Method and system for preparing spiral tensile fiber concrete with high blast resistance

Abstract

The application provides a spiral tensile fiber concrete preparation method with high anti-explosion performance, belongs to the field of fiber reinforced concrete, and comprises the following steps: according to the fiber type and mechanical performance parameters, the diameter ratio and wrapping angle of the spiral tensile fiber are predicted based on deep learning, and the spiral tensile fiber is prepared; the spiral tensile fiber is embedded into concrete with different matrix strengths, the negative Poisson's ratio of the spiral tensile fiber is determined after curing, and the concrete matrix strength grade corresponding to the maximum negative Poisson's ratio is recorded; under the concrete matrix strength grade corresponding to the maximum negative Poisson's ratio, a plurality of groups of mixing ratios are generated, the mechanical performance prediction value is obtained based on deep learning according to the spiral tensile fiber characteristic parameters and the mixing ratios, the fitness of the mechanical performance prediction value and the target value is calculated, the matrix mixing ratio is searched based on a genetic algorithm, and the matrix mixing ratio is verified; and the concrete is prepared according to the matrix mixing ratio and the spiral tensile fiber; the preparation system is also provided; and the anti-explosion performance of the concrete is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fiber reinforced concrete, in particular to a spiral tensile fiber concrete preparation method and system with high blast resistance. BACKGROUND

[0002] Concrete is a kind of artificial stone made of cement as a cementitious material, mixed with water, aggregate (sand, gravel) and admixture in proportion, and then hardened by stirring, molding and curing. Concrete is widely used in construction engineering (house beam column, floor), infrastructure (bridge, road, tunnel), water conservancy engineering (dam, dike) and other fields due to its high compressive strength, good durability (service life up to 100 years), strong plasticity, economy and practicality. As a core material for modern engineering, concrete supports the structural safety and long-term service of various civil engineering with high cost performance and multi-functionality.

[0003] However, concrete also has the disadvantages of low tensile strength (only 1 / 10 of the compressive strength), internal microcrack defects and poor fracture toughness. In addition, when subjected to impact load, concrete is prone to blast crater on the blast surface and flying debris on the back blast surface, causing harm to the internal structure personnel. Incorporating fibers is an effective means to enhance the blast resistance of concrete. At present, the incorporation of fibers into concrete is mainly single fiber incorporation, and the incorporated fibers are generally steel fibers, polyvinyl alcohol fibers, polypropylene fibers and glass fibers. Single fiber incorporation can enhance the toughness of concrete, but the enhancement effect is still not ideal. In addition, the traditional fiber concrete mix design method often focuses on empirical formulas and test data, making it difficult to meet complex and individual application requirements. In recent years, the development of deep learning technology has provided a new way to solve this problem. For example, the Chinese invention patent application CN117912594A “High-performance concrete mix design and performance prediction method based on deep learning” applies deep learning algorithm to the research and development of ultra-high performance concrete. However, this method still has the problem of not ideal enhancement effect by incorporating single fiber into concrete. SUMMARY

[0004] The technical problem to be solved by the present application is how to improve the blast resistance of concrete.

[0005] The present application solves the above technical problems by the following technical solution: a spiral tensile fiber concrete preparation method with high blast resistance, the method comprising:

[0006] Based on the deep learning, the diameter ratio and wrapping angle of the spiral tensile fiber are predicted according to the fiber type and mechanical property parameters, and the spiral tensile fiber is prepared;

[0007] The spiral expansion fiber is embedded in concrete with different matrix strengths, the negative Poisson's ratio of the spiral expansion fiber is measured after curing, and the concrete matrix strength grade corresponding to the maximum negative Poisson's ratio is recorded;

[0008] Under the concrete matrix strength grade corresponding to the maximum negative Poisson's ratio, a plurality of mix proportions are generated, the mechanical property prediction value is obtained based on the spiral expansion fiber characteristic parameters and the mix proportions based on deep learning, the fitness of the mechanical property prediction value and the target value is calculated, the matrix mix proportion is searched based on the genetic algorithm, and the matrix mix proportion is verified;

[0009] The concrete is prepared according to the matrix mix proportion and the spiral expansion fiber.

[0010] The spiral expansion fiber negative Poisson's ratio effect maximization optimization is carried out based on deep learning, the optimal diameter ratio and wrapping angle are predicted from the fiber type and mechanical property parameters, so that the predicted fiber has a negative Poisson's ratio, the spiral expansion fiber with the best fiber type, diameter ratio and wrapping angle is designed, and the spiral expansion fiber is designed based on the principle of maximizing the negative Poisson's ratio effect of the spiral expansion fiber in the concrete. The maximum negative Poisson's ratio of the spiral expansion fiber with different geometric parameters in different concrete strength grades is designed, the best matching scheme of the spiral expansion fiber and the different strength concrete matrix is designed, a plurality of mix proportions are generated under the concrete matrix strength grade corresponding to the maximum negative Poisson's ratio, the mechanical property is predicted based on the spiral expansion fiber characteristic parameters and the mix proportions based on deep learning, the fitness of the mechanical property prediction value and the target value is calculated, the best matrix mix proportion is searched, and the matrix mix proportion is verified to ensure the rationality of the predicted mix proportion. Compared with single fiber incorporation, the concrete prepared according to the matrix mix proportion and the spiral expansion fiber has high toughness, the spiral expansion fiber has good energy absorption characteristics, and the generation and development of concrete cracks can be effectively inhibited by absorbing explosion load energy, so as to improve the anti-explosion performance of the concrete and ensure the safety of the personnel inside the concrete structure under impact load. It has a wide application prospect in civil, industrial and military fields.

[0011] Preferably, the process of predicting the diameter ratio and wrapping angle of the fiber based on deep learning according to the fiber type and mechanical property parameters comprises:

[0012] The fiber type and fiber mechanical property parameters of the spiral yarn are taken as input, the structure parameters of the spiral yarn are taken as output labels, a multi-layer fully connected neural network is trained, and a deep learning model is obtained, wherein the fiber mechanical property parameters include Poisson's ratio, elastic modulus and elongation at break, and the structure parameters of the spiral yarn include diameter ratio, wrapping angle and negative Poisson's ratio effect value;

[0013] Two kinds of fibers with large difference in elastic modulus are selected, the fiber types and mechanical property parameters are input into the deep learning model, and the diameter ratio and wrapping angle of the spiral auxetic fiber are output.

[0014] Preferably, the process for preparing the spiral auxetic fiber comprises:

[0015] Two kinds of fibers with large difference in elastic modulus are selected, one type of fiber is the core fiber, and the other type of fiber is the wrapping fiber, the fiber types include polypropylene fiber, polyvinyl alcohol fiber, flax fiber, glass fiber, steel fiber;

[0016] The core fiber is wound on the fiber conveying device, so that the core fiber is in a straightened state, and the fiber conveying device uniformly conveys the core fiber;

[0017] The wrapping fiber is fixed on the winding device, and the conveying speed of the fiber conveying device and the rotation speed of the winding device are adjusted so that the wrapping fiber is wrapped around the core fiber at the predicted wrapping angle, thereby obtaining the spiral auxetic fiber;

[0018] The prepared spiral auxetic fiber is collected using a take-up device, the spiral auxetic fiber is cut, and universal glue is added at both ends of the cut spiral auxetic fiber.

[0019] Preferably, the different matrix strength concretes have different raw material proportions in the preparation process and the same preparation method, the spiral auxetic fiber is embedded in the different matrix strength concretes, and the process for measuring the negative Poisson's ratio of the spiral auxetic fiber after curing comprises:

[0020] The spiral auxetic fiber is immersed in a 3% silane coupling agent ethanol solution for 15 min, and then dried at 60°C for standby;

[0021] The coarse aggregate and fine aggregate are added to the mixer mixing container and stirred for 1 min;

[0022] The cement, slag powder, fly ash and silica fume are mixed and stirred for 2 min, then poured into the mixer mixing container and stirred for another 2 min;

[0023] The water and polycarboxylic acid water reducing agent are mixed and stirred for 30 s, then poured into the mixer mixing container and stirred for 3 min, thereby obtaining the target matrix strength concrete;

[0024] The target matrix strength concrete is poured into the mold, and the spiral auxetic fiber with a length of 30 mm is vertically embedded in the concrete matrix with an embedding depth of 20 mm, and cured for 28 days;

[0025] The spiral auxetic fiber is subjected to a pull-out load, the spiral auxetic fiber inside the concrete is scanned using in-situ CT, and the negative Poisson's ratio of the spiral auxetic fiber inside the concrete during the load action is recorded.

[0026] Preferably, the cement is PII·52.5 grade ordinary portland cement, the fine aggregate is quartz sand with a particle size of 0.15-0.6 mm, the fly ash is I grade fly ash, the silica fume has a SiO2 content of ≥90%, a chloride ion content of ≤0.1% in the silica fume, a sulfide content of ≤2% in the silica fume, a fineness that meets a specific surface area of ≥15.000 m² / kg, a water content of ≤3% in the silica fume, and the polycarboxylate superplasticizer meets a water-reducing rate of ≥35%, an air content of ≤6%, and an alkali content of ≤10%.

[0027] Preferably, based on the characteristic parameters of the spiral tensile fiber and the mixing proportion, the process of obtaining the mechanical property prediction value based on deep learning includes:

[0028] A bidirectional LSTM network is trained with the characteristic parameters of the spiral tensile fiber and the mixing proportion as inputs and the mechanical properties of the fiber concrete as output labels to obtain a performance prediction network, wherein the characteristic parameters of the spiral tensile fiber include fiber type, diameter, length, elastic modulus, and Poisson's ratio, and the mixing proportion includes water-cement ratio, sand ratio, superplasticizer content, and spiral tensile fiber volume content.

[0029] The characteristic parameters of the spiral tensile fiber and the mixing proportion are input into the performance prediction network, wherein the water-cement ratio in the mixing proportion is less than or equal to 0.3, the sand ratio ranges from 30% to 40%, and the mechanical property prediction value is output, including splitting tensile strength, toughness index, and strain energy.

[0030] Preferably, the fitness of the mechanical property prediction value and the target value is:

[0031]

[0032] wherein, , , are the target splitting tensile strength, the target toughness index, and the target strain energy, respectively, , , are the predicted splitting tensile strength, the predicted toughness index, and the predicted strain energy, respectively, and is a cost coefficient.

[0033] Preferably, the process of searching for the matrix mixing proportion based on a genetic algorithm includes:

[0034] Under the concrete matrix strength grade corresponding to the maximum negative Poisson's ratio value, a plurality of mixing proportions are randomly generated to form an initial population, and each mixing proportion is taken as an individual.

[0035] ​Calculate the fitness of each individual, screen the individuals according to the fitness, perform cross operation on the selected individuals, generate new individuals, perform mutation operation on the individuals after cross operation, produce the next generation population, and perform the cycle until the iteration is set for a certain number of times or the fitness change is less than 1e -5 , to obtain the matrix matching ratio.

[0036] Preferably, the matrix matching ratio includes water-cement ratio , sand ratio , spiral dilatancy fiber volume content , negative Poisson's ratio of spiral dilatancy fiber , water reducing agent content , and the constraint condition of the matrix matching ratio is verified as: , , , , .

[0037] The present application also provides a spiral dilatancy fiber concrete preparation system with high anti-explosion performance, which comprises:

[0038] A spiral dilatancy fiber design module is used to predict the diameter ratio and wrapping angle of the spiral dilatancy fiber based on deep learning according to the fiber type and mechanical property parameters, and to prepare the spiral dilatancy fiber.

[0039] A matrix strength determination module based on negative Poisson's ratio effect maximization is used to bury the spiral dilatancy fiber in concrete with different matrix strengths, to measure the negative Poisson's ratio of the spiral dilatancy fiber after curing, and to record the concrete matrix strength grade corresponding to the maximum negative Poisson's ratio.

[0040] A matrix matching ratio design module is used to generate multiple matching ratios under the concrete matrix strength grade corresponding to the maximum negative Poisson's ratio, to obtain the mechanical property prediction value based on deep learning according to the spiral dilatancy fiber characteristic parameters and the matching ratio, to calculate the fitness of the mechanical property prediction value and the target value, to search for the matrix matching ratio based on the genetic algorithm, and to verify the matrix matching ratio.

[0041] A concrete preparation module is used to prepare the concrete according to the matrix matching ratio and the spiral dilatancy fiber.

[0042] The spiral dilatancy fiber for concrete designed based on deep learning in the present application can meet the diversification and complication requirements of concrete, the concrete matching ratio design based on deep learning can reduce the waste of raw materials caused by concrete trial matching, greatly improve the fiber concrete matching ratio design efficiency, reduce the resource waste caused by trial matching, help low-carbon emission reduction, realize reasonable use of resources and sustainable development. BRIEF DESCRIPTION OF DRAWINGS

[0043] Figure 1 Flow chart of the method for preparing the spiral tensile fiber concrete with high anti-blast performance provided in Embodiment 1 of the present application;

[0044] Figure 2 Principle diagram of the spiral tensile fiber design suitable for concrete based on deep learning in the method for preparing the spiral tensile fiber concrete with high anti-blast performance provided in Embodiment 1 of the present application;

[0045] Figure 3 Photo of the spiral tensile fiber synthesized by the polypropylene fiber (d=0.9mm) and the steel fiber (d=0.3mm) at a wrapping angle of 15° in the method for preparing the spiral tensile fiber concrete with high anti-blast performance provided in Embodiment 1 of the present application, and the negative Poisson's ratio can reach -3.23 in the natural state;

[0046] Figure 4 Photo of the spiral tensile fiber synthesized by the polypropylene fiber (d=0.9mm) and the steel fiber (d=0.3mm) at a wrapping angle of 20° in the method for preparing the spiral tensile fiber concrete with high anti-blast performance provided in Embodiment 1 of the present application, and the negative Poisson's ratio can reach -5.2 in the natural state;

[0047] Figure 5 Photo of the spiral tensile fiber synthesized by the polypropylene fiber (d=1.2mm) and the steel fiber (d=0.3mm) at a wrapping angle of 10° in the method for preparing the spiral tensile fiber concrete with high anti-blast performance provided in Embodiment 1 of the present application, and the negative Poisson's ratio can reach -3.61 in the natural state;

[0048] Figure 6 Principle diagram of the construction of the double prediction model in the method for preparing the spiral tensile fiber concrete with high anti-blast performance provided in Embodiment 1 of the present application;

[0049] Figure 7 Reverse performance prediction network code in the method for preparing the spiral tensile fiber concrete with high anti-blast performance provided in Embodiment 1 of the present application;

[0050] Figure 8 Realization code of the genetic algorithm optimization of the mix proportion in the method for preparing the spiral tensile fiber concrete with high anti-blast performance provided in Embodiment 1 of the present application;

[0051] Figure 9 Code of the constraint verification module in the method for preparing the spiral tensile fiber concrete with high anti-blast performance provided in Embodiment 1 of the present application;

[0052] Figure 10 Schematic diagram of the spiral tensile fiber concrete preparation system with high anti-blast performance provided in Embodiment 2 of the present application. DETAILED DESCRIPTION

[0053] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions of the present application are described below in detail with reference to specific embodiments and drawings. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0054] Embodiment 1

[0055] As shown in the embodiment, a spiral tensile fiber concrete preparation method with high anti-blast performance is provided, comprising the following steps: Figure 1 Step 1, according to the fiber type and mechanical property parameters, the diameter ratio and the wrap angle of the spiral tensile fiber are predicted based on deep learning, and the spiral tensile fiber is prepared.

[0056] Referring to

[0057] , according to the fiber type and mechanical property parameters, the process of predicting the diameter ratio and the wrap angle of the fiber based on deep learning includes: Figure 2 Step 1.1.1, collect the existing negative Poisson's ratio spiral yarn related parameters and the basic mechanical properties of the fibers commonly mixed into concrete, the fiber types include: polyvinyl alcohol fiber (PVA), polypropylene fiber (PP), flax fiber, glass fiber, steel fiber, and the basic mechanical property parameters of the fiber include: Poisson's ratio, Young's modulus, elongation at break, the existing negative Poisson's ratio spiral yarn structure parameters: inner / outer diameter ratio, wrap angle, negative Poisson's ratio effect value (label and numerical judgment).

[0058] Take the fiber type and the fiber mechanical property parameters of the spiral yarn as the input, and take the structure parameters of the spiral yarn as the output label, train the multi-layer fully connected neural network (MLP Regression), which includes linear layers, ReLU activation functions, linear layers, ReLU activation functions and linear layers connected in turn. When the loss function is minimum, the deep learning model is obtained, wherein the fiber mechanical property parameters include Poisson's ratio, Young's modulus and elongation at break, and the structure parameters of the spiral yarn include diameter ratio, wrap angle and negative Poisson's ratio effect value; the loss function adopts mean square error

[0059] ​, for regression of diameter ratio and wrapping angle, and a categorical loss if distinguishing between negative and positive Poisson's ratio, where the mean squared error The calculation method is as follows:

[0060]

[0061] wherein, is the true value of the th sample, is the predicted value of the th sample, is the total number of samples.

[0062] Step 1.1.2, select two types of fibers with large difference in elastic modulus, input the fiber type and mechanical property parameters into the deep learning model, and output the diameter ratio and wrapping angle of the spiral dilational fiber.

[0063] The present application is based on deep learning to design spiral dilational fibers suitable for concrete. By constructing a deep learning model, five types of fibers are subjected to Embedding or One-Hot, and the Poisson's ratio, elastic modulus, elongation at break and target negative Poisson's ratio range of each type of fiber are combined to construct input features, with diameter ratio and wrapping angle as output labels. The diameter ratio is a non-continuous value, and the specific value is 2, 3, 4. The wrapping angle is also a non-continuous value, and the range is 10°, 15°, 20°, 25°. Based on the existing negative Poisson's ratio spiral yarn related parameters and the basic mechanical property parameters of the fibers commonly incorporated into concrete, the deep learning model is trained to obtain a trained deep learning model, which is used to predict the optimal diameter ratio and wrapping angle according to the fiber type and its performance parameters, so that the predicted yarn has a negative Poisson's ratio. The spiral dilational fiber with negative Poisson's ratio effect is a spiral structure composed of two types of fiber filaments with large difference in elastic modulus, for example, polypropylene fiber and steel fiber.

[0064] In order to verify the rationality of the spiral dilational fiber suitable for concrete designed based on deep learning, the present application compares the prediction results with the test results:

[0065] Two types of fibers, polypropylene fiber and steel fiber, are selected, and the mechanical property parameters of polypropylene fiber and steel fiber, including Poisson's ratio, elastic modulus and elongation at break, are obtained. The fiber type and mechanical property parameters are input into the trained deep learning model, and the diameter ratio and wrapping angle of the spiral dilational fiber are output, and the maximum negative Poisson's ratio value is predicted. Referring to Table 1, the predicted value of the maximum negative Poisson's ratio corresponding to the diameter ratio 1:3 and the wrapping angle 15° of the polypropylene fiber and the steel fiber is -2.59. According to the diameter ratio 1:3 and the wrapping angle 15°, the polypropylene fiber is used as the core fiber, and the steel fiber is used as the wrapping fiber, to prepare a spiral dilational fiber, Figure 3The actual preparation effect diagram of the spiral auxetic fiber is shown in FIG. 1. The maximum negative Poisson's ratio of the prepared spiral auxetic fiber is -2.68. The predicted value of the maximum negative Poisson's ratio corresponding to the diameter ratio of 1:3 and the wrapping angle of 20° of the polypropylene fiber and the steel fiber is -3.10. According to the diameter ratio of 1:3 and the wrapping angle of 20°, the polypropylene fiber is used as the core fiber, and the steel fiber is used as the wrapping fiber to prepare the spiral auxetic fiber, Figure 4 The actual preparation effect diagram of the spiral auxetic fiber is shown in FIG. 1. The maximum negative Poisson's ratio of the prepared spiral auxetic fiber is -2.68. The predicted value of the maximum negative Poisson's ratio corresponding to the diameter ratio of 1:3 and the wrapping angle of 20° of the polypropylene fiber and the steel fiber is -3.10. According to the diameter ratio of 1:3 and the wrapping angle of 20°, the polypropylene fiber is used as the core fiber, and the steel fiber is used as the wrapping fiber to prepare the spiral auxetic fiber, Figure 5 The actual preparation effect diagram of the spiral auxetic fiber is shown in FIG. 1. The maximum negative Poisson's ratio of the prepared spiral auxetic fiber is -2.68. The predicted value of the maximum negative Poisson's ratio corresponding to the diameter ratio of 1:3 and the wrapping angle of 20° of the polypropylene fiber and the steel fiber is -3.10. According to the diameter ratio of 1:3 and the wrapping angle of 20°, the polypropylene fiber is used as the core fiber, and the steel fiber is used as the wrapping fiber to prepare the spiral auxetic fiber,

[0066] Table 1 Comparison of predicted results and test results of spiral auxetic fibers suitable for concrete based on deep learning

[0067]

[0068] The process of preparing the spiral auxetic fiber includes:

[0069] Step 1.2.1, selecting two types of fibers with large difference in elastic modulus, one type of fiber as core fiber and the other type of fiber as wrapping fiber, the fiber types including polypropylene fiber, polyvinyl alcohol fiber, flax fiber, glass fiber and steel fiber;

[0070] Step 1.2.2, winding the core fiber on the fiber conveying device so that the core fiber is in a straightened state, and the fiber conveying device uniformly conveying the core fiber;

[0071] Step 1.2.3, fixing the wrapping fiber on the winding device, adjusting the fiber conveying speed of the fiber conveying device and the rotation speed of the winding device, so that the wrapping fiber is wound on the core fiber at the predicted wrapping angle to obtain the spiral auxetic fiber;

[0072] Step 1.2.4, collecting the prepared spiral auxetic fiber using the take-up device, cutting the spiral auxetic fiber, and dropping universal glue at both ends of the cut spiral auxetic fiber.

[0073] Step 2, embedding the spiral auxetic fiber into concrete with different matrix strengths, and measuring the negative Poisson's ratio of the spiral auxetic fiber after curing, and recording the concrete matrix strength grade corresponding to the maximum negative Poisson's ratio.

[0074] The raw material proportions are different in the preparation process, and the preparation methods are the same, the raw materials include cement, coarse aggregate, fine aggregate, cement, slag powder, fly ash, silica fume, water and polycarboxylic acid water reducing agent, wherein the cement is PII·52.5 grade ordinary portland cement, the fine aggregate is quartz sand with a particle size of 0.15-0.6 mm, the fly ash is I grade fly ash, the SiO2 content in the silica fume is ≥90%, the chloride ion in the silica fume is ≤0.1%, the sulfide in the silica fume is ≤2%, the fineness of the silica fume meets the specific surface area ≥15.000 m² / kg, the water content in the silica fume is ≤3%, the polycarboxylic acid water reducing agent meets: water reducing rate ≥35%, air content ≤6%, alkali content ≤10%. The water is tap water or drinking water, which meets the requirements of the Concrete Water Standard (JGJ63-2006).

[0075] The process of embedding the spiral expansion fiber into the concrete with different matrix strengths and measuring the negative Poisson's ratio of the spiral expansion fiber after curing includes:

[0076] Step 2.1, immerse the spiral expansion fiber in a 3% silane coupling agent ethanol solution for 15 min, and dry at 60℃ for standby;

[0077] Step 2.2, add the coarse aggregate and fine aggregate into the mixer mixing container and stir for 1 min;

[0078] Step 2.3, mix and stir the cement, slag powder, fly ash and silica fume for 2 min, then pour into the mixer mixing container and stir for another 2 min;

[0079] Step 2.4, mix and stir the water and polycarboxylic acid water reducing agent for 30 s, then pour into the mixer mixing container and stir for 3 min, to obtain the concrete with the target matrix strength;

[0080] Step 2.5, pour the concrete with the target matrix strength into the mold, and vertically embed the spiral expansion fiber with a length of 30 mm into the concrete matrix, with an embedding depth of 20 mm, and cure for 28 days;

[0081] Step 2.6, apply the pull-out load to the spiral expansion fiber, use in-situ CT to scan the spiral expansion fiber inside the concrete, and record the negative Poisson's ratio of the spiral expansion fiber inside the concrete during the load action.

[0082] The present application applies a load to the prepared spiral expansion fiber concrete specimen, uses in-situ CT to scan the internal fiber, and counts the change of the fiber Poisson's ratio during the load application, based on the maximum principle of the negative Poisson's ratio effect of the fiber, counts the concrete with different matrix strengths, and corresponds to the optimal fiber geometric parameters.

[0083] Step 3. Generating multiple groups of mix proportions at the concrete matrix strength grade corresponding to the maximum negative Poisson's ratio value, calculating the fitness of the predicted mechanical properties and the target values based on the deep learning of the mechanical properties according to the spiral dilatation fiber characteristic parameters and the mix proportions, searching for the matrix mix proportion based on the genetic algorithm, and verifying the matrix mix proportion.

[0084] The process of obtaining the predicted mechanical properties based on deep learning according to the spiral dilatation fiber characteristic parameters and the mix proportions comprises:

[0085] Step 3.1.1. Based on literature research and experimental data, 300 groups of original data of fiber concrete parameters and properties are obtained, which should include fiber characteristics: fiber type, diameter and length (mm), elastic modulus (MPa) and Poisson's ratio, mix proportion parameters: water-binder ratio, sand ratio, water reducing agent content, and fiber volume content. In order to quantify the toughening effect of spiral dilatation fiber on concrete, set up spiral dilatation fiber group and control group of steel fiber for comparison, except that the type of fiber is different, the rest of the variables are the same. The indicators for measuring the toughening effect of fiber concrete are: splitting tensile strength (MPa), toughness index, and strain energy (N·m). The original data is cleaned and normalized, all numerical features are scaled to [0, 1], the fiber type is converted to One-hot vector, and ±5% random disturbance is applied to the sample to realize data enhancement.

[0086] A double prediction network model is constructed, as shown in Figure 6 The model includes a performance prediction network and a mix proportion generation network. The performance prediction network is a forward network, including a bidirectional LSTM network and a fully connected layer connected in sequence, for predicting mechanical properties according to input fiber characteristics and mix proportions. The mix proportion generation network is a reverse network, including a convolutional layer and a fully connected layer connected in sequence, for outputting fiber parameters and matrix mix proportions according to input target mechanical properties, including splitting tensile strength, toughness index, and strain energy. The loss function of the reverse network is: fitness = performance prediction error (sum of errors of splitting tensile strength, toughness index, and strain energy) + fiber cost penalty term (to control economy). The reverse network code is shown in Figure 7 .

[0087] The spiral dilatation fiber characteristic parameters and the mix proportions are input, and the mechanical properties of the fiber concrete are output as labels. The bidirectional LSTM network is trained, and when the loss function is minimized, the performance prediction network is obtained. The spiral dilatation fiber characteristic parameters include fiber type, diameter, length, elastic modulus, and Poisson's ratio, and the mix proportions include water-binder ratio, sand ratio, water reducing agent content, and spiral dilatation fiber volume content. The loss function is the weighted average of splitting tensile strength, toughness index, and strain energy loss, and the calculation method is:

[0088] + +

[0089] wherein, 、 、 are the true value of the split tensile strength, the true value of the toughness index, the true value of the strain energy of the first sample, respectively, are the predicted value of the split tensile strength, the predicted value of the toughness index, the predicted value of the strain energy of the first sample, respectively. are the predicted value of the split tensile strength, the predicted value of the toughness index, the predicted value of the strain energy of the first sample, respectively. Step 3.1.2, input the helical auxetic fiber characteristic parameters and the mix proportion into the performance prediction network, wherein the water-cement ratio in the mix proportion is less than or equal to 0.3, and the sand ratio ranges from 30% to 40%, and output the mechanical performance prediction value, the mechanical performance including the split tensile strength, the toughness index and the strain energy.

[0090] The fitness of the mechanical performance prediction value and the target value is:

[0091]

[0092]

[0093] wherein, 、 、 are the target split tensile strength, the target toughness index, the target strain energy, i.e. the target value of each mechanical performance, respectively, 、 、 are the predicted split tensile strength, the predicted toughness index, the predicted strain energy, i.e. the predicted value of each mechanical performance, respectively, is the cost coefficient.

[0094] The process of searching for the matrix mix proportion based on the genetic algorithm includes:

[0095] Step 3.2.1, randomly generate multiple groups of mix proportions under the strength grade of the concrete matrix corresponding to the maximum value of the negative Poisson's ratio, in this embodiment, 100 groups of mix proportions are randomly generated (satisfying the constraints: water-cement ratio ≤ 0.3, sand ratio 30-40%), constituting an initial population, and each group of mix proportion is taken as an individual;

[0096] ​​​​Step 3.2.2, calculate the fitness of each individual, screen the individuals according to the fitness, perform crossover operation on the selected individuals, two-point crossover, probability 0.8, generate new individuals, perform mutation operation on the individuals after the crossover operation, add Gaussian disturbance, probability 0.2, standard deviation 0.05, produce the next generation population, loop execution until the iteration reaches 50 generations (or 100 generations can also be set according to the needs) or the fitness change is less than 1e -5 , and the base matrix mixing ratio is obtained. The implementation code of the genetic algorithm optimization mixing ratio is shown in Figure 8 .

[0097] Based on the provisions of “Technical Specification for Application of Fiber Reinforced Concrete” (JGJ / T221-2010) and “Mixing Ratio Design Specification for Ordinary Concrete” (JGJ55-2019), as well as the basic requirements of concrete mixing ratio and the research results of related literature. To avoid extreme unreasonable results, ensure that the water-cement ratio, sand ratio, spiral tensile fiber volume content, and water reducing agent dosage must be within the reasonable range determined by the specification or experiment, and ensure that the spiral tensile fiber can exert the maximum negative Poisson's ratio effect in the concrete matrix. The range of factors is shown in Table 2.

[0098] Table 2: Constraint conditions for multi-objective optimization design of spiral tensile fiber reinforced concrete

[0099]

[0100] As shown in Table 2, the base matrix mixing ratio includes water-cement ratio , sand ratio , spiral tensile fiber volume content , spiral tensile fiber negative Poisson's ratio value , and water reducing agent dosage . The constraint conditions for verifying the base matrix mixing ratio are: , , , , . The implementation code of the constraint verification module is shown in the attached Figure 9 .

[0101] Verification and application effect: After testing set verification, the performance prediction network R² reached 0.92 (splitting tensile strength), 0.89 (toughness index), and 0.95 (strain energy). By inputting the mechanical performance parameters of spiral tensile fiber and its mechanical performance in different strength matrices, based on the principle of maximizing the negative Poisson's ratio effect of spiral tensile fiber, the optimal solution set is obtained. This optimal solution balances multiple performance targets and is considered the solution closest to the ideal solution among all considered factors.

[0102] Table 3 shows the optimal solution set given by the deep learning model, all of which successfully meet the performance requirements and are considered to have the best matching performance of the geometric parameters of the spiral dilatancy fiber and the concrete matrix, and can maximize the toughening effect of the fiber.

[0103] Table 3 Optimal solution set given by the deep learning model

[0104]

[0105] Step 4, preparing the concrete according to the matrix mix proportion and the spiral dilatancy fiber.

[0106] The present application is based on deep learning to maximize the negative Poisson's ratio effect of spiral dilatancy fiber, and the optimal diameter ratio and wrapping angle are predicted from the fiber type and mechanical property parameters, so that the predicted fiber has a negative Poisson's ratio, and the spiral dilatancy fiber with the best fiber type, diameter ratio and wrapping angle is designed. Based on the principle of maximizing the negative Poisson's ratio effect of spiral dilatancy fiber in concrete, the maximum negative Poisson's ratio of spiral dilatancy fiber with different geometric parameters in different concrete strength grades is calculated, the best matching scheme of spiral dilatancy fiber and different strength concrete matrix is designed, and multiple groups of mix proportions are generated at the concrete matrix strength grade corresponding to the maximum negative Poisson's ratio. The mechanical properties are predicted from the spiral dilatancy fiber characteristic parameters and mix proportions based on deep learning, the fitness of the predicted value and the target value of the mechanical properties is calculated, the best matrix mix proportion is searched, and the matrix mix proportion is verified to ensure the rationality of the predicted mix proportion.

[0107] Compared with single fiber incorporation, the concrete prepared according to the matrix mix proportion and the spiral dilatancy fiber has high toughness, the spiral dilatancy fiber has good energy absorption characteristics, and can effectively inhibit the generation and development of concrete cracks by absorbing explosion load energy, thereby improving the anti-explosion performance of concrete and ensuring the safety of personnel inside the concrete structure under impact load. It has a wide application prospect in civil, industrial and military fields.

[0108] To show that the spiral dilatancy fiber has a significant toughening effect on concrete, and based on the deep learning derived optimal mix proportion solution set of spiral dilatancy fiber reinforced concrete, the spiral dilatancy fiber reinforced concrete of the above scheme is prepared according to the Technical Application Regulation of Fiber Reinforced Concrete (JGJ / T 221-2010), and the splitting tensile strength, toughness index and strain energy of the spiral dilatancy fiber reinforced concrete of each scheme are measured. In addition, the spiral dilatancy fiber of the above scheme is replaced by steel fiber with the same length-diameter ratio and the same volume fraction, and the steel fiber control group is prepared according to the Technical Application Regulation of Fiber Reinforced Concrete (JGJ / T 221-2010), and the splitting tensile strength, toughness index and strain energy of each control group are measured. The splitting tensile strength, toughness index and strain energy of the spiral dilatancy fiber reinforced concrete prepared in each scheme, and the splitting tensile strength, toughness index and strain energy of the control group are shown in Table 4.

[0109] Table 4 Mechanical property test results of each group

[0110]

[0111] As shown in Table 4, in group No. 1, compared with the control group of steel fiber, the spiral dilatancy fiber with the same mix proportion makes the splitting tensile strength, toughness index and strain energy of the concrete increase by 28.4%, 24.2% and 27.0% respectively; in group No. 2, compared with the control group of steel fiber, the spiral dilatancy fiber with the same mix proportion makes the splitting tensile strength, toughness index and strain energy of the concrete increase by 28.7%, 25.7% and 23.0% respectively; in group No. 3, compared with the control group of steel fiber, the spiral dilatancy fiber with the same mix proportion makes the splitting tensile strength, toughness index and strain energy of the concrete increase by 19.5%, 38.2% and 21.3% respectively, and in addition, compared with the orthogonal experiment method, the time cost is reduced by 90%, and the fiber utilization rate is increased by 15-30%.

[0112] In summary, based on the principle of maximizing the negative Poisson's ratio effect of spiral dilatancy fiber in concrete and the deep learning design, the spiral dilatancy fiber reinforced concrete has high toughness, which can greatly improve the impact load performance of concrete. Compared with the existing single fiber mixed into concrete, the present application can enhance the toughening effect of the fiber by changing the geometric structure of the two fibers mixed into the concrete. The present application combines deep learning method with spiral dilatancy fiber reinforced concrete material design, greatly improves the efficiency of fiber reinforced concrete mix proportion design, reduces the resource waste caused by trial production, helps low-carbon emission reduction, and realizes the rational use of resources. Compared with the traditional method, the preparation method of the high blast-resistant performance concrete of the present application not only has higher precision and higher efficiency, but also has stronger generalization ability.

[0113] Example 2

[0114] Reference Figure 10The embodiment provides a spiral tensile fiber concrete preparation system with high anti-blast performance, and comprises the following modules:

[0115] A spiral tensile fiber design module is used for predicting a diameter ratio and a wrapping angle of the spiral tensile fiber based on deep learning according to fiber types and mechanical property parameters, and preparing the spiral tensile fiber.

[0116] The process of predicting the diameter ratio and the wrapping angle of the fiber based on deep learning according to the fiber types and the mechanical property parameters comprises the following steps:

[0117] A multi-layer fully connected neural network is trained by taking the fiber types and the fiber mechanical property parameters of the spiral yarn as inputs and taking the structural parameters of the spiral yarn as output labels, to obtain a deep learning model, wherein the fiber mechanical property parameters comprise a Poisson's ratio, an elastic modulus and a breaking elongation, and the structural parameters of the spiral yarn comprise the diameter ratio, the wrapping angle and a negative Poisson's ratio effect value;

[0118] Two fibers with large differences in elastic modulus are selected, and the fiber types and the mechanical property parameters are input into the deep learning model to output the diameter ratio and the wrapping angle of the spiral tensile fiber.

[0119] The process of preparing the spiral tensile fiber comprises the following steps:

[0120] Two fibers with large differences in elastic modulus are selected, one type of fiber is a core fiber, and the other type of fiber is a wrapping fiber, and the fiber types comprise polypropylene fiber, polyvinyl alcohol fiber, flax fiber, glass fiber and steel fiber.

[0121] The core fiber is wound on a fiber conveying device, so that the core fiber is in a straightened state, and the fiber conveying device uniformly conveys the core fiber.

[0122] The wrapping fiber is fixed on a winding device, the conveying speed of the fiber conveying device is adjusted to be the same as the rotating speed of the winding device, so that the wrapping fiber is wound on the core fiber at the predicted wrapping angle, and the spiral tensile fiber is obtained.

[0123] The prepared spiral tensile fiber is collected by using a take-up device, the spiral tensile fiber is cut, and universal glue is added at both ends of the cut spiral tensile fiber.

[0124] A matrix strength determination module based on negative Poisson's ratio effect maximization is used for embedding the spiral tensile fiber in concrete with different matrix strengths, measuring the negative Poisson's ratio value of the spiral tensile fiber after curing, and recording the concrete matrix strength grade corresponding to the maximum negative Poisson's ratio value.

[0125] Wherein, the different base strength of concrete in the preparation process of raw material ratio, preparation method is same, will spiral dilatancy fiber buried in different base strength of concrete, after curing to measure the negative poisson ratio of spiral dilatancy fiber value process includes:

[0126] The spiral dilatancy fiber is immersed in 3% silane coupling agent ethanol solution for 15 min, and dried at 60°C for standby;

[0127] The coarse aggregate and fine aggregate are added into the mixer mixing container and stirred for 1 min;

[0128] The cement, slag powder, fly ash and silica fume are mixed and stirred for 2 min, then poured into the mixer mixing container and stirred for 2 min;

[0129] The water and polycarboxylic acid water reducing agent are mixed and stirred for 30 s, then poured into the mixer mixing container and stirred for 3 min to obtain the target base strength of concrete;

[0130] The target base strength of concrete is poured into the mold, and the spiral dilatancy fiber with a length of 30 mm is vertically buried in the concrete matrix with a burial depth of 20 mm, and cured for 28 days;

[0131] The spiral dilatancy fiber is subjected to pull-out load, and the spiral dilatancy fiber inside the concrete is scanned by in-situ CT, and the negative poisson ratio of the spiral dilatancy fiber inside the concrete during the load action is recorded.

[0132] The cement is PII·52.5 grade ordinary portland cement, the fine aggregate is quartz sand with a particle size of 0.15-0.6 mm, the fly ash is I-class fly ash, the silica fume contains SiO2≥90%, the chloride ion in the silica fume is ≤0.1%, the sulfide in the silica fume is ≤2%, the fineness of the silica fume meets the specific surface area≥15.000 m² / kg, the water content in the silica fume is ≤3%, the polycarboxylic acid water reducing agent meets: water reducing rate≥35%, air content≤6%, alkali content≤10%.

[0133] The base mix proportion design module is used to generate multiple groups of mix proportions under the concrete matrix strength grade corresponding to the maximum negative poisson ratio value, calculate the fitness of the mechanical property prediction value and the target value based on the spiral dilatancy fiber characteristic parameters and the mix proportions, search the base mix proportion based on the genetic algorithm, and verify the base mix proportion.

[0134] Wherein, the process of obtaining the mechanical property prediction value based on the spiral dilatancy fiber characteristic parameters and the mix proportions includes:

[0135] The performance prediction network is trained by taking the helical auxetic fiber characteristic parameters and the mixing ratio as inputs and taking the mechanical properties of the fiber concrete as output labels, wherein the helical auxetic fiber characteristic parameters include fiber type, diameter, length, elastic modulus, and Poisson's ratio, and the mixing ratio includes water-cement ratio, sand ratio, water reducing agent content, and helical auxetic fiber volume content;

[0136] The helical auxetic fiber characteristic parameters and the mixing ratio are input into the performance prediction network, wherein the water-cement ratio in the mixing ratio is less than or equal to 0.3, the sand ratio ranges from 30% to 40%, and the mechanical property prediction value is output, and the mechanical properties include splitting tensile strength, toughness index, and strain energy.

[0137] The fitness of the mechanical property prediction value and the target value is:

[0138]

[0139] wherein, , , are the target splitting tensile strength, the target toughness index, and the target strain energy, respectively, , , are the predicted splitting tensile strength, the predicted toughness index, and the predicted strain energy, respectively, is a cost coefficient.

[0140] The process of searching for the matrix mixing ratio based on the genetic algorithm includes:

[0141] Under the concrete matrix strength grade corresponding to the maximum negative Poisson's ratio value, a plurality of mixing ratios are randomly generated to form an initial population, and each mixing ratio serves as an individual;

[0142] The fitness of each individual is calculated, the individuals are selected according to the fitness, the selected individuals are subjected to a crossover operation to generate new individuals, the individuals after the crossover operation are subjected to a mutation operation to generate the next generation population, and the cycle is executed until the iteration is set to a certain number of times or the fitness change is less than 1e -5 , and the matrix mixing ratio is obtained.

[0143] The matrix mixing ratio includes water-cement ratio , sand ratio , helical auxetic fiber volume content , helical auxetic fiber negative Poisson's ratio , and water reducing agent content , and the constraint condition of the matrix mixing ratio is verified as: , , , , . ​

[0144] A concrete preparation module for preparing concrete from a matrix mix ratio and a spiral auxetic fiber.

[0145] To ensure the normal operation of the model and sufficient data storage space, the software and hardware configuration adopted by the present application is as follows: Windows 11 operating system is equipped with Python 3.7.2 and Matlab R2021b, which can provide strong programming and scientific computing support; an AMD Ryzen 7 5800H with Radeon Graphics processor is internally mounted, which is matched with an NVIDIA GeForce RTX 3050-6G graphics processor to ensure the computing power and graphics processing capability; in addition, 16GB of memory and 520GB of solid state disk are equipped to ensure the multitasking processing capability and efficient data access.

[0146] The test results of the maximum negative Poisson's ratio of the spiral auxetic fiber in the different strength concrete matrix in Example 1 are introduced as follows through a plurality of comparative examples:

[0147] Comparative Example 1

[0148] The preparation process of the spiral auxetic fiber with negative Poisson's ratio effect is as follows: 1, winding the core fiber polypropylene fiber (diameter 0.9mm) on the fiber conveying device, ensuring that the core fiber is in a straight state, and ensuring that the fiber conveying device can uniformly convey the polypropylene fiber; 2, fixing the steel fiber (diameter 0.3mm) on the winding device (i.e. hollow spindle), adjusting the fiber conveying device conveying fiber speed and hollow spindle rotating speed, so that the steel fiber is wound on the polypropylene fiber at a wrapping angle of 15°; 3, using the take-up device to collect the prepared spiral auxetic fiber, then cutting the fiber to 30mm long, and dropping universal glue on both ends of the cut fiber to ensure the stability of the fiber.

[0149] The method for measuring the Poisson's ratio of the spiral auxetic fiber in the concrete with different matrix strengths is as follows:

[0150] (1) Surface modification: immerse the spiral auxetic fiber in a 3% silane coupling agent (KH-550) ethanol solution for 15 min, and dry at 60°C before use.

[0151] (2) Add 1100 parts of coarse aggregate and 700 parts of fine aggregate to the mixer mixing container and stir for 1 min.

[0152] (3) Mix 300 parts of cement, 120 parts of slag powder, 50 parts of fly ash and 30 parts of silica fume for 2 min, then pour into the above aggregate mixture and stir for another 2 min.

[0153] (4) 140 parts of water and 14 parts of polycarboxylate superplasticizer were mixed and stirred for 30 s, and then poured into the mixer mixing container, and stirred for 3 min.

[0154] (5) The concrete was poured into the mold, and the spiral tensar fiber with a length of 30 mm was vertically embedded in the concrete matrix with an embedding depth of 20 mm, and then cured for 28 d according to the requirements of the national standard, and the concrete matrix strength grade was C60.

[0155] (6) The pull-out load was applied to the spiral tensar fiber, and the spiral tensar fiber inside the concrete was scanned in situ using CT, and the Poisson's ratio of the spiral tensar fiber inside the concrete during the load action was recorded.

[0156] Comparative Example 2

[0157] The preparation process of the spiral tensar fiber with negative Poisson's ratio effect is as follows: 1, the core fiber polypropylene fiber (diameter 0.9 mm) is wound on the fiber conveying device, ensuring that the core fiber is in a straight state, and ensuring that the fiber conveying device can uniformly convey the polypropylene fiber; 2, the steel fiber (diameter 0.3 mm) is fixed on the winding device (i.e. hollow spindle), and the fiber conveying device conveying speed and the hollow spindle rotating speed are adjusted, so that the steel fiber is wound on the polypropylene fiber at a wrapping angle of 15°; 3, the prepared spiral tensar fiber is collected using the take-up device, and then the fiber is cut to a length of 30 mm, and universal glue is added at both ends of the cut fiber to ensure the stability of the fiber.

[0158] The Poisson's ratio measurement method of the spiral tensar fiber in different matrix strength concrete is as follows:

[0159] (1) Surface modification: the spiral tensar fiber was immersed in a 3% silane coupling agent (KH-550) ethanol solution for 15 min, and then dried at 60°C for standby.

[0160] (2) 1100 parts of coarse aggregate and 650 parts of fine aggregate were added to the mixer mixing container and stirred for 1 min.

[0161] (3) 300 parts of cement, 150 parts of slag powder, 30 parts of fly ash and 50 parts of silica ash were mixed and stirred for 2 min, and then poured into the above aggregate mixture and stirred for another 2 min.

[0162] (4) 140 parts of water and 14 parts of polycarboxylate superplasticizer were mixed and stirred for 30 s, and then poured into the mixer mixing container, and stirred for 3 min.

[0163] (5) The concrete was poured into the mold, and the spiral tensar fiber with a length of 30 mm was vertically embedded in the concrete matrix with an embedding depth of 20 mm, and then cured for 28 d according to the requirements of the national standard, and the concrete matrix strength grade was C80.

[0164] (6) Apply the pull-out load to the helical strain-hardening fiber, use in-situ CT to scan the helical strain-hardening fiber inside the concrete, and record the Poisson's ratio of the helical strain-hardening fiber inside the concrete during the load action.

[0165] Comparative Example 3

[0166] The preparation process of the helical strain-hardening fiber with negative Poisson's ratio effect is as follows: 1, the core fiber polypropylene fiber (diameter 0.9 mm) is wound on the fiber conveying device, to ensure that the core fiber is in a straight state, and to ensure that the fiber conveying device can uniformly convey the polypropylene fiber; 2, the steel fiber (diameter 0.3 mm) is fixed on the winding device (i.e. hollow spindle), and the fiber conveying device conveying speed and the hollow spindle rotating speed are adjusted, so that the steel fiber is wound on the polypropylene fiber at a wrapping angle of 15°; 3, the prepared helical strain-hardening fiber is collected using the take-up device, and then the fiber is cut to a length of 30 mm, and universal glue is added at both ends of the cut fiber to ensure the stability of the fiber.

[0167] The Poisson's ratio measurement method of the helical strain-hardening fiber in different matrix strength concrete is as follows:

[0168] (1) Surface modification: immerse the helical strain-hardening fiber in a 3% silane coupling agent (KH-550) ethanol solution for 15 min, and dry at 60°C before use.

[0169] (2) Add 1050 parts of coarse aggregate and 650 parts of fine aggregate to the mixer mixing container and stir for 1 min.

[0170] (3) Mix 280 parts of cement, 220 parts of slag powder, 60 parts of fly ash and 80 parts of silica fume for 2 min, then pour into the above aggregate mixture and stir for another 2 min.

[0171] (4) Mix 140 parts of water and 20 parts of polycarboxylic acid water reducing agent for 30 s, then pour into the mixer mixing container and stir for 3 min.

[0172] (5) Pour the concrete into the mold, and vertically embed the helical strain-hardening fiber with a length of 30 mm into the concrete matrix, with an embedding depth of 20 mm, and then cure for 28 d according to the national standard requirements, and the concrete matrix strength grade is C100.

[0173] (6) Apply the pull-out load to the helical strain-hardening fiber, use in-situ CT to scan the helical strain-hardening fiber inside the concrete, and record the Poisson's ratio of the helical strain-hardening fiber inside the concrete during the load action.

[0174] Comparative Example 4

[0175] The preparation process of the helical auxetic fiber with negative Poisson's ratio effect is as follows: 1. The core fiber polypropylene fiber (diameter 0.9 mm) is wound on the fiber conveying device, ensuring that the core fiber is in a straightened state, and ensuring that the fiber conveying device can uniformly convey the polypropylene fiber; 2. The steel fiber (diameter 0.3 mm) is fixed on the winding device (i.e. hollow spindle), and the fiber conveying device conveying speed and the hollow spindle rotating speed are adjusted, so that the steel fiber is wound on the polypropylene fiber at a wrapping angle of 20°; 3. The prepared helical auxetic fiber is collected using the take-up device, and then the fiber is cut to a length of 30 mm, and universal glue is added at both ends of the cut fiber to ensure the stability of the fiber.

[0176] The method for measuring the Poisson's ratio of the helical auxetic fiber in different matrix strength concrete is as follows:

[0177] (1) Surface modification: immerse the helical auxetic fiber in a 3% silane coupling agent (KH-550) ethanol solution for 15 min, and dry at 60°C before use.

[0178] (2) Add 1100 parts of coarse aggregate and 700 parts of fine aggregate to the mixer mixing container and stir for 1 min.

[0179] (3) Mix 300 parts of cement, 120 parts of slag powder, 50 parts of fly ash and 30 parts of silica fume for 2 min, then pour into the above aggregate mixture and stir for another 2 min.

[0180] (4) Mix 150 parts of water and 7.5 parts of polycarboxylic acid water reducer for 30 s, then pour into the mixer mixing container and stir for 3 min.

[0181] (5) Pour the concrete into the mold, and vertically embed the helical auxetic fiber with a length of 30 mm into the concrete matrix, with an embedding depth of 20 mm, then cure for 28 d according to the national standard requirements, and the concrete matrix strength grade is C60.

[0182] (6) Apply the pull-out load to the helical auxetic fiber, use in-situ CT to scan the helical auxetic fiber inside the concrete, and record the Poisson's ratio of the helical auxetic fiber inside the concrete during the load action.

[0183] Comparative Example 5

[0184] The preparation process of the helical auxetic fiber with negative Poisson's ratio effect is as follows: 1. The core fiber polypropylene fiber (diameter 0.9 mm) is wound on the fiber conveying device, ensuring that the core fiber is in a straightened state, and ensuring that the fiber conveying device can uniformly convey the polypropylene fiber; 2. The steel fiber (diameter 0.3 mm) is fixed on the winding device (i.e. hollow spindle), and the fiber conveying device conveying speed and the hollow spindle rotating speed are adjusted, so that the steel fiber is wound on the polypropylene fiber at a wrapping angle of 20°; 3. The prepared helical auxetic fiber is collected using the take-up device, and then the fiber is cut to a length of 30 mm, and universal glue is added at both ends of the cut fiber to ensure the stability of the fiber.

[0185] The method for measuring the Poisson's ratio of the helical auxetic fiber in the different matrix strength concrete is as follows:

[0186] (1) Surface modification: immerse the helical auxetic fiber in a 3% silane coupling agent (KH-550) ethanol solution for 15 min, and dry at 60°C for standby.

[0187] (2) Add 1100 parts of coarse aggregate and 650 parts of fine aggregate to the mixer mixing container and stir for 1 min.

[0188] (3) Mix 300 parts of cement, 150 parts of slag powder, 30 parts of fly ash and 50 parts of silica fume for 2 min, then pour into the above aggregate mixture and stir for another 2 min.

[0189] (4) Mix 140 parts of water and 14 parts of polycarboxylic acid water reducer for 30 s, then pour into the mixer mixing container and stir for 3 min.

[0190] (5) Pour the concrete into the mold, and vertically bury the helical auxetic fiber with a length of 30 mm in the concrete matrix, with a burial depth of 20 mm, and then cure for 28 d according to the national standard requirements. The concrete matrix strength grade is C80.

[0191] (6) Apply the pull-out load to the helical auxetic fiber, use in-situ CT to scan the helical auxetic fiber inside the concrete, and record the Poisson's ratio of the helical auxetic fiber inside the concrete during the load action.

[0192] Comparative Example 6

[0193] The preparation process of the helical auxetic fiber with negative Poisson's ratio effect is as follows: 1. The core fiber polypropylene fiber (diameter 0.9 mm) is wound on the fiber conveying device, ensuring that the core fiber is in a straightened state, and ensuring that the fiber conveying device can uniformly convey the polypropylene fiber; 2. The steel fiber (diameter 0.3 mm) is fixed on the winding device (i.e. hollow spindle), and the fiber conveying device conveying speed and the hollow spindle rotating speed are adjusted, so that the steel fiber is wound on the polypropylene fiber at a wrapping angle of 20°; 3. The prepared helical auxetic fiber is collected using the take-up device, and then the fiber is cut to a length of 30 mm, and universal glue is added at both ends of the cut fiber to ensure the stability of the fiber.

[0194] The method for measuring the Poisson's ratio of the helical auxetic fiber in the different matrix strength concrete is as follows:

[0195] (1) Surface modification: immerse the helical auxetic fiber in a 3% silane coupling agent (KH-550) ethanol solution for 15 min, and dry at 60°C for standby.

[0196] (2) Add 1050 parts of coarse aggregate and 650 parts of fine aggregate into the mixer mixing container and stir for 1 min.

[0197] (3) Mix and stir 280 parts of cement, 220 parts of slag powder, 60 parts of fly ash and 80 parts of silica fume for 2 min, then pour into the above aggregate mixture and stir for another 2 min.

[0198] (4) Mix and stir 140 parts of water and 20 parts of polycarboxylic acid water reducer for 30 s, then pour into the mixer mixing container and stir for 3 min.

[0199] (5) Pour the concrete into the mold, and vertically bury the helical auxetic fiber with a length of 30 mm in the concrete matrix, with a burial depth of 20 mm, then cure for 28 d according to the national standard requirements, and the concrete matrix strength grade is C100.

[0200] (6) Apply the pull-out load to the helical auxetic fiber, use in-situ CT to scan the helical auxetic fiber inside the concrete, and record the Poisson's ratio of the helical auxetic fiber inside the concrete during the load action.

[0201] Comparative Example 7

[0202] The preparation process of the helical auxetic fiber with negative Poisson's ratio effect is as follows: 1. The core fiber polypropylene fiber (diameter 1.2 mm) is wound on the fiber conveying device, ensuring that the core fiber is in a straightened state, and ensuring that the fiber conveying device can uniformly convey the polypropylene fiber; 2. The steel fiber (diameter 0.3 mm) is fixed on the winding device (i.e. hollow spindle), and the fiber conveying device conveying speed and the hollow spindle rotating speed are adjusted, so that the steel fiber is wound on the polypropylene fiber at a wrapping angle of 10°; 3. The prepared helical auxetic fiber is collected using the take-up device, and then the fiber is cut to a length of 30 mm, and universal glue is added at both ends of the cut fiber to ensure the stability of the fiber.

[0203] The method for measuring the Poisson's ratio of the helical auxetic fiber in different matrix strength concrete is as follows:

[0204] (1) Surface modification: immerse the helical auxetic fiber in a 3% silane coupling agent (KH-550) ethanol solution for 15 min, and dry at 60°C for standby.

[0205] (2) Add 1100 parts of coarse aggregate and 700 parts of fine aggregate to the mixer mixing container and stir for 1 min.

[0206] (3) Mix 300 parts of cement, 120 parts of slag powder, 50 parts of fly ash and 30 parts of silica fume for 2 min, then pour into the above aggregate mixture and stir for another 2 min.

[0207] (4) Mix 150 parts of water and 7.5 parts of polycarboxylic acid water reducer for 30 s, then pour into the mixer mixing container and stir for 3 min.

[0208] (5) Pour the concrete into the mold, and vertically embed the helical auxetic fiber with a length of 30 mm into the concrete matrix, with an embedding depth of 20 mm, and then cure for 28 d according to the national standard requirements. The concrete matrix strength grade is C60.

[0209] (6) Apply the pull-out load to the helical auxetic fiber, use in-situ CT to scan the helical auxetic fiber inside the concrete, and record the Poisson's ratio of the helical auxetic fiber inside the concrete during the load action.

[0210] Comparative Example 8

[0211] The preparation process of the helical auxetic fiber with negative Poisson's ratio effect is as follows: 1. The core fiber polypropylene fiber (diameter 1.2 mm) is wound on the fiber conveying device, ensuring that the core fiber is in a straightened state, and ensuring that the fiber conveying device can uniformly convey the polypropylene fiber; 2. The steel fiber (diameter 0.3 mm) is fixed on the winding device (i.e. hollow spindle), and the fiber conveying device conveying speed and the hollow spindle rotating speed are adjusted, so that the steel fiber is wound on the polypropylene fiber at a wrapping angle of 10°; 3. The prepared helical auxetic fiber is collected using the take-up device, and then the fiber is cut to a length of 30 mm, and universal glue is added at both ends of the cut fiber to ensure the stability of the fiber.

[0212] The method for measuring the Poisson's ratio of the helical auxetic fiber in the different matrix strength concrete is as follows:

[0213] (1) Surface modification: immerse the helical auxetic fiber in a 3% silane coupling agent (KH-550) ethanol solution for 15 min, and dry at 60°C for standby.

[0214] (2) Add 1100 parts of coarse aggregate and 650 parts of fine aggregate to the mixer mixing container and stir for 1 min.

[0215] (3) Mix 300 parts of cement, 150 parts of slag powder, 30 parts of fly ash and 50 parts of silica fume for 2 min, then pour into the above aggregate mixture and stir for another 2 min.

[0216] (4) Mix 140 parts of water and 14 parts of polycarboxylic acid water reducer for 30 s, then pour into the mixer mixing container and stir for 3 min.

[0217] (5) Pour the concrete into the mold, and vertically bury the helical auxetic fiber with a length of 30 mm in the concrete matrix, with a burial depth of 20 mm, and then cure for 28 d according to the national standard requirements. The concrete matrix strength grade is C80.

[0218] (6) Apply the pull-out load to the helical auxetic fiber, use in-situ CT to scan the helical auxetic fiber inside the concrete, and record the Poisson's ratio of the helical auxetic fiber inside the concrete during the load action.

[0219] Comparative Example 9

[0220] The preparation process of the helical auxetic fiber with negative Poisson's ratio effect is as follows: 1, winding the core fiber polypropylene fiber (diameter 1.2 mm) on the fiber conveying device, ensuring that the core fiber is in a straight state, and ensuring that the fiber conveying device can uniformly convey the polypropylene fiber; 2, fixing the steel fiber (diameter 0.3 mm) on the winding device (i.e. hollow spindle), adjusting the fiber conveying speed of the fiber conveying device and the rotation speed of the hollow spindle, so that the steel fiber is wound on the polypropylene fiber at a wrapping angle of 10°; 3, collecting the prepared helical auxetic fiber using the take-up device, then cutting the fiber into a length of 30 mm, and dropping universal glue at both ends of the cut fiber to ensure the stability of the fiber.

[0221] The method for measuring the Poisson's ratio of the helical auxetic fiber in different matrix strength concrete is as follows:

[0222] (1) Surface modification: immerse the helical auxetic fiber in a 3% silane coupling agent (KH-550) ethanol solution for 15 min, and then dry at 60°C for standby.

[0223] (2) Add 1050 parts of coarse aggregate and 650 parts of fine aggregate into the mixer mixing container and stir for 1 min.

[0224] (3) Mix and stir 280 parts of cement, 220 parts of slag powder, 60 parts of fly ash and 80 parts of silica fume for 2 min, then pour into the above aggregate mixture and stir for another 2 min.

[0225] (4) Mix and stir 140 parts of water and 20 parts of polycarboxylic acid water reducing agent for 30 s, then pour into the mixer mixing container and stir for 3 min.

[0226] (5) Pour the concrete into the mold, and vertically bury the helical auxetic fiber with a length of 30 mm in the concrete matrix, with a burial depth of 20 mm, then cure for 28 d according to the national standard requirements, and the concrete matrix strength grade is C100.

[0227] (6) Apply the pull-out load to the helical auxetic fiber, use in-situ CT to scan the helical auxetic fiber inside the concrete, and record the Poisson's ratio of the helical auxetic fiber inside the concrete during the load action.

[0228] The detection results of the maximum negative Poisson's ratio of the helical auxetic fiber in different matrix strength concrete are shown in Table 5: from Table 5, based on the principle of maximizing the negative Poisson's ratio effect of the helical auxetic fiber, the helical auxetic fiber with a diameter ratio of 1-3 and a wrapping angle of 15° has the best matching performance with C60 concrete; the helical auxetic fiber with a diameter ratio of 1-3 and a wrapping angle of 20° has the best matching performance with C80 concrete; the helical auxetic fiber with a diameter ratio of 1-4 and a wrapping angle of 10° has the best matching performance with C100 concrete.

[0229] Table 5. Statistical analysis of the maximum negative Poisson's ratio of helical tensile fibers in concrete with different matrix strengths.

[0230]

[0231] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for the production of spiral tensile fiber concrete with high blast resistance, characterized in that: The method comprises: According to the fiber type and the mechanical property parameter, the diameter ratio and the wrapping angle of the spiral auxetic fiber are predicted based on deep learning, and the spiral auxetic fiber is prepared; The spiral auxetic fiber is embedded in concrete with different matrix strengths, and the negative Poisson's ratio of the spiral auxetic fiber is measured after curing, and the maximum negative Poisson's ratio corresponding to the concrete matrix strength grade is recorded; Under the concrete matrix strength grade corresponding to the maximum negative Poisson's ratio value, a plurality of groups of mix proportions are generated, the mechanical property prediction values are obtained based on deep learning according to the characteristics parameters of the helical tensile fiber and the mix proportions, the fitness of the mechanical property prediction values and the target values is calculated, the matrix mix proportion is searched based on the genetic algorithm, and the matrix mix proportion is verified; the fitness of the mechanical property prediction values and the target values is is: wherein, , , are the target split tensile strength, target toughness index, target strain energy, respectively, , , are the predicted split tensile strength, predicted toughness index, predicted strain energy, respectively, is a cost coefficient; According to the matrix mixing ratio and the spiral auxetic fiber, the concrete is prepared.

2. The method of manufacturing the spiral tensylon fiber concrete having high blast resistance according to claim 1, wherein: The process of predicting the diameter ratio and the wrapping angle of the fiber based on deep learning according to the fiber type and the mechanical property parameter comprises: Taking the fiber type and the fiber mechanical property parameter of the spiral yarn as the input and the structure parameter of the spiral yarn as the output label, a multi-layer fully connected neural network is trained to obtain a deep learning model, wherein the fiber mechanical property parameter comprises Poisson's ratio, elastic modulus and elongation at break, and the structure parameter of the spiral yarn comprises diameter ratio, wrapping angle and negative Poisson's ratio effect value; Two fibers with different elastic moduli are selected, and the fiber type and the mechanical property parameter are input into the deep learning model to output the diameter ratio and the wrapping angle of the spiral auxetic fiber.

3. The method of claim 1, wherein the spiral tensylon fiber concrete having high blast resistance is characterized by: The process of preparing the spiral auxetic fiber comprises: Two fibers with different elastic moduli are selected, one type of fiber is a core fiber, and the other type of fiber is a wrapping fiber, and the fiber type comprises polypropylene fiber, polyvinyl alcohol fiber, flax fiber, glass fiber and steel fiber; The core fiber is wound on a fiber conveying device so that the core fiber is in a straightened state, and the fiber conveying device uniformly conveys the core fiber; The wrapping fiber is fixed on a winding device, the conveying speed of the fiber conveying device is adjusted to be the same as the rotating speed of the winding device, the wrapping fiber is wrapped around the core fiber at the predicted wrapping angle, and the spiral auxetic fiber is obtained; The prepared spiral auxetic fiber is collected by using a take-up device, the spiral auxetic fiber is cut, and universal glue is added at both ends of the cut spiral auxetic fiber.

4. The method of claim 1, wherein the spiral tensylon fiber concrete having high blast resistance is characterized by: Different matrix strengths of concrete are prepared by different proportions of raw materials and the same preparation method, the spiral auxetic fiber is embedded in concrete with different matrix strengths, and the negative Poisson's ratio of the spiral auxetic fiber is measured after curing. The spiral auxetic fiber is immersed in a 3% silane coupling agent ethanol solution for 15 min, and then dried at 60°C for standby; The coarse aggregate and the fine aggregate are added to the mixer mixing container and stirred for 1 min; The cement, slag powder, fly ash and silica ash are mixed and stirred for 2 min, then poured into the mixer mixing container and stirred for another 2 min; The water and the polycarboxylic acid water reducing agent are mixed and stirred for 30 s, then poured into the mixer mixing container and stirred for 3 min to obtain the concrete with the target matrix strength; The concrete with the target matrix strength is poured into a mold, and the spiral auxetic fiber with a length of 30 mm is vertically embedded in the concrete matrix with an embedding depth of 20 mm, and cured for 28 days; The spiral auxetic fiber is pulled out under load, the spiral auxetic fiber inside the concrete is scanned by in-situ CT, and the negative Poisson's ratio of the spiral auxetic fiber inside the concrete during the load action is recorded.

5. The method of claim 4, wherein the spiral tensylon fiber concrete having high blast resistance is characterized by: The cement is PII·52.5 grade ordinary portland cement, the fine aggregate is quartz sand with a particle size of 0.15-0.6 mm, the fly ash is I grade fly ash, the silica ash contains SiO2≥90%, the chloride ion in the silica ash is ≤0.1%, the sulfide in the silica ash is ≤2%, the fineness of the silica ash meets the specific surface area≥15.000 m² / kg, the water content in the silica ash is ≤3%, and the polycarboxylic acid water reducing agent meets the water reducing rate≥35%, the air content≤6% and the alkali content≤10%.

6. The method of manufacturing the spiral tensylon fiber concrete having high blast resistance according to claim 1, wherein: According to the characteristic parameters of the spiral tensar fiber and the mixing proportion, the process of obtaining the mechanical property prediction value based on deep learning includes: Taking the characteristic parameters of the spiral tensar fiber and the mixing proportion as inputs and the mechanical property of the fiber concrete as an output label, a bidirectional LSTM network is trained to obtain a performance prediction network, wherein the characteristic parameters of the spiral tensar fiber include fiber type, diameter, length, elastic modulus and Poisson's ratio, and the mixing proportion includes water-cement ratio, sand ratio, water reducing agent content and spiral tensar fiber volume content; The characteristic parameters of the spiral tensar fiber and the mixing proportion are input into the performance prediction network, wherein the water-cement ratio in the mixing proportion is less than or equal to 0.3, the sand ratio ranges from 30% to 40%, and the mechanical property prediction value is output, including splitting tensile strength, toughness index and strain energy.

7. A method of preparing a spiral tensylon fibre concrete with high blast resistance according to claim 6, characterized in that: Loss function for training a bidirectional LSTM network is: + + wherein, , , are the true values of the split tensile strength, the toughness index, the strain energy, respectively, of the first sample, , , are the predicted values of the split tensile strength, the toughness index, the strain energy, respectively, of the first sample.

8. The method of manufacturing the spiral tensylon fiber concrete having high blast resistance according to claim 1, wherein: The process of searching for the matrix mixing proportion based on a genetic algorithm includes: Under the strength grade of the concrete matrix corresponding to the maximum negative Poisson's ratio value, a plurality of mixing proportions are randomly generated to form an initial population, and each mixing proportion is taken as an individual. The fitness of each individual is calculated, the individuals are screened according to the fitness, the selected individuals are subjected to crossover operation to generate new individuals, the individuals after the crossover operation are subjected to mutation operation to produce a next generation population, and the cycle is executed until a preset number of iterations is met or the fitness change is less than 1e -5 , and the base body mixing ratio is obtained.

9. The method of manufacturing the spiral tensylon fiber concrete having high blast resistance according to claim 1, wherein: The matrix matching ratio includes water-cement ratio , sand ratio , spiral tensile fiber volume content , negative Poisson's ratio of spiral tensile fiber , water reducing agent content , the constraint condition for verifying the matrix matching ratio is: , , , , .

10. A system for the production of spiral tensile fiber concrete with high blast resistance, characterized by: The system includes: A spiral tensar fiber design module for predicting the diameter ratio and wrapping angle of the spiral tensar fiber based on deep learning according to the fiber type and mechanical property parameters, and preparing the spiral tensar fiber; A matrix strength determination module based on the maximization of the negative Poisson's ratio effect, for embedding the spiral tensar fiber in concrete with different matrix strengths, measuring the negative Poisson's ratio of the spiral tensar fiber after curing, and recording the strength grade of the concrete matrix corresponding to the maximum negative Poisson's ratio value; The base mixture ratio design module is used to generate multiple groups of mixture ratios under the concrete base strength grade corresponding to the maximum negative Poisson's ratio value, calculate the fitness of the mechanical property prediction value and the target value based on the spiral dilatation fiber characteristic parameters and the mixture ratio, obtain the base mixture ratio based on the genetic algorithm, and verify the base mixture ratio; the fitness of the mechanical property prediction value and the target value is: wherein, , , are the target split tensile strength, target toughness index, target strain energy, respectively, , , are the predicted split tensile strength, predicted toughness index, predicted strain energy, respectively, is a cost coefficient; A concrete preparation module for preparing concrete according to the matrix mixing proportion and the spiral tensar fiber.

Images