical Field

[0001] One or more embodiments of this specification relate to the technical field of building structure design, and in particular, to a design method, device, medium, and program product for prestressed concrete slabs. Background Art

[0002] Due to its good load-bearing capacity and durability, concrete slabs are widely used in the design of roof slabs and floor slabs of industrial plants, warehouses, parking lots, etc. The concrete roof system plays multiple roles such as load-bearing, protection, and space shaping in the building structure, and its design needs to comprehensively consider various factors such as structural safety, durability, and economy.

[0003] Traditional concrete design methods mainly rely on manual experience for structural design, making it difficult to select the optimal parameters, resulting in poor performance in terms of economy and material utilization of the design results.

[0004] In recent years, the finite element analysis method has been introduced into the design of concrete slabs. The finite element analysis method can more accurately analyze the mechanical properties and deformation behavior of the structure through numerical simulation, thus providing a more scientific basis for the design. However, in the design of concrete slabs, there are a large number of parameter inputs and complex calculation processes, resulting in a large time overhead for design and low calculation efficiency, making it difficult to meet the rapid design requirements in actual projects. Summary of the Invention

[0005] In view of this, one or more embodiments of this specification provide the following technical solutions:

[0006] According to the first aspect of one or more embodiments of this specification, a design method for prestressed concrete slabs is proposed, and the method includes:

[0007] Obtain the cost optimization function of the prestressed concrete slab, where the cost optimization function includes a cost item and a load constraint item, the cost item takes the design parameters of the prestressed concrete slab as input variables, and the load constraint item takes the load response of the prestressed concrete slab as input variables;

[0008] Iteratively execute the following steps until the stop condition is met:

[0009] Determine the current design parameters used in this round of iteration;

[0010] Predict the current load response corresponding to the current design parameters based on the load prediction model;

[0011] Use the current design parameters and the current load response to perform this round of iteration on the cost optimization function to optimize the current design parameters and obtain the optimized design parameters;

[0012] Update the current design parameters to the optimized design parameters and perform the next iteration;

[0013] Determine the latest optimized design parameters obtained after the iteration stops as the design result of the prestressed concrete slab.

[0014] Optionally, the construction process of the cost optimization function includes:

[0015] Construct a cost calculation function for the prestressed concrete slab;

[0016] Construct a constrained optimization problem with the goal of minimizing the cost calculation function, and the constraint conditions of the constrained optimization problem include that the load response of the prestressed concrete slab does not exceed a threshold;

[0017] Construct the cost optimization function based on the constrained optimization problem, where the cost term in the cost optimization function corresponds to the cost calculation function, and the constraint term corresponds to the constraint condition.

[0018] Optionally, the training process of the load prediction model includes:

[0019] Construct sample data for training the load prediction model, where the sample data uses the design parameters of the prestressed concrete slab as features and the load response corresponding to the design parameters as labels;

[0020] Train the initial load prediction model using the sample data to obtain a trained load prediction model.

[0021] Optionally, the construction of the sample data for training the load prediction model includes:

[0022] Obtain several groups of design parameters of the prestressed concrete slab;

[0023] Use the finite element analysis method to determine the load response corresponding to each group of design parameters respectively;

[0024] Normalize the design parameters and the load response respectively, and construct the sample data based on the normalized design parameters and their corresponding load responses.

[0025] Optionally, the stop condition includes:

[0026] Compared with the current design parameters used in the first iteration, the cost of the prestressed concrete slab corresponding to the optimized design parameters has been reduced by a preset proportion.

[0027] Optionally, the process of determining the latest optimized design parameters obtained after the iteration stops as the design result of the prestressed concrete slab includes:

[0028] Verify the latest optimized design parameters;

[0029] When the latest optimized design parameters pass the verification, determine the latest optimized design parameters as the design result of the prestressed concrete slab.

[0030] According to the second aspect of one or more embodiments of this specification, a design device for a prestressed concrete slab is proposed. The device includes:

[0031] A function acquisition unit that acquires a cost optimization function for the prestressed concrete slab. The cost optimization function includes a cost item and a load constraint item. The cost item takes the design parameters of the prestressed concrete slab as input variables, and the load constraint item takes the load response of the prestressed concrete slab as input variables;

[0032] An iterative optimization unit that iteratively executes the following steps until a stop condition is met:

[0033] Determine the current design parameters used in this round of iteration;

[0034] Predict the current load response corresponding to the current design parameters based on a load prediction model;

[0035] Use the current design parameters and the current load response to perform this round of iteration on the cost optimization function to optimize the current design parameters and obtain optimized optimized design parameters;

[0036] Update the current design parameters to the optimized design parameters and perform the next round of iteration;

[0037] A result determination unit that determines the latest optimized design parameters obtained after the iteration stops as the design result of the prestressed concrete slab.

[0038] According to the third aspect of one or more embodiments of this specification, an electronic device is proposed, including: a processor; a memory for storing processor-executable instructions; wherein, the processor runs the executable instructions to implement the steps of the foregoing method.

[0039] According to the fourth aspect of one or more embodiments of this specification, a computer-readable storage medium is proposed, on which computer instructions are stored, and when the instructions are executed by a processor, the steps of the foregoing method are implemented.

[0040] According to the fifth aspect of one or more embodiments of this specification, a computer program product is proposed, including a computer program / instructions, and when the computer program / instructions are executed by a processor, the steps of the foregoing method are implemented.

[0041] As can be seen from the above embodiments, by adopting the above design solution provided in this specification, a cost optimization function including a cost item and a load constraint item can be obtained, and the current load response corresponding to the current design parameters can be predicted. Then, the cost optimization function is iteratively solved using the current design parameters and the current load response to optimize the current design parameters to obtain optimized design parameters, and the latest optimized design parameters obtained after the iteration stops can be determined as the design result of the prestressed concrete slab. By adopting such a design solution for the prestressed concrete slab, the intelligent optimization of the design parameters can be automatically carried out, which can not only improve the design efficiency of the prestressed concrete slab, but also improve the utilization rate of materials and reduce the material cost through the iterative solution of the cost optimization function, thereby improving the economy of the structure. Description of the Drawings

[0042] Figure 1 FIG. is a schematic structural diagram of a design system for a prestressed concrete slab provided by an exemplary embodiment.

[0043] Figure 2 FIG. is a flowchart of a design method for a prestressed concrete slab provided by an exemplary embodiment.

[0044] Figure 3 FIG. is a flowchart of a training method for a load prediction model provided by an exemplary embodiment.

[0045] Figure 4 FIG. is a flowchart of a method for constructing sample data provided by an exemplary embodiment.

[0046] Figure 5 FIG. is a flowchart of a method for constructing a cost optimization function provided by an exemplary embodiment.

[0047] Figure 6 FIG. is a schematic diagram of an iterative solution process of a cost optimization function provided by an exemplary embodiment.

[0048] Figure 7 FIG. is a schematic diagram of another iterative solution process of a cost optimization function provided by an exemplary embodiment.

[0049] Figure 8 FIG. is a schematic structural diagram of a device provided by an exemplary embodiment.

[0050] Figure 9 FIG. is a block diagram of a design device for a prestressed concrete slab provided by an exemplary embodiment. Detailed Embodiments

[0051] Due to its good load-bearing capacity and durability, concrete slabs are widely used in the design of roof slabs and floor slabs of industrial plants, warehouses, parking lots and other buildings. The concrete roof system plays multiple roles in building structures, such as load-bearing, protection and space shaping. Its design needs to comprehensively consider various factors such as structural safety, durability and economy.

[0052] Traditional concrete design methods mainly rely on manual experience for structural design, making it difficult to select the optimal parameters, resulting in poor performance in terms of economy and material utilization of the design results.

[0053] In recent years, the finite element analysis method has been introduced into the design of concrete slabs. Through numerical simulation, the finite element analysis method can more accurately analyze the mechanical properties and deformation behavior of the structure, thus providing a more scientific basis for the design. However, in the design of concrete slabs, there are a large number of parameter inputs and complex calculation processes, resulting in a large time overhead for design and low calculation efficiency, making it difficult to meet the rapid design requirements in actual projects.

[0054] This specification provides a design scheme for prestressed concrete slabs, which can determine the design parameters of prestressed concrete slabs based on deep learning technology, not only improving the design efficiency, but also enhancing the material utilization rate, reducing the material cost at the same time, and improving the economy of the structure.

[0055] The prestressed concrete slabs described in this specification can be prestressed slabs of ordinary concrete, or prestressed UHPC slabs using UHPC (Ultra-High Performance Concrete) instead of ordinary concrete. The prestressed concrete slabs can be of double-T structure or C-shaped structure. For prestressed concrete slabs of different structures, although the design parameters may not be exactly the same, the design scheme provided in this specification can be used for design.

[0056] The design solution of the prestressed concrete slab provided in this specification can be applied to electronic devices, such as: PC (Personal Computer), mobile phones, tablet devices, PDAs (Personal Digital Assistants), wearable devices (such as smart glasses, smart watches, etc.). During operation, the electronic device can run the design client program of the prestressed concrete slab to implement the relevant functions of this application. For example, when the electronic device runs the design program of the prestressed concrete slab, it can implement the client for the design service of the prestressed concrete slab. Among them, the application program of the client for the design service of the prestressed concrete slab can be started and run on the electronic device. The client program can be a native application installed on the electronic device, or the program on the client side can be a mini-program, a quick application, or other similar forms. Of course, when using web technologies such as HTML5 or similar, the relevant functions can be implemented through the page displayed by the browser. Here, the browser can be an independent browser application or a browser module embedded in some applications.

[0057] The design solution of the prestressed concrete slab provided in this specification can also be implemented in cooperation with an electronic device and a server. Please refer to Figure 1 the schematic diagram of the architecture of the design system of the prestressed concrete slab shown. The system can include a server 11, a network 12, and several electronic devices, such as a PC (Personal Computer) 13, a mobile phone 14, etc. Using Figure 1 this front-end and back-end separation architecture, the design program of the prestressed concrete slab on the client side in the electronic device can be responsible for user interaction. For example, it can receive the initial design parameters, the maximum allowable stress of the design, the deflection, etc. uploaded by the user. The design program of the prestressed concrete slab on the back end in the server 11 is responsible for determining the design result and can return the design result to the electronic device for presentation, etc. Among them, the server 11 can be a physical server including an independent host, or the server 11 can be a virtual server hosted by a host cluster. This specification does not make special restrictions on this.

[0058] Figure 2 is a flowchart of a design method of a prestressed concrete slab provided by an exemplary embodiment.

[0059] Please refer to Figure 2 , the design method of the prestressed concrete slab can include the following steps:

[0060] Step 202: Obtain the cost optimization function of the prestressed concrete slab. The cost optimization function includes a cost item and a load constraint item. The cost item takes the design parameters of the prestressed concrete slab as input variables, and the load constraint item takes the load response of the prestressed concrete slab as input variables.

[0061] In this embodiment, the design parameters of the prestressed concrete slab may include material design parameters, slab design parameters, and prestress design parameters. Taking the prestressed UHPC slab with a double-T structure as an example, the material design parameters may include: UHPC compressive strength, tensile strength, ultimate tensile strain, etc.; the slab design parameters may include: top slab thickness, web thickness, web height, longitudinal and transverse rib spacing, longitudinal and transverse rib dimensions, etc.; the prestress design parameters may include: prestressed tendon position, prestressed steel strand diameter, total cross-sectional area of prestressed steel strand tension, etc.

[0062] In this embodiment, the load response of the prestressed concrete slab may include the maximum stress (which may include the maximum tensile stress and the maximum compressive stress), the maximum deflection, etc. The load response corresponds to the design parameters, and different design parameters usually correspond to different load responses.

[0063] In this embodiment, the cost optimization function may be an unconstrained optimization function, such as the Lagrangian function, etc. The cost optimization function includes a cost item and a load constraint item. Among them, the cost item takes the design parameters of the prestressed concrete slab as input variables, and can represent the cost of manufacturing the prestressed concrete slab with the corresponding design parameters, and can be the sum of the cost of the concrete slab and the prestress cost. The constraint item takes the load response of the prestressed concrete slab as input variables, and can represent the constraint that the load response of the prestressed concrete slab does not exceed the load response allowed by the design.

[0064] Step 204: Iteratively execute the following steps until the stop condition is met: Determine the current design parameters used in this round of iteration; Predict the current load response corresponding to the current design parameters based on the load prediction model; Use the current design parameters and the current load response to perform this round of iteration on the cost optimization function to optimize the current design parameters, and obtain the optimized design parameters after optimization; Update the current design parameters to the optimized design parameters, and execute the next round of iteration.

[0065] In this embodiment, the cost optimization function can be iteratively solved to optimize the design parameters, which can ensure that the corresponding load response does not exceed the load response allowed by the design while reducing the cost of the prestressed concrete slab.

[0066] In this embodiment, a set of initial design parameters can be preset in advance, and a load prediction model can be used to predict the load response corresponding to this set of initial design parameters as the initial load response. Then, this set of initial design parameters and the corresponding initial load response are used to iteratively solve the cost optimization function to update and optimize the initial design parameters, obtaining optimized design parameters. If the iteration stop condition corresponding to the cost optimization function is not satisfied, the cost optimization function can be continuously iteratively solved based on the optimized design parameters.

[0067] Among them, the load prediction model can be a trained deep learning model that can be used to predict the load response of a prestressed concrete slab, or a finite element analysis model. The finite element analysis model can use the finite element analysis (Finite Element Analysis, FEA) method to determine the load response corresponding to the design parameters. Of course, in other examples, the load prediction model can also be other models, and this specification does not make special restrictions on this.

[0068] In this embodiment, the stop condition can be that compared with the current design parameters (i.e., the initial design parameters) used in the first round of iteration, the cost of the prestressed concrete slab corresponding to the optimized design parameters has been reduced by a preset proportion. That is, after each round of iteration to update the current design parameters to obtain optimized design parameters, the cost of the optimized prestressed concrete slab corresponding to the optimized design parameters can be calculated, and then the cost of the optimized prestressed concrete slab is compared with the cost of the initial prestressed concrete slab corresponding to the initial optimized design parameters to determine whether the cost of the optimized prestressed concrete slab has been reduced by a preset proportion (for example: 20%, 30%, etc.) compared with the cost of the initial prestressed concrete slab to determine whether the stop condition is satisfied.

[0069] In this embodiment, the stop condition can also be that the cost optimization function has reached a preset number of iterative solutions, such as 1000 times, 2000 times, etc.

[0070] In this embodiment, the stop condition can also be that compared with the initial design parameters, the cost of the prestressed concrete slab corresponding to the optimized design parameters has been reduced by a preset proportion, and the cost optimization function has reached a preset number of iterative solutions, and the first one to reach is used as the stop condition. This specification does not make special restrictions on this.

[0071] Step 206: Determine the latest optimized design parameters obtained after the iteration stop as the design result of the prestressed concrete slab.

[0072] Based on the iterative solution in the foregoing step 204, the optimized design parameters (referred to as the latest optimized design parameters) obtained after the iteration stops can be determined as the target design parameters of the prestressed concrete slab, that is, the design result. Using such a design result to manufacture the corresponding prestressed concrete slab can greatly improve the material utilization rate, reduce the material cost, and improve the economy of the structure.

[0073] As can be seen from the above embodiments, by adopting the above design scheme provided in this specification, a cost optimization function including a cost item and a load constraint item can be obtained, and the current load response corresponding to the current design parameters can be predicted. Then, the current design parameters and the current load response are used to perform iterative solution on the cost optimization function to optimize the current design parameters to obtain optimized design parameters, and the latest optimized design parameters obtained after the iteration stops can be determined as the design result of the prestressed concrete slab. Using such a prestressed concrete slab design scheme, the intelligent optimization of design parameters can be automatically performed, which can not only improve the design efficiency of the prestressed concrete slab, but also improve the material utilization rate and reduce the material cost through the iterative solution of the cost optimization function, thereby improving the economy of the structure.

[0074] The following will introduce the specific implementation of this specification in detail from two aspects: the load prediction model and the cost optimization function.

[0075] I. Load Prediction Model

[0076] In this embodiment, taking the load prediction model as a deep learning model as an example, a multi-layer perceptron (MLP) neural network can be selected as the load prediction model, and then sample data can be used to train, validate, and test the initial load prediction model.

[0077] Please refer to Figure 3 , the training process of the load prediction model may include the following steps:

[0078] Step 302, construct sample data for training the load prediction model, where the sample data uses the design parameters of the prestressed concrete slab as features and the load response corresponding to the design parameters as labels.

[0079] In this embodiment, sample data for training the load prediction model can be constructed first. Please refer to Figure 4 , and the following method can be used to construct the sample data.

[0080] Step 3022, obtain several groups of design parameters of the prestressed concrete slab.

[0081] In this embodiment, multiple sets of design parameters for prestressed concrete slabs can be pre-constructed. Taking the prestressed concrete slab with a double-T structure as an example, the design parameters may include: top plate thickness, web thickness, web height, top plate span, compressive strength of the concrete slab, position of prestressed steel strands, number of prestressed steel strands, diameter of prestressed steel strands, total cross-sectional area of prestressed steel strand tension, etc.

[0082] Step 3024: Use the finite element analysis method to determine the load responses corresponding to each set of design parameters respectively.

[0083] In this embodiment, the finite element analysis method can be used to determine the load responses corresponding to each set of design parameters. For example, for each set of design parameters including top plate thickness, web thickness, web height, top plate span, compressive strength of the concrete slab, position of prestressed steel strands, number of prestressed steel strands, diameter of prestressed steel strands, and total cross-sectional area of prestressed steel strand tension, the finite element analysis method can be used to determine the load responses of the prestressed concrete slab corresponding to these design parameters. The load responses may include maximum stress, maximum deflection, etc.

[0084] For example, assuming that 200 sets of design parameters are pre-constructed, the finite element analysis method can be used to determine 200 sets of load responses corresponding to these 200 sets of design parameters.

[0085] Step 3026: Normalize the design parameters and the load responses respectively, and construct the sample data based on the normalized design parameters and their corresponding load responses.

[0086] In this embodiment, each design parameter and load response can be normalized respectively, so as to ensure that the constructed samples are not affected by the feature scale, and thus more meaningful features can be extracted from them.

[0087] Exemplarily, when performing the normalization process, for each type of design parameter, the maximum-minimum normalization method can be used. Specifically, the following formula can be adopted:

[0088]

[0089] where x represents the value of the original design parameter, x max represents the maximum value of the design parameter, x min represents the minimum value of the design parameter, represents the value of the design parameter obtained after the normalization process. Taking the design parameter of the top plate thickness as an example, the difference between the maximum top plate thickness and the minimum top plate thickness can be calculated. Then, for each top plate thickness, the quotient of this top plate thickness and this difference can be calculated, and thus the result after the normalization process of this top plate thickness can be obtained. The other design parameters are similar to the top plate thickness and will not be elaborated here.

[0090] Exemplarily, when performing normalization processing, for each load response, a reference value scaling normalization method can be adopted, and the following formula can be specifically used:

[0091]

[0092] Where y represents the value of the original load response determined by the finite element analysis method, y ref represents the reference value of the load response, which can be preset and can be the maximum possible value of the response load response. y represents the value of the load response obtained after normalization processing. Taking the maximum stress as an example, the reference stress can be obtained, and then for each maximum stress, the quotient of the maximum stress and the reference stress can be calculated, and then the result after normalization processing of the maximum stress can be obtained. Load responses such as the maximum deflection are similar to the maximum stress and will not be elaborated here.

[0093] In this embodiment, after separately performing normalization processing on the design parameters and the load responses, sample construction can be performed. For example, the Kernel PCA (Kernel Principal Component Analysis) method based on the kernel function can be used to perform corresponding feature extraction from the design parameters, and then each group of design parameter features obtained can be used as sample features, and the corresponding load response can be used as the sample label of the corresponding design parameter, and then multiple pieces of sample data for training the load prediction model can be obtained. Of course, in other examples, other methods can also be used for sample feature extraction, and this specification does not make special restrictions on this.

[0094] Step 304: Use the sample data to train the initial load prediction model to obtain the trained load prediction model.

[0095] Based on the foregoing step 302, after constructing the sample data, the sample data can be used to train the initial load prediction model.

[0096] Exemplarily, the constructed sample data can be first divided into a training set, a validation set, and a test set. For example, the sample data can be divided into a training set, a validation set, and a test set by means of random sampling to ensure that the distribution of various design parameters among these three data sets is basically uniform.

[0097] In this embodiment, the sample data in the training set can be used to train the initial load prediction model, and then the sample data in the validation set can be used to verify the trained load prediction model. After passing the verification, the sample data in the test set can also be used for testing.

[0098] Optionally, for hyperparameters such as the hidden layer depth, dimension, and learning rate of the load prediction model, multiple different combinations of hyperparameters can be preset in advance, and the initial load prediction models formed by each combination of hyperparameters can be trained separately to obtain multiple trained load prediction models. Then, based on the sample data in the validation set, an optimal load prediction model can be selected from the multiple trained load prediction models for subsequent testing. For example, the optimal load prediction model can be selected based on the sample data in the validation set using the Root Mean Square Error (RMSE) method, etc. This specification does not impose special restrictions on this.

[0099] In this embodiment, the load prediction model that passes the test is determined as the trained load prediction model for subsequent prediction of load response.

[0100] It can be seen that this specification can train a deep learning model as a load prediction model to predict load response. Compared with the finite element calculation method, the time cost can be greatly shortened, thereby improving the design efficiency.

[0101] II. Cost Optimization Function

[0102] 1. Construction of the Cost Optimization Function

[0103] Please refer to Figure 5 , the construction method of the cost optimization function provided in this embodiment may include the following steps:

[0104] Step 502, construct a cost calculation function for the prestressed concrete slab.

[0105] In this embodiment, when designing a prestressed concrete slab, economy is often considered. That is, under the condition of meeting the requirements, the lower the cost of the prestressed concrete slab, the better. Especially for the prestressed UHPC slab made of UHPC material, due to its high cost, more attention needs to be paid to cost considerations during design.

[0106] In this embodiment, a cost calculation function for calculating the cost of the prestressed concrete slab can be constructed first. The cost calculation function can be the sum of the concrete slab cost and the prestress cost. Taking the prestressed UHPC slab with a double T-shaped structure as an example, its cost calculation function can refer to the following formula:

[0107] Q = (t 1 × l + t 2 × h) × C UHPC + S × C S

[0108] Where Q represents the cost calculation function, t 1 represents the top plate thickness, l represents the span, t 2represents the web thickness, h represents the web height, (t 1 × l + t 2 × h) represents the amount of UHPC used, C UHPC represents the unit price of UHPC, S represents the total cross-sectional area of prestressed steel strand tensioning (i.e., the amount of prestress), C S represents the unit price of prestress.

[0109] Step 504, construct a constrained optimization problem with the minimization of the cost calculation function as the objective, and the constraint conditions of the constrained optimization problem include that the load response of the prestressed concrete slab does not exceed a threshold value.

[0110] Based on the foregoing step 502, a constrained optimization problem can be constructed with the minimization of the cost calculation function as the objective, and the constraint condition of this constrained optimization problem is that the load response of the prestressed concrete slab does not exceed a threshold value. For example, the maximum stress of the prestressed concrete slab does not exceed the allowable stress threshold of the design, and the maximum deflection does not exceed the deflection threshold of the design operation, etc.

[0111] Still taking the prestressed UHPC slab as an example, the following constrained optimization problem can be constructed:

[0112] min Q = (t 1 × l + t 2 × h) × C UHPC + S × C S

[0113] s.t. σ max ≤ [σ]

[0114] w max ≤ [w]

[0115] where, σ max represents the maximum stress of the prestressed concrete slab, [σ] represents the allowable stress of the design, that is, the stress threshold, w max represents the maximum deflection of the prestressed concrete slab, [w] represents the allowable deflection of the design, that is, the deflection threshold.

[0116] Step 506, construct the cost optimization function based on the constrained optimization problem, where the cost items in the cost optimization function correspond to the cost calculation function, and the constraint items correspond to the constraint conditions.

[0117] In this embodiment, the Lagrange multiplier method can be used to convert the constrained optimization problem in the foregoing step 504 into an unconstrained optimization problem, and then construct the corresponding cost optimization function to facilitate subsequent solution.

[0118] Exemplarily, the constructed cost optimization function can refer to the following formula:

[0119]

[0120] ReLU(x) = max(x, 0)

[0121]

[0122] Wherein, L(x, λ) represents the cost optimization function to be solved represents some design parameters, d represents the web spacing, and f t represents the tensile strength of UHPC

[0123] t 1 ×l + t 2 ×h + S is the cost item of the cost optimization function, which corresponds to the foregoing cost calculation function

[0124] is the constraint item of the cost optimization function, which corresponds to the foregoing constraint condition. ReLU(x) is an activation function that can be used to convert the constraint condition into a non - negative penalty term

[0125] In this cost optimization function, σ i represents the maximum stress corresponding to a prestressed concrete slab of a certain design parameter, and w i represents the maximum deflection corresponding to a prestressed concrete slab of a certain design parameter. Both σ i and w i can be predicted based on the design parameters using the foregoing load prediction model. λ 1 and λ 2 represent penalty coefficients

[0126] Wherein, the number of iterative updates of the penalty coefficient can be M, and the number of updates of the design parameters can be N

[0127] So far, the constrained optimization problem used in the design process of the prestressed concrete slab has been converted into an unconstrained optimization problem, which facilitates the iterative optimization solution of subsequent design parameters

[0128] Of course, in other examples, when constructing the cost optimization function, the penalty coefficient may not be introduced, and this specification does not make special restrictions on this

[0129] 2. Iterative solution of the cost optimization function

[0130] Please refer to Figure 6 and Figure 7 , the iterative solution process of the cost optimization function provided in this embodiment may include the following steps

[0131] Step 602, obtain the initial design parameters of the prestressed concrete slab as the current design parameters

[0132] In this embodiment, when solving the cost optimization function of the prestressed concrete slab, the initial design parameters can be obtained first, and the initial design parameters can be set in advance based on design requirements. The initial design parameters are the current design parameters used in the first round of iteration. In non-first-round iterations, the current design parameters are the optimized design parameters obtained by optimizing and updating in the previous round of iteration.

[0133] Step 604, use the load prediction model to predict the initial load response corresponding to the initial design parameters as the current load response.

[0134] Based on the foregoing step 602, after obtaining the current design parameters, the trained load prediction model can be used to predict the current load response corresponding to the current design parameters.

[0135] In the first round of iteration, use the load prediction model to predict the initial load response corresponding to the initial design parameters as the current load response used in the first round of iteration.

[0136] In non-first-round iterations, use the load prediction model to predict the current load response corresponding to the current design parameters used in this round of iteration.

[0137] Step 606, use the current design parameters and the current load response to iterate the cost optimization function to optimize the current design parameters, and obtain the optimized design parameters after optimization.

[0138] In this embodiment, based on the current design parameters and their corresponding current load responses, the cost optimization function is iteratively solved to update and optimize the current design parameters. For example, the dual gradient ascent method can be used for solution. Specifically, the optimized design parameters after each round of iteration update are:

[0139]

[0140] where x (i) is the design parameter used in this round of iteration, x (i+1) is the design parameter after optimization and update in this round of iteration, α is the step size for updating the design parameter, is the gradient of the cost optimization function with respect to the design parameter x.

[0141] In another embodiment, when using the penalty coefficient, the penalty coefficient can also be iteratively updated:

[0142]

[0143] where λ (j) is the penalty coefficient used in this round of iteration, λ (j+1) is the penalty coefficient after update in this round of iteration, β is the step size for updating the penalty coefficient, is the gradient of the cost optimization function with respect to the penalty coefficient λ, where j ranges from 1 to M.

[0144] In this embodiment, the values of the update steps α and β, as well as the values of the design parameter update times N and the penalty coefficient update times M can all be preset in advance. The calculation of the above gradients can be obtained through the automatic differentiation algorithm of the deep learning model, and this specification does not impose special restrictions on this.

[0145] Step 608: Determine whether the stopping condition is satisfied. If not, determine the optimized design parameters as the current design parameters for the next iteration and return to Step 602. If satisfied, execute Step 610.

[0146] In this embodiment, after iteratively solving the cost optimization function to obtain the optimized design parameters updated in each iteration, it can be determined whether the stopping condition is satisfied.

[0147] In one example, it can be determined whether the cost of the prestressed concrete slab corresponding to the optimized design parameters after optimization update has been reduced by a preset ratio compared to the cost of the prestressed concrete slab corresponding to the initial design parameters. If the preset ratio has been reduced, it can be determined that the stopping condition is satisfied, and thus the iteration can be stopped and Step 610 can be executed. If the stopping condition is not satisfied, for example, the cost reduction has not reached the preset ratio, the next iteration solution can be continued.

[0148] Among them, the cost of the prestressed concrete slab corresponding to the design parameters can be calculated using the aforementioned cost calculation function. For example, after determining the initial design parameters, the corresponding initial cost can be calculated using this cost calculation function and saved. Subsequently, after obtaining the optimized design parameters, the optimized cost can be calculated using this cost calculation function for comparison with the initial cost, etc.

[0149] In another example, it can be determined whether the preset number of iteration solutions has been reached. In the case of introducing a penalty coefficient, this number of iteration solutions can be the iteration times threshold of the penalty coefficient. If the number of iteration solutions has not been reached, the next iteration solution can be continued. If the number of iteration solutions has been reached, the iteration can be stopped and Step 610 can be executed.

[0150] In yet another example, it can be determined whether the preset number of iteration solutions has been reached while determining whether the cost has been reduced by a preset ratio. When the cost has been reduced by the preset ratio or the preset number of iteration solutions has been reached, it can be determined that the stopping condition is satisfied. If neither is satisfied, it can be determined that the stopping condition is not satisfied.

[0151] Step 610: Verify the latest optimized design parameters obtained after the iteration stops, and when the latest optimized design parameters pass the verification, determine the latest optimized design parameters as the design result of the prestressed concrete slab.

[0152] Based on the judgment result of the foregoing step 608, if the iteration stop condition is satisfied, it can be determined whether the load response corresponding to the latest optimized design parameters obtained after the iteration stops exceeds the corresponding threshold to verify the latest optimized design parameters.

[0153] Specifically, the finite element analysis method can be used to determine the load response corresponding to the latest optimized design parameters, and then for each load response, the magnitude relationship between it and the corresponding load response threshold can be determined respectively. For example, it can be determined whether the maximum stress corresponding to the latest optimized design parameters exceeds the design allowable stress (i.e., the stress threshold), and it can also be determined whether the maximum deflection corresponding to the latest optimized design parameters exceeds the design allowable deflection (i.e., the deflection threshold). If the maximum stress does not exceed the stress threshold and at the same time the maximum deflection does not exceed the deflection threshold, it can be determined that the latest optimized design parameters pass the verification, and the latest optimized design parameters can be determined as the design result of the prestressed concrete slab. If the maximum stress exceeds the stress threshold, or the maximum deflection exceeds the deflection threshold, it can be determined that the latest optimized design parameters do not pass the verification. At this time, the initial design parameters can be re-entered to re-optimize the design parameters. This specification does not make special restrictions on this.

[0154] It can be seen that in this embodiment, after the iteration stops to obtain the latest optimized design parameters, the latest optimized design parameters can also be verified, and after the latest optimized design parameters pass the verification, they can be determined as the design result for the design of the prestressed concrete slab, thereby ensuring the structural safety of the prestressed concrete slab designed with this design result and meeting the design specifications.

[0155] Figure 8 is a schematic structural diagram of a device provided by an exemplary embodiment. Please refer to Figure 8 , at the hardware level, the device includes a processor 802, an internal bus 804, a network interface 806, a memory 808, and a non-volatile memory 810. Of course, there may also be other hardware required for other functions. One or more embodiments of this specification can be implemented in a software manner. For example, the processor 802 reads the corresponding computer program from the non-volatile memory 810 into the memory 808 and then runs it. Of course, in addition to the software implementation manner, one or more embodiments of this specification do not exclude other implementation manners, such as logic devices or a combination of software and hardware, etc. That is to say, the execution subject of the following processing flow is not limited to each logic unit, and can also be hardware or a logic device.

[0156] Please refer to Figure 9 , the design device 900 of the prestressed concrete slab can be applied to such as Figure 8In the device shown, the technical solution of this specification is implemented. Among them, the design device 900 of the prestressed concrete slab may include:

[0157] A function acquisition unit 902 that acquires a cost optimization function for the prestressed concrete slab. The cost optimization function includes a cost item and a load constraint item. The cost item uses the design parameters of the prestressed concrete slab as input variables, and the load constraint item uses the load response of the prestressed concrete slab as input variables;

[0158] An iterative optimization unit 904 that iteratively executes the following steps until a stop condition is met:

[0159] Determine the current design parameters used in the current iteration;

[0160] Predict the current load response corresponding to the current design parameters based on the load prediction model;

[0161] Use the current design parameters and the current load response to perform this iteration of the cost optimization function to optimize the current design parameters and obtain optimized design parameters;

[0162] Update the current design parameters to the optimized design parameters and perform the next iteration;

[0163] A result determination unit 906 that determines the latest optimized design parameters obtained after the iteration stops as the design result of the prestressed concrete slab.

[0164] Optionally, the construction process of the cost optimization function includes:

[0165] Construct a cost calculation function for the prestressed concrete slab;

[0166] Construct a constrained optimization problem with the minimization of the cost calculation function as the goal. The constraint conditions of the constrained optimization problem include that the load response of the prestressed concrete slab does not exceed a threshold;

[0167] Construct the cost optimization function based on the constrained optimization problem. Among them, the cost item in the cost optimization function corresponds to the cost calculation function, and the constraint item corresponds to the constraint condition.

[0168] Optionally, the training process of the load prediction model includes:

[0169] Construct sample data for training the load prediction model. The sample data uses the design parameters of the prestressed concrete slab as features and the load response corresponding to the design parameters as labels;

[0170] Use the sample data to train an initial load prediction model to obtain a trained load prediction model.

[0171] Optionally, the sample data for training the load prediction model includes:

[0172] Obtain several groups of design parameters of prestressed concrete slabs;

[0173] Use the finite element analysis method to respectively determine the load responses corresponding to each group of design parameters;

[0174] Normalize the design parameters and the load responses respectively, and construct the sample data based on the normalized design parameters and their corresponding load responses.

[0175] Optionally, the stopping condition includes:

[0176] Compared with the current design parameters used in the first round of iteration, the cost of the prestressed concrete slab corresponding to the optimized design parameters has been reduced by a preset ratio.

[0177] Optionally, the process by which the result determination unit 906 determines the latest optimized design parameters obtained after the iteration stops as the design result of the prestressed concrete slab includes:

[0178] Verify the latest optimized design parameters;

[0179] When the latest optimized design parameters pass the verification, determine the latest optimized design parameters as the design result of the prestressed concrete slab.

[0180] Based on the same concept as the above method, this specification also provides an electronic device, including: a processor; a memory for storing processor-executable instructions; wherein, the processor realizes the steps of the method as described in any one of the above embodiments by running the executable instructions.

[0181] Based on the same concept as the above method, this specification also provides a computer-readable storage medium, on which computer instructions are stored, and when the instructions are executed by a processor, the steps of the method as described in any one of the above embodiments are realized.

[0182] Based on the same concept as the above method, this specification also provides a computer program product, including computer programs / instructions, and when the computer programs / instructions are executed by a processor, the steps of the method as described in any one of the above embodiments are realized.

Claims

1. A design method for a prestressed concrete slab, characterized in that: The method comprises: Obtaining a cost optimization function of a prestressed concrete slab, wherein the cost optimization function includes a cost item and a load constraint item, wherein the cost item uses a design parameter of the prestressed concrete slab as an input variable, and the load constraint item uses a load response of the prestressed concrete slab as an input variable; Iterate the following steps until the stopping condition is met: Determine the current design parameters used in this iteration; Predicting a current load response corresponding to the current design parameters based on a load prediction model; Using the current design parameters and the current load response to perform this round of iteration on the cost optimization function to optimize the current design parameters, and obtain optimized design parameters; Updating the current design parameters to the optimized design parameters and performing the next round of iteration; The latest optimized design parameters obtained after the iteration stops are determined as the design results of the prestressed concrete slab.

2. The method according to claim 1, characterized in that The construction process of the cost optimization function includes: Cost calculation function for constructing prestressed concrete slabs; Constructing a constrained optimization problem with the goal of minimizing the cost calculation function, wherein the constraint condition of the constrained optimization problem includes that the load response of the prestressed concrete slab does not exceed a threshold value; The cost optimization function is constructed based on the constrained optimization problem, wherein the cost item in the cost optimization function corresponds to the cost calculation function, and the constraint item corresponds to the constraint condition.

3. The method according to claim 1, characterized in that The training process of the load prediction model includes: Constructing sample data for training the load prediction model, wherein the sample data uses design parameters of the prestressed concrete slab as features and load responses corresponding to the design parameters as labels; The sample data is used to train the initial load prediction model to obtain a trained load prediction model.

4. The method according to claim 3, characterized in that The sample data constructed for training the load prediction model includes: Obtaining several sets of design parameters of prestressed concrete slabs; The finite element analysis method is used to determine the load response corresponding to each set of design parameters; The design parameters and the load responses are normalized respectively, and the sample data is constructed based on the normalized design parameters and the corresponding load responses.

5. The method according to claim 1, characterized in that The stop conditions include: Compared with the current design parameters used in the first round of iterations, the cost of the prestressed concrete slab corresponding to the optimized design parameters has been reduced by a preset proportion.

6. The method according to claim 1, characterized in that The process of determining the latest optimized design parameters obtained after the iteration stops as the design result of the prestressed concrete slab includes: Verifying the latest optimized design parameters; When the latest optimized design parameters pass the verification, the latest optimized design parameters are determined as the design results of the prestressed concrete slab.

7. A design device for a prestressed concrete slab, characterized in that: The device comprises: A function acquisition unit, which acquires a cost optimization function of a prestressed concrete slab, wherein the cost optimization function includes a cost item and a load constraint item, wherein the cost item uses a design parameter of the prestressed concrete slab as an input variable, and the load constraint item uses a load response of the prestressed concrete slab as an input variable; Iterate the optimization unit and iterate the following steps until the stopping condition is met: Determine the current design parameters used in this iteration; Predicting a current load response corresponding to the current design parameters based on a load prediction model; Using the current design parameters and the current load response to perform this round of iteration on the cost optimization function to optimize the current design parameters, and obtain optimized design parameters; Updating the current design parameters to the optimized design parameters and performing the next round of iteration; The result determination unit determines the latest optimized design parameters obtained after the iteration stops as the design results of the prestressed concrete slab.

8. An electronic device, characterized in that: include: processor; A memory for storing processor-executable instructions; wherein the processor implements the steps of the method according to any one of claims 1 to 6 by executing the executable instructions.

9. A computer-readable storage medium, characterized in that: Computer instructions are stored thereon, and when the instructions are executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that The method comprises a computer program / instruction, which, when executed by a processor, implements the steps of the method according to any one of claims 1 to 6.

Images