Fitting method and device for peak strain expression of concrete member
By constructing and training the prediction model of concrete components, the peak strain expression is fitted, which solves the problem of difficulty in estimating peak strain in the existing technology, and achieves fast and accurate strain estimation, which improves the efficiency of engineering design and quality inspection.
Patent Information
- Application Number
- CN202411091298.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-09
- Publication Date
- 2024-12-06
AI Technical Summary
The prior art lacks expressions for peak strain of concrete components, which makes it difficult to quickly estimate peak strain in the case of incomplete design parameters, increasing the difficulty of engineering design and quality inspection.
Construct a prediction model for predicting the peak strain of concrete components, and fit the peak strain expression with the characteristic value of the hoop and the characteristic value of the reinforcement as independent variables, and shape it as εcc = ε0(1+α*λt +β*λs, optimize the characteristic coefficient of the coupling α* and the characteristic coefficient of the reinforcement β*.
It provides a method to quickly estimate the peak strain of concrete components, which reduces the work difficulty of engineering design and quality inspection personnel and improves work efficiency.
Smart Images

Figure CN119089079A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of data processing, and in particular to a fitting method and device for a peak strain expression of a concrete component. Background Art
[0002] The peak strain ε of reinforced concrete components (referred to as concrete components) cc The main design parameters with the most significant impact are: stirrup / longitudinal reinforcement ratio, stirrup / longitudinal reinforcement strength and concrete strength. The corresponding stirrup characteristic value λ can be calculated from the stirrup reinforcement ratio, stirrup strength and concrete strength. t The corresponding reinforcement characteristic value λ can be calculated from the longitudinal reinforcement ratio, longitudinal reinforcement strength and concrete strength s ; That is, the peak strain ε cc and the stirrup eigenvalue λ t and reinforcement characteristic value λ s There is a certain mapping relationship between them. If this mapping relationship can be expressed as the peak strain expression (λ t , s is the independent variable, ε cc As the dependent variable), it can help quality inspectors or engineering designers to calculate the design parameters based on only three variables (λ t , s , ε cc ) can quickly estimate another unknown quantity. Unfortunately, there is no similar expression used in practical engineering applications. The technical problem to be solved by the present invention is how to set the peak strain expression and how to confirm the variable coefficients of the peak strain expression. Summary of the invention
[0003] The purpose of the present invention is to provide a fitting method, device, electronic device and computer-readable storage medium for the peak strain expression of a concrete member in view of the defects of the prior art. The present invention first constructs a first prediction model for predicting the peak strain according to the design parameter list of the concrete member, and trains the first prediction model based on a data set (first data set) with a verification label; and sets the peak strain expression with the stirrup characteristic value and the reinforcement characteristic value as independent variables and the peak strain as the dependent variable: ε cc =ε0(1+αλ t +βλ s ); and based on a simulation data set without verification labels (the second data set) and the first prediction model, a data set (the third data set) for expression coefficient fitting operation is constructed; then, based on the third data set, the optimal values of the stirrup characteristic coefficient α and the reinforcement characteristic coefficient β of the peak strain expression are fitted; finally, based on the optimal stirrup characteristic coefficient α *and the optimal reinforcement characteristic coefficient β * The peak strain expression is shaped as: cc =ε0(1+α*λ t +β*λ s The peak strain expression given by the present invention can be used in practical engineering applications to help quality inspectors or engineering designers to calculate the peak strain based on only three variables (λ t , s , ε cc ) can be used to quickly estimate another unknown quantity from any two known quantities, thereby reducing the difficulty of personnel's work and improving work efficiency.
[0004] To achieve the above object, a first aspect of an embodiment of the present invention provides a method for fitting a peak strain expression of a concrete member, the method comprising:
[0005] Constructing a first prediction model for predicting peak strain according to a design parameter list of a concrete component; and training the first prediction model based on a preset first data set; the design parameter list includes nineteen types of component design parameters;
[0006] Construct a linear expression with the stirrup eigenvalue and the reinforcement eigenvalue as independent variables and the peak strain as the dependent variable as the corresponding peak strain expression; the peak strain expression is ε cc =ε0(1+αλ t +βλ s ), where ε cc is the peak strain, ε0 is the preset plain concrete peak strain constant, λ t is the stirrup eigenvalue, λ s is the reinforcement characteristic value, α is the stirrup characteristic coefficient, and β is the reinforcement characteristic coefficient;
[0007] Based on the preset second data set and the first prediction model, a data set for fitting operation is constructed and recorded as a corresponding third data set;
[0008] Based on the third data set, the optimal values of the stirrup characteristic coefficient α and the reinforcement characteristic coefficient β of the peak strain expression are fitted to obtain the corresponding optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * ;
[0009] Based on the optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * Set the fitting result of the peak strain expression; the fitting result is ε cc =ε0(1+α*λ t +β*λ s);The optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * A set of coefficient values is: α * =1.997, β * =0.869.
[0010] Preferably, the component design parameters include stirrup minor axis size, stirrup major axis size, stirrup cross-sectional area, stirrup cross-sectional moment of inertia, stirrup shape code, stirrup spacing, stirrup reinforcement ratio, stirrup strength, number of stirrup limbs, concrete strength, cross-sectional size, concrete cover thickness, component longitudinal size, component aspect ratio, component cross-sectional shape, longitudinal reinforcement strength, longitudinal reinforcement ratio, stirrup characteristic value and reinforcement characteristic value;
[0011] The first data set includes a plurality of first data records; the first data record includes a first training parameter list and a first label peak strain; the first training parameter list includes nineteen categories of component design parameters;
[0012] The second data set is a simulation data set consisting of a plurality of second data records; each of the second data records consists of nineteen types of simulated component design parameters;
[0013] In the second data set, all simulated stirrup minor axis sizes are within a preset size range and satisfy the normal distribution, all simulated stirrup major axis sizes are within a preset size range and satisfy the normal distribution, all simulated stirrup spacings are within a preset spacing range and satisfy the normal distribution, all simulated stirrup strengths are within a preset strength range and satisfy the normal distribution, all simulated concrete strengths are within a preset strength range and satisfy the normal distribution, all simulated cross-sectional sizes are within a preset size range and satisfy the normal distribution, all simulated concrete cover thicknesses are within a preset thickness range and satisfy the normal distribution, all simulated longitudinal reinforcement strengths are within a preset strength range and satisfy the normal distribution, and all simulated longitudinal reinforcement ratios are within a preset reinforcement ratio range and satisfy the normal distribution;
[0014] In the second data set, all simulated stirrup shape codes are within a preset code range and are evenly distributed, all simulated stirrup leg numbers are within a preset leg number range and are evenly distributed, and all simulated component aspect ratios are within a preset aspect ratio range and are evenly distributed;
[0015] In the second data set, the simulated stirrup cross-sectional area in each of the second data records is calculated based on the simulated stirrup long / short axis size in the current record; the simulated stirrup cross-sectional moment of inertia in each of the second data records is calculated based on the simulated stirrup cross-sectional area in the current record; the simulated stirrup reinforcement ratio in each of the second data records is calculated based on the simulated stirrup cross-sectional area, stirrup spacing and number of stirrup legs in the current record; the simulated longitudinal dimension of the component in each of the second data records is calculated based on the simulated cross-sectional dimension and the aspect ratio of the component in the current record; the simulated cross-sectional shape of the component in each of the second data records is a random code within a preset shape coding range; the simulated stirrup characteristic value in each of the second data records is calculated based on the simulated stirrup reinforcement ratio, stirrup strength and concrete strength in the current record; the simulated reinforcement characteristic value in each of the second data records is calculated based on the simulated longitudinal reinforcement ratio, longitudinal reinforcement strength and concrete strength in the current record;
[0016] The third data set includes a plurality of third data records; the third data records include a first stirrup characteristic value, a first reinforcement characteristic value and a second label peak strain.
[0017] Preferably, the first prediction model comprises a normalization module, a BP network and a denormalization module; the output of the normalization module is connected to the input of the BP network, and the output of the BP network is connected to the input of the denormalization module;
[0018] The normalization module is used to perform normalization calculation on each type of component design parameter in the design parameter list input by the model to obtain corresponding normalized parameters; and the corresponding first input vector is composed of the obtained 19 types of normalized parameters and sent to the BP network;
[0019] The BP network includes an input layer, a hidden layer and an output layer;
[0020] The input layer is composed of nineteen parallel input layer units; each of the input layer units is used to receive a type of normalized parameter in the first input vector as a corresponding first input unit value;
[0021] The hidden layer is composed of a plurality of parallel hidden layer units, each of which is fully connected to all the input layer units; each of the hidden layer units is used to perform a fully connected calculation on all the first input unit values to obtain a corresponding first fully connected vector, and perform an activation calculation on the first fully connected vector based on a predicted hidden layer activation function to obtain a corresponding first activation vector; the hidden layer activation function includes at least a sigmoid function;
[0022] The output layer has only one output layer unit, and the output layer unit is fully connected to all the hidden layer units; the output layer unit is used to perform a fully connected calculation on all the first activation vectors to obtain a corresponding second fully connected vector, and perform an activation calculation on the second fully connected vector based on the predicted output layer activation function to obtain a corresponding first activation data, and send the first activation data as the corresponding normalized peak strain to the denormalization module; the output layer activation function at least includes a sigmoid function;
[0023] The denormalization module is used to perform denormalization calculation on the normalized peak strain to obtain the corresponding predicted peak strain and output it.
[0024] Preferably, the training of the first prediction model based on a preset first data set specifically includes:
[0025] The input layer width w of the BP network of the first prediction model is in Set to 19, output layer width w out Set to 1, and based on Get multiple hidden layer widths w h The width range of the hidden layer composed of INT () is a rounding function, the value range of △w is [1, N], and the preset number N is an even number greater than or equal to 2; the hidden layer width range includes N hidden layer widths w h ;
[0026] Based on the width w of each hidden layer h The total number of hidden layer units of the BP network of the first prediction model is set to obtain a corresponding first structure BP network; and the first prediction model with each of the first structure BP networks is recorded as a corresponding first structure prediction model;
[0027] Iteratively optimizing the model parameters of all the first structure prediction models according to the first data set based on the GP algorithm and taking the first structure prediction model with the best performance as the corresponding base model when the number of iterations exceeds a preset iteration number threshold;
[0028] Based on the gradient descent algorithm, further model parameter optimization training is performed on the base model according to the first data set; and the base model obtained at the end of the training is used as the final first prediction model and output.
[0029] Further, the iterative optimization of the model parameters of all the first structure prediction models based on the GP algorithm according to the first data set and taking the first structure prediction model with the best performance as the corresponding base model when the number of iterations exceeds a preset iteration number threshold specifically includes:
[0030] Step 51, randomly extracting a plurality of first data records from the first data set based on a preset first extraction ratio to form a corresponding first evaluation set;
[0031] wherein the first extraction ratio is a percentage less than or equal to 1; the total number of records in the first data set matches the product of the first extraction ratio and the total number of records in the first evaluation set;
[0032] Step 52, initializing the network parameter set of the first structure BP network of each of the first structure prediction models; performing parameter vector conversion on the current network parameter set of each of the first structure BP networks to obtain a corresponding first structure parameter vector; setting a corresponding first model error for each of the first structure prediction models, and initializing each of the first model errors to 1; and initializing a first iteration counter to 1;
[0033] Step 53, taking each of the first structural prediction models as the corresponding current prediction model; and traversing all the first data records of the first evaluation set; and taking the first data record currently traversed as the corresponding current data record during the traversal; and inputting the first training parameter list of the current data record into the current prediction model for prediction processing to obtain the corresponding first training prediction peak strain; and forming a corresponding first prediction label pair from the first training prediction peak strain and the first label peak strain of the current data record; and at the end of this traversal, bringing all the first prediction label pairs corresponding to the current prediction model into a preset first error function for error calculation to obtain the corresponding first error value, and resetting the first model error corresponding to the current prediction model to the corresponding first error value;
[0034] Wherein, the first error function is an MSE error function;
[0035] Step 54, taking the inverse of the first model error corresponding to each of the first structure prediction models as the corresponding first model fitness; summing up all the first model fitnesses obtained to obtain the corresponding first fitness sum; taking the ratio of each of the first model fitnesses to the first fitness sum as the corresponding first model selection probability; and sorting all the first structure prediction models in descending order of the first model selection probability to obtain the corresponding first structure prediction model sequence;
[0036] Step 55, the first and second first structure prediction models of the first structure prediction model sequence are combined into a corresponding first model pair, the third and fourth first structure prediction models are combined into a corresponding first model pair, and so on, until the last two first structure prediction models are combined into a corresponding first model pair; and based on a preset individual crossover algorithm and a preset first crossover probability, vector data crossover processing is performed on the first structure parameter vectors of the two first structure prediction models of each first model pair to obtain two corresponding first structure crossover parameter vectors;
[0037] The individual crossover algorithm is a type of chromosome crossover algorithm in the GA algorithm, and the chromosome crossover algorithm includes at least a single-point crossover algorithm, a multi-point crossover algorithm, and an arithmetic crossover algorithm;
[0038] Step 56, performing vector data mutation processing on the first structural cross parameter vector of each of the first structural prediction models based on a preset individual mutation algorithm and a preset first mutation probability to obtain a corresponding first structural mutation parameter vector;
[0039] The individual mutation algorithm is a type of chromosome mutation algorithm in the GA algorithm, and the chromosome mutation algorithm at least includes a basic position mutation algorithm, a uniform mutation algorithm, a non-uniform mutation algorithm, and a Gaussian approximate mutation algorithm;
[0040] Step 57, identifying whether the first iteration counter exceeds the iteration number threshold; if so, going to step 58; if not, resetting the network parameter set of the first structural BP network of the corresponding first structural prediction model based on each of the first structural variation parameter vectors, adding 1 to the first iteration counter, and returning to step 53;
[0041] Step 58: taking the first first structure prediction model in the latest first structure prediction model sequence as the corresponding base model, and performing parameter solidification on the network parameter set of the first structure BP network of the base model.
[0042] Furthermore, the step of further optimizing the model parameters of the base model according to the first data set based on the gradient descent algorithm specifically includes:
[0043] Step 61, randomly dividing the first data set into data subsets based on a preset first ratio to obtain a corresponding first training set and a second evaluation set;
[0044] Wherein, both the first training set and the second evaluation set include a plurality of the first data records; the ratio of the total number of records in the first training set to the total number of records in the second evaluation set satisfies the first ratio;
[0045] Step 62, extracting the first first data record of the first training set as the corresponding current training record;
[0046] Step 63, inputting the first training parameter list of the current training record into the base model for prediction processing to obtain the corresponding second training prediction peak strain;
[0047] Step 64, bringing the second training predicted peak strain and the first label peak strain of the current training record into a preset first model loss function to calculate and obtain a corresponding first loss value;
[0048] Wherein, the first model loss function at least includes an L1 loss function and an L2 loss function;
[0049] Step 65, identifying whether the first loss value satisfies a preset first loss value range; if the first loss value satisfies the first loss value range, identifying whether the current training record is the last first data record of the first training set, if so, turning to step 66, if not, extracting the next first data record of the first training set as the new current training record and returning to step 63; if the first loss value does not satisfy the first loss value range, performing a round of model parameter modulation on the base model in the direction of minimizing the first model loss function based on a preset first model parameter optimizer, and returning to step 63 at the end of this round of parameter modulation;
[0050] Wherein, the first model parameter optimizer includes at least an SGD optimizer and an ADAM optimizer based on a gradient descent algorithm;
[0051] Step 66, traverse all the first data records of the second evaluation set; and during the traversal, use the first data record currently traversed as the corresponding current evaluation record; and input the first training parameter list of the current evaluation record into the base model for prediction processing to obtain the corresponding third training predicted peak strain; and form a corresponding second predicted label pair with the third training predicted peak strain and the first label peak strain of the current evaluation record;
[0052] Step 67, after traversing all the first data records of the second evaluation set, all the obtained second prediction label pairs are brought into a preset second error function to calculate corresponding second error values;
[0053] The second error function at least includes a MAE error function, an MSE error function and a RMSE error function;
[0054] Step 68, identifying whether the second error value is greater than a preset first error threshold; if so, returning to step 61; if not, performing a parameter solidification on the network parameter set of the first structure BP network of the base model and confirming that the model training of the base model is completed.
[0055] Preferably, the step of constructing a data set for fitting operation based on the preset second data set and the first prediction model as a corresponding third data set specifically includes:
[0056] All the second data records of the second data set are traversed; and during the traversal, the second data record currently traversed is used as the corresponding current data record; and the component design parameters of the nineteen types of simulations in the current data record are extracted to form a corresponding second training parameter list; and the second training parameter list is input into the first prediction model for prediction processing to obtain the corresponding first predicted peak strain, and the first predicted peak strain is used as the corresponding second label peak strain; and the stirrup characteristic value and the reinforcement characteristic value of the current data record are extracted as the corresponding first stirrup characteristic value and the first reinforcement characteristic value; and the obtained first stirrup characteristic value, the first reinforcement characteristic value and the second label peak strain form a corresponding third data record; and at the end of the traversal, all the obtained third data records form the corresponding third data set.
[0057] Preferably, the optimal stirrup characteristic coefficient α and the reinforcement characteristic coefficient β of the peak strain expression are fitted based on the third data set to obtain the corresponding optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * , specifically including:
[0058] Step 81: Count the total number of the third data records of the third data set to obtain a corresponding total number M; and record each of the third data records as a corresponding record R. i , 1≤index i≤M; and each of the records R i The first stirrup characteristic value is recorded as the corresponding λ t,i , the first reinforcement characteristic value is recorded as the corresponding λ s,i , the second label peak strain is recorded as the corresponding and setting a corresponding objective function F and a corresponding objective function constraint condition based on the peak strain expression;
[0059] The objective function F is:
[0060]
[0061] εcc,i To t,i , s,i Substituting into the peak strain expression ε cc =ε0(1+αλ t +βλ s ) calculated peak strain;
[0062] The objective function constraints are:
[0063]
[0064] is cc,i Less than the corresponding probability;
[0065] Here, it should be noted that in order to further improve the fitting accuracy, customized penalty terms can be added to the objective function according to actual application requirements, and the three constraints can be further refined;
[0066] Step 82, based on the COBYLA algorithm, according to the objective function F and the corresponding objective function constraint condition, the optimal values of the stirrup characteristic coefficient α and the reinforcement characteristic coefficient β of the peak strain expression are fitted to obtain the corresponding optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * .
[0067] A second aspect of an embodiment of the present invention provides a device for implementing the fitting method of the peak strain expression of a concrete member described in the first aspect, the device comprising: a model building and training module, an expression building module, a fitting operation data set preparation module, a fitting operation module and an expression finalization module;
[0068] The model building and training module is used to build a first prediction model for predicting peak strain according to a design parameter list of concrete components; and to train the first prediction model based on a preset first data set; the design parameter list includes nineteen types of component design parameters;
[0069] The expression building module is used to build a linear expression with the stirrup characteristic value and the reinforcement characteristic value as independent variables and the peak strain as the dependent variable as the corresponding peak strain expression; the peak strain expression is ε cc =ε0(1+αλ t +βλ s ), where ε cc is the peak strain, ε0 is the preset plain concrete peak strain constant, λ t is the stirrup eigenvalue, λ sis the reinforcement characteristic value, α is the stirrup characteristic coefficient, and β is the reinforcement characteristic coefficient;
[0070] The fitting operation data set preparation module is used to construct a data set for fitting operation based on the preset second data set and the first prediction model as a corresponding third data set;
[0071] The fitting operation module is used to perform a fitting operation on the optimal values of the stirrup characteristic coefficient α and the reinforcement characteristic coefficient β of the peak strain expression based on the third data set to obtain the corresponding optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * ;
[0072] The expression finalization module is used to determine the optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * Set the fitting result of the peak strain expression; the fitting result is ε cc =ε0(1+α*λ t +β*λ s );The optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * A set of coefficient values is: α * =1.997, β * =0.869.
[0073] A third aspect of an embodiment of the present invention provides an electronic device, including: a memory, a processor, and a transceiver;
[0074] The processor is used to couple with the memory, read and execute instructions in the memory, so as to implement the method steps described in the first aspect above;
[0075] The transceiver is coupled to the processor, and the processor controls the transceiver to send and receive messages.
[0076] A fourth aspect of an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions. When the computer instructions are executed by a computer, the computer executes the instructions of the method described in the first aspect above.
[0077] The embodiment of the present invention provides a fitting method, device, electronic device and computer-readable storage medium for a peak strain expression of a concrete component. As can be seen from the above content, the embodiment of the present invention first constructs a first prediction model for predicting the peak strain according to the design parameter list of the concrete component, and trains the first prediction model based on a data set (first data set) with a verification label; and sets the peak strain expression with the stirrup characteristic value and the reinforcement characteristic value as independent variables and the peak strain as the dependent variable: ε cc =ε0(1+αλ t +βλ s ); and based on a simulation data set without verification labels (the second data set) and the first prediction model, a data set (the third data set) for expression coefficient fitting operation is constructed; then, based on the third data set, the optimal values of the stirrup characteristic coefficient α and the reinforcement characteristic coefficient β of the peak strain expression are fitted; finally, based on the optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * The peak strain expression is shaped as: cc =ε0(1+α*λ t +β*λ s ). The peak strain expression given in the embodiment of the present invention is used in actual engineering applications to help quality inspectors or engineering designers to calculate the peak strain based on only three variables (λ t , s , ε cc ) can quickly estimate another unknown quantity from any two known quantities; the embodiments of the present invention can reduce the work difficulty of quality inspectors or engineering designers and improve work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0078] Figure 1 A schematic diagram of a fitting method for a peak strain expression of a concrete component provided in the first embodiment of the present invention;
[0079] Figure 2 A module structure diagram of a fitting device for a peak strain expression of a concrete member provided in the second embodiment of the present invention;
[0080] Figure 3 A schematic diagram of the structure of an electronic device provided in Embodiment 3 of the present invention. DETAILED DESCRIPTION
[0081] In order to make the purpose, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0082] Embodiment 1 of the present invention provides a fitting method for a peak strain expression of a concrete component, such as Figure 1 As shown in the schematic diagram of a fitting method for a peak strain expression of a concrete component provided in the first embodiment of the present invention, the method mainly comprises the following steps:
[0083] Step 1, constructing a first prediction model for predicting peak strain according to a design parameter list of a concrete component; and training the first prediction model based on a preset first data set;
[0084] Specifically, it includes: step 11, constructing a first prediction model for predicting peak strain according to a design parameter list of a concrete component;
[0085] The design parameter list includes 19 types of component design parameters; the component design parameters include stirrup minor axis size, stirrup major axis size, stirrup cross-sectional area, stirrup cross-sectional inertia moment, stirrup shape code, stirrup spacing, stirrup reinforcement ratio, stirrup strength, number of stirrup legs, concrete strength, cross-sectional size, concrete cover thickness, component longitudinal size, component aspect ratio, component cross-sectional shape, longitudinal reinforcement strength, longitudinal reinforcement ratio, stirrup characteristic value and reinforcement characteristic value;
[0086] Here, the first prediction model of the embodiment of the present invention includes a normalization module, a BP network and a denormalization module; the output of the normalization module is connected to the input of the BP network, and the output of the BP network is connected to the input of the denormalization module;
[0087] The normalization module is used to perform normalization calculation on each type of component design parameter in the design parameter list input by the model to obtain the corresponding normalized parameter; and the corresponding first input vector is composed of the obtained 19 types of normalized parameters and sent to the BP network;
[0088] BP network includes input layer, hidden layer and output layer;
[0089] The input layer is composed of nineteen parallel input layer units; each input layer unit is used to receive a type of normalized parameter in the first input vector as a corresponding first input unit value;
[0090] The hidden layer is composed of a plurality of parallel hidden layer units, each hidden layer unit is fully connected to all input layer units; each hidden layer unit is used to perform a fully connected calculation on all first input unit values to obtain a corresponding first fully connected vector, and perform an activation calculation on the first fully connected vector based on a predicted hidden layer activation function to obtain a corresponding first activation vector; the hidden layer activation function includes at least a sigmoid function;
[0091] The output layer has only one output layer unit, and the output layer unit is fully connected to all hidden layer units; the output layer unit is used to perform full connection calculation on all first activation vectors to obtain corresponding second fully connected vectors, and perform activation calculation on the second fully connected vector based on the predicted output layer activation function to obtain corresponding first activation data, and send the first activation data as the corresponding normalized peak strain to the denormalization module; the output layer activation function at least includes a sigmoid function;
[0092] The denormalization module is used to perform denormalization calculation on the normalized peak strain to obtain the corresponding predicted peak strain and output it;
[0093] Step 12, training the first prediction model based on the preset first data set;
[0094] The first data set includes a plurality of first data records; the first data record includes a first training parameter list and a first label peak strain; the first training parameter list includes nineteen types of component design parameters;
[0095] Specifically, step 121: set the input layer width w of the BP network of the first prediction model in Set to 19, output layer width w out Set to 1, and based on Get multiple hidden layer widths w h The width range of the hidden layers that it consists of;
[0096] Among them, F INT () is a floor function, the value range of △w is [1, N], and the preset number N is an even number greater than or equal to 2; the hidden layer width range includes N hidden layer widths w h ;
[0097] Step 122, based on the width w of each hidden layer h The total number of hidden layer units of the BP network of the first prediction model is set to obtain a corresponding first structure BP network; and the first prediction model with each first structure BP network is recorded as the corresponding first structure prediction model;
[0098] Step 123, iteratively optimizing the model parameters of all first structure prediction models according to the first data set based on the GP algorithm and taking the first structure prediction model with the best performance as the corresponding base model when the number of iterations exceeds a preset iteration number threshold;
[0099] Specifically, it includes: step 1231, randomly extracting a plurality of first data records from a first data set based on a preset first extraction ratio to form a corresponding first evaluation set;
[0100] wherein the first extraction ratio is a percentage less than or equal to 1; the total number of records in the first data set matches the product of the first extraction ratio and the total number of records in the first evaluation set;
[0101] Step 1232, initializing the network parameter set of the first structure BP network of each first structure prediction model; performing parameter vector conversion on the current network parameter set of each first structure BP network to obtain a corresponding first structure parameter vector; setting a corresponding first model error for each first structure prediction model, and initializing each first model error to 1; and initializing the first iteration counter to 1;
[0102] Step 1233, taking each first structural prediction model as the corresponding current prediction model; and traversing all the first data records of the first evaluation set; and taking the currently traversed first data record as the corresponding current data record during the traversal; and inputting the first training parameter list of the current data record into the current prediction model for prediction processing to obtain the corresponding first training prediction peak strain; and forming a corresponding first prediction label pair from the first training prediction peak strain and the first label peak strain of the current data record; and at the end of this traversal, bringing all the first prediction label pairs corresponding to the current prediction model into the preset first error function for error calculation to obtain the corresponding first error value, and resetting the first model error corresponding to the current prediction model to the corresponding first error value;
[0103] Wherein, the first error function is the MSE error function;
[0104] Step 1234, taking the inverse of the first model error corresponding to each first structure prediction model as the corresponding first model fitness; summing up all the obtained first model fitnesses to obtain the corresponding first fitness sum; taking the ratio of each first model fitness to the first fitness sum as the corresponding first model selection probability; and sorting all the first structure prediction models in descending order of the first model selection probability to obtain the corresponding first structure prediction model sequence;
[0105] Step 1235, the 1st and 2nd first structure prediction models of the first structure prediction model sequence are combined into a corresponding first model pair, the 3rd and 4th first structure prediction models are combined into a corresponding first model pair, and so on, until the last two first structure prediction models are combined into a corresponding first model pair; and based on a preset individual crossover algorithm and a preset first crossover probability, vector data crossover processing is performed on the first structure parameter vectors of the two first structure prediction models of each first model pair to obtain two corresponding first structure crossover parameter vectors;
[0106] The individual crossover algorithm is a type of chromosome crossover algorithm in the GA algorithm, and the chromosome crossover algorithm includes at least a single-point crossover algorithm, a multi-point crossover algorithm, and an arithmetic crossover algorithm; the first crossover probability is a preset probability value, such as 0.4;
[0107] Step 1236, performing vector data mutation processing on the first structure cross parameter vectors of each first structure prediction model based on a preset individual mutation algorithm and a preset first mutation probability to obtain a corresponding first structure mutation parameter vector;
[0108] The individual mutation algorithm is a type of chromosome mutation algorithm in the GA algorithm, and the chromosome mutation algorithm includes at least a basic position mutation algorithm, a uniform mutation algorithm, a non-uniform mutation algorithm, and a Gaussian approximate mutation algorithm; the first mutation probability is a preset probability value, such as 0.01;
[0109] Step 1237, identifying whether the first iteration counter exceeds the iteration number threshold; if so, going to step 1238; if not, resetting the network parameter set of the first structure BP network of the corresponding first structure prediction model based on each first structure variation parameter vector, adding 1 to the first iteration counter, and returning to step 1233;
[0110] Here, the iteration number threshold is a preset integer;
[0111] Step 1238, taking the first first structure prediction model in the latest first structure prediction model sequence as the corresponding base model, and performing parameter solidification on the network parameter set of the first structure BP network of the base model;
[0112] Step 124, performing further model parameter optimization training on the base model according to the first data set based on a gradient descent algorithm;
[0113] Specifically, it includes: step 1241, randomly dividing the first data set into data subsets based on a preset first ratio to obtain a corresponding first training set and a second evaluation set;
[0114] The first ratio is a preset ratio, such as 8:2; the first training set and the second evaluation set both include a plurality of first data records; the ratio of the total number of records in the first training set to the total number of records in the second evaluation set satisfies the first ratio;
[0115] Step 1242, extracting the first first data record of the first training set as the corresponding current training record;
[0116] Step 1243, inputting the first training parameter list of the current training record into the base model for prediction processing to obtain the corresponding second training prediction peak strain;
[0117] Step 1244, bringing the second training predicted peak strain and the first label peak strain of the current training record into a preset first model loss function to calculate and obtain a corresponding first loss value;
[0118] Wherein, the first model loss function includes at least an L1 loss function and an L2 loss function;
[0119] Step 1245, identifying whether the first loss value satisfies a preset first loss value range; if the first loss value satisfies the first loss value range, identifying whether the current training record is the last first data record of the first training set, if so, turning to step 1246, if not, extracting the next first data record of the first training set as a new current training record and returning to step 1243; if the first loss value does not satisfy the first loss value range, performing a round of model parameter modulation on the base model in the direction of minimizing the first model loss function based on a preset first model parameter optimizer, and returning to step 1243 at the end of this round of parameter modulation;
[0120] Here, the first loss value range is a preset loss value range; the first model parameter optimizer includes at least an SGD optimizer and an ADAM optimizer based on a gradient descent algorithm;
[0121] Step 1246, traverse all first data records of the second evaluation set; and during the traversal, use the currently traversed first data record as the corresponding current evaluation record; and input the first training parameter list of the current evaluation record into the base model for prediction processing to obtain the corresponding third training predicted peak strain; and form a corresponding second predicted label pair from the third training predicted peak strain and the first label peak strain of the current evaluation record;
[0122] Step 1247, after traversing all the first data records of the second evaluation set, all the obtained second prediction label pairs are brought into a preset second error function to calculate and obtain corresponding second error values;
[0123] The second error function includes at least a MAE error function, an MSE error function and a RMSE error function;
[0124] Step 1248, identifying whether the second error value is greater than a preset first error threshold; if so, returning to step 1241; if not, performing a parameter solidification on the network parameter set of the first structure BP network of the base model and confirming that the model training of the base model is completed;
[0125] Here, the first error threshold is a preset error value;
[0126] Step 125, the base model obtained at the end of the training is used as the final first prediction model and output.
[0127] Step 2: Construct a linear expression with the stirrup eigenvalue and the reinforcement eigenvalue as independent variables and the peak strain as the dependent variable as the corresponding peak strain expression.
[0128] The peak strain expression is ε cc =ε0(1+αλ t +βλ s ),
[0129] ε cc is the peak strain,
[0130] ε0 is the preset peak strain constant of plain concrete,
[0131] λ t is the stirrup eigenvalue,
[0132] λ s is the reinforcement characteristic value,
[0133] α is the stirrup characteristic coefficient,
[0134] β is the reinforcement characteristic coefficient.
[0135] Here, λ t , s is the independent variable, ε cc is the dependent variable, ε0 is a preset known constant; and the stirrup characteristic coefficient α and the reinforcement characteristic coefficient β are two variable coefficients that need to be further confirmed in the embodiment of the present invention.
[0136] Step 3, constructing a data set for fitting operation based on the preset second data set and the first prediction model and recording it as the corresponding third data set;
[0137] Here, the second data set of the embodiment of the present invention is a simulated data set, which is composed of multiple second data records; each second data record is composed of nineteen types of simulated component design parameters; the second data set of the embodiment of the present invention can be obtained by random sampling and combination of the first data set, or can be simulated and generated by a data generator. The embodiment of the present invention does not limit the specific construction process of the data set, but imposes the following restrictions on the data features of the data set:
[0138] In the second data set, all simulated stirrup minor axis sizes are within a preset size range and satisfy the normal distribution, all simulated stirrup major axis sizes are within a preset size range and satisfy the normal distribution, all simulated stirrup spacings are within a preset spacing range and satisfy the normal distribution, all simulated stirrup strengths are within a preset strength range and satisfy the normal distribution, all simulated concrete strengths are within a preset strength range and satisfy the normal distribution, all simulated cross-sectional sizes are within a preset size range and satisfy the normal distribution, all simulated concrete cover thicknesses are within a preset thickness range and satisfy the normal distribution, all simulated longitudinal reinforcement strengths are within a preset strength range and satisfy the normal distribution, and all simulated longitudinal reinforcement ratios are within a preset reinforcement ratio range and satisfy the normal distribution;
[0139] In the second data set, all simulated stirrup shape codes are within a preset code range and are evenly distributed, all simulated stirrup leg numbers are within a preset leg number range and are evenly distributed, and all simulated component aspect ratios are within a preset aspect ratio range and are evenly distributed;
[0140] In the second data set, the simulated stirrup cross-sectional area in each second data record is calculated based on the simulated stirrup long / short axis size in the current record; the simulated stirrup cross-sectional moment of inertia in each second data record is calculated based on the simulated stirrup cross-sectional area in the current record; the simulated stirrup reinforcement ratio in each second data record is calculated based on the simulated stirrup cross-sectional area, stirrup spacing and number of stirrup limbs in the current record; the simulated longitudinal dimension of the component in each second data record is calculated based on the simulated cross-sectional dimension and the aspect ratio of the component in the current record; the simulated cross-sectional shape of the component in each second data record is a random code within a preset shape coding range; the simulated stirrup characteristic value in each second data record is calculated based on the simulated stirrup reinforcement ratio, stirrup strength and concrete strength in the current record; the simulated reinforcement characteristic value in each second data record is calculated based on the simulated longitudinal reinforcement ratio, longitudinal reinforcement strength and concrete strength in the current record;
[0141] The third data set of the embodiment of the present invention includes a plurality of third data records; the third data record includes a first stirrup characteristic value, a first reinforcement characteristic value, and a second label peak strain;
[0142] The processing steps of the current step 3 specifically include:
[0143] All second data records of the second data set are traversed; and during the traversal, the second data record currently traversed is used as the corresponding current data record; and the design parameters of the component of the nineteen types of simulations in the current data record are extracted to form a corresponding second training parameter list; and the second training parameter list is input into the first prediction model for prediction processing to obtain the corresponding first predicted peak strain, and the first predicted peak strain is used as the corresponding second label peak strain; and the stirrup characteristic value and the reinforcement characteristic value of the current data record are extracted as the corresponding first stirrup characteristic value and the first reinforcement characteristic value; and a corresponding third data record is formed by the obtained first stirrup characteristic value, the first reinforcement characteristic value and the second label peak strain; and at the end of the traversal, all the obtained third data records are formed into a corresponding third data set.
[0144] Step 4: Based on the third data set, the optimal values of the stirrup characteristic coefficient α and the reinforcement characteristic coefficient β of the peak strain expression are fitted to obtain the corresponding optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * ;
[0145] Specifically, the step 81 includes: counting the total number of third data records of the third data set to obtain a corresponding total number M; and recording each third data record as a corresponding record R i , 1≤index i≤M; and each record R i The first stirrup eigenvalue of t,i , the first reinforcement characteristic value is recorded as the corresponding λ s,i , the second label peak strain is recorded as the corresponding And set the corresponding objective function F and the corresponding objective function constraint condition based on the peak strain expression;
[0146] Here, the objective function F of the embodiment of the present invention is:
[0147]
[0148] ε cc,i To t,i , s,i Substitute the peak strain expression ε into cc =ε0(1+αλ t +βλ s ) calculated peak strain;
[0149] The objective function constraint conditions of the embodiment of the present invention are:
[0150]
[0151] is cc,i Less than the corresponding probability;
[0152] Step 82, based on the Constrained Optimization BY Linear Approximations (COBYLA) algorithm, the optimal values of the stirrup characteristic coefficient α and the reinforcement characteristic coefficient β of the peak strain expression are fitted according to the objective function F and the corresponding objective function constraints to obtain the corresponding optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * .
[0153] Step 5: Based on the optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * Set the fitting result of the peak strain expression;
[0154] Here, the fitting result is also the peak strain expression finally obtained in the embodiment of the present invention, which is ε cc =ε0(1+α*λ t +α*λ s ).
[0155] It should be noted that a set of optimal stirrup characteristic coefficients α obtained based on the embodiment of the present invention is * and the optimal reinforcement characteristic coefficient β * The coefficient value is: α * =1.997, β * =0.869.
[0156] Figure 2 This is a module structure diagram of a fitting device for a peak strain expression of a concrete component provided in the second embodiment of the present invention. The device is a terminal device or a server that implements the aforementioned method embodiment, and can also be a device that enables the aforementioned terminal device or server to implement the aforementioned method embodiment. For example, the device can be a device or chip system of the aforementioned terminal device or server. Figure 2 As shown, the device includes: a model building and training module 201, an expression building module 202, a fitting operation data set preparation module 203, a fitting operation module 204 and an expression finalization module 205.
[0157] The model building and training module 201 is used to build a first prediction model for predicting peak strain according to a design parameter list of a concrete component; and to train the first prediction model based on a preset first data set; the design parameter list includes nineteen categories of component design parameters.
[0158] The expression building module 202 is used to build a linear expression with the stirrup characteristic value and the reinforcement characteristic value as independent variables and the peak strain as the dependent variable as the corresponding peak strain expression; the peak strain expression is ε cc =ε0(1+αλ t +βλ s ), where ε cc is the peak strain, ε0 is the preset plain concrete peak strain constant, λ t is the stirrup eigenvalue, λ s is the characteristic value of reinforcement, α is the characteristic coefficient of stirrups, and β is the characteristic coefficient of reinforcement.
[0159] The fitting operation data set preparation module 203 is used to construct a data set for fitting operation based on the preset second data set and the first prediction model, which is recorded as the corresponding third data set.
[0160] The fitting operation module 204 is used to perform a fitting operation on the optimal values of the stirrup characteristic coefficient α and the reinforcement characteristic coefficient β of the peak strain expression based on the third data set to obtain the corresponding optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * .
[0161] The expression finalization module 205 is used to calculate the optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * Set the fitting result of the peak strain expression; the fitting result is ε cc =ε0(1+α*λ t +β*λ s );Optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * A set of coefficient values is: α * =1.997, β * =0.869.
[0162] The embodiment of the present invention provides a fitting device for a peak strain expression of a concrete member, which can execute the method steps in the above method embodiment. Its implementation principle and technical effects are similar and will not be repeated here.
[0163] It should be noted that it should be understood that the division of the various modules of the above device is only a division of logical functions. In actual implementation, they can be fully or partially integrated into one physical entity, or they can be physically separated. And these modules can all be implemented in the form of software called by processing elements; they can also be all implemented in the form of hardware; some modules can also be implemented in the form of software called by processing elements, and some modules can be implemented in the form of hardware. For example, the model building and training module can be a separately established processing element, or it can be integrated in a chip of the above device. In addition, it can also be stored in the memory of the above device in the form of program code, and called and executed by a processing element of the above device. The function of the above-mentioned module is determined. The implementation of other modules is similar. In addition, all or part of these modules can be integrated together, or they can be implemented independently. The processing element described here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each module above can be completed by an integrated logic circuit of hardware in the processor element or instructions in software form.
[0164] For example, the above modules may be one or more integrated circuits configured to implement the above methods, such as one or more application specific integrated circuits (ASIC), or one or more digital signal processors (DSP), or one or more field programmable gate arrays (FPGA). For another example, when a module above is implemented in the form of a processing element scheduling program code, the processing element may be a general-purpose processor, such as a central processing unit (CPU) or other processor that can call program code. For another example, these modules may be integrated together and implemented in the form of a system-on-a-chip (SOC).
[0165] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments may be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the above method embodiments are generated. The above computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The above-mentioned computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the above-mentioned computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, Bluetooth, microwave, etc.) methods. The above-mentioned computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media integrated. The above-mentioned available medium can be a magnetic medium (such as a floppy disk, a hard disk, a tape), an optical medium (such as a DVD), or a semiconductor medium (such as a solid state disk (SSD)), etc.
[0166] Figure 3 This is a schematic diagram of the structure of an electronic device provided in Embodiment 3 of the present invention. The electronic device may be a terminal device or a server for implementing the method of the aforementioned embodiment, or may be a terminal device or a server for implementing the method of the aforementioned embodiment connected to the aforementioned terminal device or server. Figure 3 As shown, the electronic device may include: a processor 301 (such as a CPU), a memory 302, and a transceiver 303; the transceiver 303 is coupled to the processor 301, and the processor 301 controls the transceiver 303. Various instructions may be stored in the memory 302 to complete various processing functions and implement the processing steps described in the aforementioned embodiment method. Preferably, the electronic device involved in the embodiment of the present invention also includes: a power supply 304, a system bus 305 and a communication port 306. The system bus 305 is used to realize the communication connection between components. The above-mentioned communication port 306 is used for connecting and communicating between the electronic device and other peripherals.
[0167] exist Figure 3The system bus 305 mentioned in the figure can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The system bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus. The communication interface is used to realize the communication between the database access device and other devices (such as clients, read-write libraries, and read-only libraries). The memory may include random access memory (RAM), and may also include non-volatile memory (Non-Volatile Memory), such as at least one disk storage.
[0168] The above-mentioned processor can be a general-purpose processor, including a central processing unit CPU, a network processor (Network Processor, NP), a graphics processing unit (Graphics Processing Unit, GPU), etc.; it can also be a digital signal processor DSP, an application-specific integrated circuit ASIC, a field programmable gate array FPGA or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.
[0169] It should be noted that an embodiment of the present invention further provides a computer-readable storage medium, in which instructions are stored. When the computer-readable storage medium is run on a computer, the computer executes the method and processing process provided in the above embodiments.
[0170] An embodiment of the present invention further provides a chip for executing instructions, wherein the chip is used to execute the processing steps described in the aforementioned method embodiment.
[0171] The embodiment of the present invention provides a fitting method, device, electronic device and computer-readable storage medium for a peak strain expression of a concrete component. As can be seen from the above content, the embodiment of the present invention first constructs a first prediction model for predicting the peak strain according to the design parameter list of the concrete component, and trains the first prediction model based on a data set (first data set) with a verification label; and sets the peak strain expression with the stirrup characteristic value and the reinforcement characteristic value as independent variables and the peak strain as the dependent variable: ε cc =ε0(1+αλ t +βλ s); and based on a simulation data set without verification labels (the second data set) and the first prediction model, a data set (the third data set) for expression coefficient fitting operation is constructed; then, based on the third data set, the optimal values of the stirrup characteristic coefficient α and the reinforcement characteristic coefficient β of the peak strain expression are fitted; finally, based on the optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * The peak strain expression is shaped as: cc =ε0(1+α*λ t +β*λ s ). The peak strain expression given in the embodiment of the present invention is used in actual engineering applications to help quality inspectors or engineering designers to calculate the peak strain based on only three variables (λ t , s , ε cc ) can quickly estimate another unknown quantity from any two known quantities; the embodiments of the present invention can reduce the work difficulty of quality inspectors or engineering designers and improve work efficiency.
[0172] The professionals should further realize that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to the function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.
[0173] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
[0174] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. A fitting method for peak strain expression of concrete components, characterized in that: The method comprises: Constructing a first prediction model for predicting peak strain according to a design parameter list of a concrete component; and training the first prediction model based on a preset first data set; the design parameter list includes nineteen types of component design parameters; Construct a linear expression with stirrup eigenvalue and reinforcement eigenvalue as independent variables and peak strain as dependent variable as the corresponding peak strain expression; the peak strain expression is ε cc =ε0(1+αλ t +βλ s ), where ε cc is the peak strain, ε0 is the preset plain concrete peak strain constant, λ t is the stirrup eigenvalue, λ s is the reinforcement characteristic value, α is the stirrup characteristic coefficient, and β is the reinforcement characteristic coefficient; Based on the preset second data set and the first prediction model, a data set for fitting operation is constructed and recorded as a corresponding third data set; Based on the third data set, the optimal values of the stirrup characteristic coefficient α and the reinforcement characteristic coefficient β of the peak strain expression are fitted to obtain the corresponding optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * ; Based on the optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * Set the fitting result of the peak strain expression; the fitting result is ε cc =ε0(1+α * λ t +β * λ s );The optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * A set of coefficient values is: α * =1.997, β * =0.
869.
2. The fitting method of the peak strain expression of a concrete member according to claim 1 is characterized in that: The component design parameters include stirrup minor axis size, stirrup major axis size, stirrup cross-sectional area, stirrup cross-sectional moment of inertia, stirrup shape code, stirrup spacing, stirrup reinforcement ratio, stirrup strength, number of stirrup limbs, concrete strength, cross-sectional size, concrete cover thickness, component longitudinal size, component aspect ratio, component cross-sectional shape, longitudinal reinforcement strength, longitudinal reinforcement ratio, stirrup characteristic value and reinforcement characteristic value; The first data set includes a plurality of first data records; the first data record includes a first training parameter list and a first label peak strain; the first training parameter list includes nineteen categories of component design parameters; The second data set is a simulation data set consisting of a plurality of second data records; each of the second data records consists of nineteen types of simulated component design parameters; In the second data set, all simulated stirrup minor axis sizes are within a preset size range and satisfy the normal distribution, all simulated stirrup major axis sizes are within a preset size range and satisfy the normal distribution, all simulated stirrup spacings are within a preset spacing range and satisfy the normal distribution, all simulated stirrup strengths are within a preset strength range and satisfy the normal distribution, all simulated concrete strengths are within a preset strength range and satisfy the normal distribution, all simulated cross-sectional sizes are within a preset size range and satisfy the normal distribution, all simulated concrete cover thicknesses are within a preset thickness range and satisfy the normal distribution, all simulated longitudinal reinforcement strengths are within a preset strength range and satisfy the normal distribution, and all simulated longitudinal reinforcement ratios are within a preset reinforcement ratio range and satisfy the normal distribution; In the second data set, all simulated stirrup shape codes are within a preset code range and are evenly distributed, all simulated stirrup leg numbers are within a preset leg number range and are evenly distributed, and all simulated component aspect ratios are within a preset aspect ratio range and are evenly distributed; In the second data set, the cross-sectional area of the stirrups simulated in each of the second data records is calculated based on the major / minor axis size of the stirrups simulated in the current record; The stirrup cross-sectional inertia moment simulated in each of the second data records is calculated based on the stirrup cross-sectional area simulated in the current record; the stirrup reinforcement ratio simulated in each of the second data records is calculated based on the stirrup cross-sectional area, stirrup spacing and number of stirrup legs simulated in the current record; the component longitudinal size simulated in each of the second data records is calculated based on the cross-sectional size and component aspect ratio simulated in the current record; the component cross-sectional shape simulated in each of the second data records is a random code within a preset shape coding range; The stirrup characteristic values simulated in each of the second data records are calculated based on the stirrup reinforcement ratio, stirrup strength and concrete strength simulated in the current record; the reinforcement characteristic values simulated in each of the second data records are calculated based on the longitudinal reinforcement ratio, longitudinal reinforcement strength and concrete strength simulated in the current record; The third data set includes a plurality of third data records; the third data records include a first stirrup characteristic value, a first reinforcement characteristic value and a second label peak strain.
3. The fitting method of the peak strain expression of a concrete member according to claim 2 is characterized in that: The first prediction model includes a normalization module, a BP network and a denormalization module; the output of the normalization module is connected to the input of the BP network, and the output of the BP network is connected to the input of the denormalization module; The normalization module is used to perform normalization calculation on each type of component design parameters in the design parameter list input by the model to obtain corresponding normalized parameters; The obtained 19 types of normalized parameters form a corresponding first input vector and send it to the BP network; The BP network includes an input layer, a hidden layer and an output layer; The input layer is composed of nineteen parallel input layer units; each of the input layer units is used to receive a type of normalized parameter in the first input vector as a corresponding first input unit value; The hidden layer is composed of a plurality of parallel hidden layer units, each of which is fully connected to all the input layer units; each of the hidden layer units is used to perform a fully connected calculation on all the first input unit values to obtain a corresponding first fully connected vector, and perform an activation calculation on the first fully connected vector based on a predicted hidden layer activation function to obtain a corresponding first activation vector; the hidden layer activation function includes at least a sigmoid function; The output layer has only one output layer unit, and the output layer unit is fully connected to all the hidden layer units; the output layer unit is used to perform a fully connected calculation on all the first activation vectors to obtain a corresponding second fully connected vector, and perform an activation calculation on the second fully connected vector based on the predicted output layer activation function to obtain a corresponding first activation data, and send the first activation data as the corresponding normalized peak strain to the denormalization module; The output layer activation function at least includes a sigmoid function; The denormalization module is used to perform denormalization calculation on the normalized peak strain to obtain the corresponding predicted peak strain and output it.
4. The fitting method of the peak strain expression of a concrete member according to claim 3 is characterized in that: The training of the first prediction model based on the preset first data set specifically includes: The input layer width w of the BP network of the first prediction model is in Set to 19, output layer width w out Set to 1, and based on Get multiple hidden layer widths w h The width range of the hidden layer composed of INT () is a rounding function, the value range of △w is [1, N], and the preset number N is an even number greater than or equal to 2; the hidden layer width range includes N hidden layer widths w h ; Based on the width w of each hidden layer h The total number of hidden layer units of the BP network of the first prediction model is set to obtain a corresponding first structure BP network; and the first prediction model with each of the first structure BP networks is recorded as a corresponding first structure prediction model; Iteratively optimizing the model parameters of all the first structure prediction models according to the first data set based on the GP algorithm and taking the first structure prediction model with the best performance as the corresponding base model when the number of iterations exceeds a preset iteration number threshold; Based on the gradient descent algorithm, further model parameter optimization training is performed on the base model according to the first data set; and the base model obtained at the end of the training is used as the final first prediction model and output.
5. The fitting method of the peak strain expression of a concrete member according to claim 4, characterized in that: The iterative optimization of the model parameters of all the first structure prediction models according to the first data set based on the GP algorithm and taking the first structure prediction model with the best performance as the corresponding base model when the number of iterations exceeds a preset iteration number threshold specifically includes: Step 51, randomly extracting a plurality of first data records from the first data set based on a preset first extraction ratio to form a corresponding first evaluation set; wherein the first extraction ratio is a percentage less than or equal to 1; the total number of records in the first data set matches the product of the first extraction ratio and the total number of records in the first evaluation set; Step 52, initializing the network parameter set of the first structure BP network of each of the first structure prediction models; performing parameter vector conversion on the current network parameter set of each of the first structure BP networks to obtain a corresponding first structure parameter vector; setting a corresponding first model error for each of the first structure prediction models, and initializing each of the first model errors to 1; and initializing a first iteration counter to 1; Step 53, taking each of the first structural prediction models as the corresponding current prediction model; and traversing all the first data records of the first evaluation set; and taking the first data record currently traversed as the corresponding current data record during the traversal; and inputting the first training parameter list of the current data record into the current prediction model for prediction processing to obtain the corresponding first training prediction peak strain; and forming a corresponding first prediction label pair from the first training prediction peak strain and the first label peak strain of the current data record; and at the end of this traversal, bringing all the first prediction label pairs corresponding to the current prediction model into a preset first error function for error calculation to obtain the corresponding first error value, and resetting the first model error corresponding to the current prediction model to the corresponding first error value; Wherein, the first error function is an MSE error function; Step 54, taking the inverse of the first model error corresponding to each of the first structure prediction models as the corresponding first model fitness; summing up all the first model fitnesses obtained to obtain the corresponding first fitness sum; taking the ratio of each of the first model fitnesses to the first fitness sum as the corresponding first model selection probability; and sorting all the first structure prediction models in descending order of the first model selection probability to obtain the corresponding first structure prediction model sequence; Step 55, the first and second first structure prediction models of the first structure prediction model sequence are combined into a corresponding first model pair, the third and fourth first structure prediction models are combined into a corresponding first model pair, and so on, until the last two first structure prediction models are combined into a corresponding first model pair; and based on a preset individual crossover algorithm and a preset first crossover probability, vector data crossover processing is performed on the first structure parameter vectors of the two first structure prediction models of each first model pair to obtain two corresponding first structure crossover parameter vectors; The individual crossover algorithm is a type of chromosome crossover algorithm in the GA algorithm, and the chromosome crossover algorithm includes at least a single-point crossover algorithm, a multi-point crossover algorithm, and an arithmetic crossover algorithm; Step 56, performing vector data mutation processing on the first structural cross parameter vector of each of the first structural prediction models based on a preset individual mutation algorithm and a preset first mutation probability to obtain a corresponding first structural mutation parameter vector; The individual mutation algorithm is a type of chromosome mutation algorithm in the GA algorithm, and the chromosome mutation algorithm at least includes a basic position mutation algorithm, a uniform mutation algorithm, a non-uniform mutation algorithm, and a Gaussian approximate mutation algorithm; Step 57, identifying whether the first iteration counter exceeds the iteration number threshold; if so, going to step 58; if not, resetting the network parameter set of the first structural BP network of the corresponding first structural prediction model based on each of the first structural variation parameter vectors, adding 1 to the first iteration counter, and returning to step 53; Step 58: taking the first first structure prediction model in the latest first structure prediction model sequence as the corresponding base model, and performing parameter solidification on the network parameter set of the first structure BP network of the base model.
6. The fitting method of the peak strain expression of a concrete member according to claim 4, characterized in that: The further model parameter optimization training of the base model based on the first data set based on the gradient descent algorithm specifically includes: Step 61, randomly dividing the first data set into data subsets based on a preset first ratio to obtain a corresponding first training set and a second evaluation set; Wherein, both the first training set and the second evaluation set include a plurality of the first data records; the ratio of the total number of records in the first training set to the total number of records in the second evaluation set satisfies the first ratio; Step 62, extracting the first first data record of the first training set as the corresponding current training record; Step 63, inputting the first training parameter list of the current training record into the base model for prediction processing to obtain the corresponding second training prediction peak strain; Step 64, bringing the second training predicted peak strain and the first label peak strain of the current training record into a preset first model loss function to calculate and obtain a corresponding first loss value; Wherein, the first model loss function at least includes an L1 loss function and an L2 loss function; Step 65, identifying whether the first loss value satisfies a preset first loss value range; if the first loss value satisfies the first loss value range, identifying whether the current training record is the last first data record of the first training set, if so, turning to step 66, if not, extracting the next first data record of the first training set as the new current training record and returning to step 63; if the first loss value does not satisfy the first loss value range, performing a round of model parameter modulation on the base model in the direction of minimizing the first model loss function based on a preset first model parameter optimizer, and returning to step 63 at the end of this round of parameter modulation; Wherein, the first model parameter optimizer includes at least an SGD optimizer and an ADAM optimizer based on a gradient descent algorithm; Step 66, traverse all the first data records of the second evaluation set; and during the traversal, use the first data record currently traversed as the corresponding current evaluation record; and input the first training parameter list of the current evaluation record into the base model for prediction processing to obtain the corresponding third training predicted peak strain; and form a corresponding second predicted label pair with the third training predicted peak strain and the first label peak strain of the current evaluation record; Step 67, after traversing all the first data records of the second evaluation set, all the obtained second prediction label pairs are brought into a preset second error function to calculate corresponding second error values; The second error function at least includes a MAE error function, an MSE error function and a RMSE error function; Step 68, identifying whether the second error value is greater than a preset first error threshold; if so, returning to step 61; if not, performing a parameter solidification on the network parameter set of the first structure BP network of the base model and confirming that the model training of the base model is completed.
7. The fitting method of the peak strain expression of a concrete member according to claim 3, characterized in that: The step of constructing a data set for fitting operation based on the preset second data set and the first prediction model as a corresponding third data set specifically includes: All the second data records of the second data set are traversed; and during the traversal, the second data record currently traversed is used as the corresponding current data record; and the component design parameters of the nineteen types of simulations in the current data record are extracted to form a corresponding second training parameter list; and the second training parameter list is input into the first prediction model for prediction processing to obtain the corresponding first predicted peak strain, and the first predicted peak strain is used as the corresponding second label peak strain; and the stirrup characteristic value and the reinforcement characteristic value of the current data record are extracted as the corresponding first stirrup characteristic value and the first reinforcement characteristic value; and the obtained first stirrup characteristic value, the first reinforcement characteristic value and the second label peak strain form a corresponding third data record; and at the end of the traversal, all the obtained third data records form the corresponding third data set.
8. The fitting method of the peak strain expression of a concrete member according to claim 2, characterized in that: The optimal stirrup characteristic coefficient α and the reinforcement characteristic coefficient β of the peak strain expression are fitted based on the third data set to obtain the corresponding optimal stirrup characteristic coefficient α. * and the optimal reinforcement characteristic coefficient β * , specifically including: Step 81: Count the total number of the third data records of the third data set to obtain a corresponding total number M; and record each of the third data records as a corresponding record R. i , 1≤index i≤M; and each of the records R i The first stirrup characteristic value is recorded as the corresponding λ t,i , the first reinforcement characteristic value is recorded as the corresponding λ s,i , the second label peak strain is recorded as the corresponding ε * cc,i ; and setting a corresponding objective function F and a corresponding objective function constraint condition based on the peak strain expression; The objective function F is: ε cc,i To t,i , s,i Substituting into the peak strain expression ε cc =ε0(1+αλ t +βλ s ) calculated peak strain; The objective function constraints are: α>0,β>0, is cc,i Less than the corresponding probability; Step 82, based on the COBYLA algorithm, according to the objective function F and the corresponding objective function constraint condition, the optimal values of the stirrup characteristic coefficient α and the reinforcement characteristic coefficient β of the peak strain expression are fitted to obtain the corresponding optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * .
9. A device for executing the fitting method of the peak strain expression of a concrete member according to any one of claims 1 to 8, characterized in that: The device comprises: a model building and training module, an expression building module, a fitting operation data set preparation module, a fitting operation module and an expression finalization module; The model building and training module is used to build a first prediction model for predicting peak strain according to a design parameter list of concrete components; and to train the first prediction model based on a preset first data set; the design parameter list includes nineteen types of component design parameters; The expression building module is used to build a linear expression with the stirrup characteristic value and the reinforcement characteristic value as independent variables and the peak strain as the dependent variable as the corresponding peak strain expression; the peak strain expression is ε cc =ε0(1+αλ t +βλ s ), where ε cc is the peak strain, ε0 is the preset plain concrete peak strain constant, λ t is the stirrup eigenvalue, λ s is the reinforcement characteristic value, α is the stirrup characteristic coefficient, and β is the reinforcement characteristic coefficient; The fitting operation data set preparation module is used to construct a data set for fitting operation based on the preset second data set and the first prediction model as a corresponding third data set; The fitting operation module is used to perform a fitting operation on the optimal values of the stirrup characteristic coefficient α and the reinforcement characteristic coefficient β of the peak strain expression based on the third data set to obtain the corresponding optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * ; The expression finalization module is used to determine the optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * Set the fitting result of the peak strain expression; the fitting result is ε cc =ε0(1+α * λ t +β * λ s );The optimal stirrup characteristic coefficient α * and the optimal reinforcement characteristic coefficient β * A set of coefficient values is: α * =1.997, β * =0.
869.
10. An electronic device, characterized in that: include: memory, processors, and transceivers; The processor is used to couple with the memory, read and execute instructions in the memory, so as to implement the method according to any one of claims 1 to 8; The transceiver is coupled to the processor, and the processor controls the transceiver to send and receive messages.
11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a computer, the computer is enabled to execute the method according to any one of claims 1 to 8.
