Concrete mix proportion design method capable of meeting multi-target requirements

Abstract

The invention relates to the field of civil engineering and water conservancy, in particular to a concrete mix proportion design method capable of meeting multi-target requirements, which comprises the following steps: initially setting a concrete proportion according to engineering requirements and material characteristics to obtain a mix proportion meeting the working performance and strength of concrete; establishing a finite element stress simulation model of the structure according to the structure geometric model, the initial condition, the service environment and the construction progress; constructing a concrete mix proportion and maximum stress machine learning model; establishing a target function; obtaining the optimal mix proportion of concrete based on an optimization algorithm; according to the method, the concrete mix proportion meeting the multi-target requirements of the structure is obtained based on the structural stress and data driving, the mix proportion design is optimized on the premise of meeting the multi-target requirements of the structure, the method has universal adaptability, the production cost can be reduced to the maximum extent, and higher social comprehensive benefits can be obtained.

Description

Technical Field

[0001] The present invention relates to the field of civil engineering and water conservancy, and particularly relates to a method for designing concrete mix proportion to meet multi-objective requirements. Background Art

[0002] Ordinary concrete is generally an artificial stone formed by mixing cement, coarse aggregate (crushed stone or pebble), fine aggregate (sand), admixture and water in an appropriate proportion and hardening. Among them, sand and stone play a skeletal role in concrete, inhibiting the shrinkage of cement; cement and water form cement paste, which wraps around the surface of coarse and fine aggregates and fills the voids between aggregates; the cement paste plays a lubricating role before hardening, making the concrete mixture have good workability, and after hardening, it cements the aggregates together to form a whole.

[0003] The properties of concrete include the workability of the concrete mixture, concrete strength, deformation and durability, etc. Among them, the fluidity of concrete is generally represented by slump, and the greater the slump, the better the fluidity; strength is the main mechanical property of hardened concrete, reflecting the quantitative ability of concrete to resist loads; concrete strength includes compressive, tensile, shear, flexural, fracture and bond strength; the properties of concrete are directly related to the mix proportion of concrete constituent materials. At present, the design of concrete mix proportion is still limited to the material level, and experimental personnel need to design the mix proportion according to the material indexes proposed by the structural design. Therefore, a large number of tests are required, and the obtained mix proportion results are not necessarily the best. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for designing concrete mix proportion to meet multi-objective requirements, which optimizes the mix proportion design on the premise of meeting the concrete strength, fluidity, structural safety and economy, has universal adaptability, and can significantly reduce the experimental amount and reduce the project cost.

[0005] To achieve the above purpose, the present invention provides a method for designing concrete mix proportion to meet multi-objective requirements, including:

[0006] Initial-set the concrete mix proportion according to the project requirements and material characteristics to obtain a mix proportion that meets the workability and strength of the concrete;

[0007] Establish a finite element temperature stress simulation model of the structure according to the structural geometric model, initial conditions, service environment and construction progress;

[0008] Construct a machine learning model of concrete mix proportion and maximum stress;

[0009] Establish an objective function;

[0010] Based on the optimization algorithm, obtain the concrete mix proportion that meets the multi-objective requirements of the structure.

[0011] Among them, in the concrete mix ratio that meets the multi-objective requirements of the structure obtained based on the optimization algorithm, the algorithm uses an improved algorithm based on the gradient boosting decision tree.

[0012] Among them, in the concrete mix ratio that meets the multi-objective requirements of the structure obtained based on the optimization algorithm, the most basic unit of the algorithm is one of the regression tree or the decision tree.

[0013] Among them, the specific operation process of constructing the machine learning model of the concrete mix ratio and the maximum stress includes:

[0014] Initialize the particle swarm, including random positions and velocities;

[0015] Evaluate the fitness of each particle;

[0016] For each particle, compare its fitness value with the best position it has passed. If it is better, then use it as the current best position;

[0017] Adjust the particle velocities and positions until a preset end condition is reached.

[0018] A method for designing a concrete mix ratio that meets multi-objective requirements according to the present invention first preliminarily sets the concrete mix ratio according to the characteristics of on-site concrete materials and the structural characteristics to obtain the mechanical parameters of the concrete materials under different mix ratios, then constructs a finite element model to calculate the stresses of concrete structures with different mix ratios, and then determines the input parameters required by the present invention according to the service environment, structural characteristics, and construction progress, constructs the relationship between the input parameters and the structural stress based on machine learning. After that, a target function is constructed according to the requirements of concrete structure safety, material quality, construction efficiency, and project cost, and the optimal mix ratio of concrete is obtained by combining the optimization algorithm; the present invention obtains a concrete mix ratio design that meets the structural requirements based on structural stress and data-driven. Therefore, on the premise of calculating with the present invention to ensure that the concrete strength and workability meet the design requirements, a concrete mix ratio that meets the multi-objective requirements such as structural safety, construction efficiency, and project cost is obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0020] Figure 1 is a flowchart of a method for designing a concrete mix ratio that meets multi-objective requirements according to the present invention.

[0021] Figure 2It is the specific operation flow chart of the concrete mix ratio and maximum stress integration algorithm model of the present invention.

[0022] Figure 3 It is the training flow chart of a concrete mix ratio design method that meets multi-objective requirements of the present invention.

[0023] Figure 4 It is a partial model flow chart of a concrete mix ratio design method that meets multi-objective requirements of the present invention.

[0024] Figure 5 It is the schematic diagram of the calculation example of the present invention. Specific embodiments

[0025] The following details the embodiments of the present invention. The examples of the embodiments are shown in the drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions from beginning to end. The embodiments described below with reference to the drawings are exemplary and are intended to explain the present invention and should not be construed as limiting the present invention.

[0026] Please refer to Figures 1 to 5 , the present invention provides a design method for a concrete mix ratio design method that meets multi-objective requirements, including:

[0027] S1: Initially set the concrete mix ratio according to the engineering requirements and material characteristics to obtain a mix ratio that meets the workability and strength of the concrete;

[0028] S2: Establish a finite element stress simulation model of the structure according to the structural geometric model, initial conditions, service environment, and construction progress;

[0029] S3: Construct a machine learning model for the concrete mix ratio and the maximum stress;

[0030] S4: Establish an objective function;

[0031] S5: Based on the optimization algorithm, obtain a concrete mix ratio that meets the multi-objective requirements of the structure.

[0032] In the concrete mix ratio that meets the multi-objective requirements of the structure obtained based on the optimization algorithm, the algorithm uses an improved algorithm based on gradient boosting decision trees; the most basic unit of the algorithm is one of regression trees or decision trees.

[0033] For a concrete mix ratio design method that meets multi-objective requirements of the present invention, first, the data collected in steps S1 and S2 is randomly divided into a training set and a test set. Generally, a ratio of 80% (for training) - 20% (for testing) is used. Note that the data is usually normalized to avoid scale effects.

[0034] Secondly, the grid search method and k-fold cross-validation (CV) are used in the training stage to find the best hyperparameters. Key model parameters in the ensemble learning algorithm are selected for optimization. In general research, k = 10 is recommended.

[0035] Finally, the performance of the model needs to be evaluated through certain quantitative indicators. In this paper, the coefficient of determination R 2 is used as the indicator:

[0036]

[0037] where: T i is the true value, P i is the predicted value, and i = 1, 2..., m is the sample number is the sample mean.

[0038] Step S4 solves the optimal mix ratio according to the model constructed in Step S3;

[0039] The objective functions are respectively:

[0040] OPt(MPC = ∑c i q i 7 i＝1 + MPC0)

[0041] OPt(TS = f(mix))

[0042] where: c i , i = 1, 2.......7 are the unit prices of cement, blast furnace slag, fly ash, water, water reducer, coarse aggregate, and fine aggregate in turn. MPC0 is the average mixing and transportation cost per cubic meter of concrete, which usually decreases with the expansion of the concrete production scale. Here, it is taken as a fixed value MPC0 = 25. MPC is the production cost per cubic meter of concrete.

[0043] where TS is the peak stress of the concrete, and mix is the concrete mix ratio.

[0044] The following explains some models of the present invention

[0045] The improved algorithm based on the gradient boosting decision tree, its most basic unit is the regression tree, and the model is expressed as:

[0046]

[0047] where, is the prediction result of sample i after the t-th iteration, f t (x i ) is the model prediction result of the t-th regression tree, is the prediction result of the t = 1st regression tree. The objective function of this algorithm can be divided into two parts, namely the loss function and the positive part, and its objective function can be expressed by the following formula:

[0048]

[0049] Where is the loss function of the model, y i is the predicted value of the i-th sample of the model, is the regularization term in the function.

[0050] By splitting the regularization term into the first t - 1 terms and the t-th term, the loss function can be expressed as:

[0051]

[0052] Where is the model prediction of the previous t - 1 rounds, f t (x i ) is the new function, and constant is the constant term.

[0053] Perform a second-order Taylor expansion on the objective function. Define g i and h i as the first-order derivative and the second-order derivative of the loss function respectively, which can be expressed as:

[0054]

[0055] Furthermore, the following expression of the objective function can be obtained:

[0056]

[0057] The model formula is as follows:

[0058] v i = ωv i + c1 × rand() × (pbest i - x i ) + c2 × rand() × (gbest i - x i ) (7)

[0059] x i = x i + v1 (8)

[0060] Where: i = 1, 2, …, N, and N is the total number of particles;

[0061] v i : is the velocity of the particle;

[0062] x i: is the current position of the particle;

[0063] c1, c2: are learning factors;

[0064] rand(): is a random number between 0 and 1;

[0065] ω: is the inertia factor, and its value is non - negative;

[0066] The specific operation process of constructing the machine learning model of concrete mix ratio and maximum stress includes:

[0067] S11 Initialize the particle swarm, including random positions and velocities;

[0068] Initialize a group of particles (the group size is N), including random positions and velocities.

[0069] S12 Evaluate the fitness of each particle;

[0070] S13 For each particle, compare its fitness value with the best position it has passed. If it is better, then use it as the current best position;

[0071] For each particle, compare its fitness value with the best position pbest it has passed. If it is better, then use it as the current best position pbest.

[0072] S14 Adjust the particle velocities and positions until the preset end condition is reached;

[0073] Adjust the particle velocities and positions according to formulas (2) and (3). If the end condition is not reached, then go to step S12.

[0074] Calculation examples are Figure 5 as shown.

[0075] A method for designing concrete mix ratio to meet multi - objective requirements of the present invention first preliminarily sets the concrete mix ratio according to the characteristics of on - site concrete materials and structural characteristics to obtain the mechanical parameters of concrete materials under different mix ratios, then constructs a finite - element model to calculate the stresses of concrete structures with different mix ratios, and then determines the input parameters required by the present invention according to the service environment, structural characteristics, and construction progress. Based on machine learning, the relationship between the input parameters and structural stresses is constructed. After that, an objective function is constructed according to the requirements of concrete structure safety, material quality, construction efficiency, and project cost, and the optimal concrete mix ratio is obtained by combining with an intelligent optimization algorithm; the present invention obtains the concrete mix ratio design that meets the structural requirements based on structural stresses and data - driven. Therefore, on the premise of calculating to ensure that the concrete strength and workability meet the design requirements by using the present invention, a concrete mix ratio that meets the multi - objective requirements such as structural safety, construction efficiency, and project cost is obtained.

[0076] The above-disclosed is only a preferred embodiment of the present invention. Of course, the scope of the rights of the present invention cannot be limited thereby. Those of ordinary skill in the art can understand all or part of the processes of implementing the above embodiments, and the equivalent changes made according to the claims of the present invention still fall within the scope covered by the invention.

Claims

1. A method for designing the concrete mix proportion to meet multi-objective requirements, characterized in that Including: Initial set the concrete mix ratio according to engineering requirements and material properties to obtain a mix ratio that meets the workability and strength of the concrete; Establish a finite element stress simulation model of the structure based on the structural geometric model, initial conditions, service environment and construction progress; Construct a machine learning model of concrete mix ratio and maximum stress; Establish an objective function; Based on the optimization algorithm, obtain the concrete mix ratio that meets the multi-objective requirements of the structure.

2. A method for designing a concrete mix ratio that meets multi-objective requirements according to claim 1, characterized in that In the concrete mix ratio obtained based on the optimization algorithm to meet the multi-objective requirements of the structure, the algorithm uses an improved algorithm based on gradient boosting decision tree.

3. A method for designing a concrete mix ratio that meets multi-objective requirements according to claim 2, characterized in that In the concrete mix ratio obtained based on the optimization algorithm to meet the multi-objective requirements of the structure, the most basic unit of the algorithm is one of regression tree or decision tree.

4. A method for designing a concrete mix ratio that meets multi-objective requirements according to claim 3, characterized in that The specific operation process of constructing the machine learning model of concrete mix ratio and maximum stress includes: Initialize the particle swarm, including random positions and velocities; Evaluate the fitness of each particle; For each particle, compare its fitness value with its best position passed through. If it is better, then use it as the current best position; Adjust the particle velocities and positions until a preset end condition is reached.

Images