An intelligent design method and system for fiber composite anchor bolts

Abstract

The present invention relates to the technical fields of civil engineering and composite materials, and in particular to an intelligent design method and system for fiber composite bolts. The method includes the following steps: making a carbon-glass hybrid bolt core using carbon fiber and glass fiber, simultaneously using LCP fiber bundles to braid the surface of the carbon-glass hybrid bolt core to form a braided layer, and then using LCP fiber bundles to wind around the surface of the braided layer to form LCP fiber winding ribs to obtain a fiber composite bolt; adjusting the adjustable parameters of the fiber composite bolt to produce multiple fiber composite bolts, and then constructing a production parameter prediction model, and using the production parameter prediction model to obtain the production parameter prediction value of the fiber composite bolt when given bolt performance parameters; producing the fiber composite bolt according to the production parameter prediction value. Using the present invention, fiber composite bolts with high mechanical properties and economy can be produced, and the production efficiency can be improved and the production cost can be reduced.

Description

Technical Field

[0001] The present invention relates to the technical fields of civil engineering and composite materials, and in particular, to an intelligent design method and system for fiber composite bolts. Background Art

[0002] Bolting support is a commonly used form of geotechnical anchoring and can be used in underground engineering, slope stability, tunnel construction, mine roadways, deep foundation pit support, and various geotechnical reinforcement projects. By burying bolts in the geotechnical body and utilizing the adhesive force and frictional force between the bolts and the surrounding geotechnical body, stable supporting force is provided, thereby enhancing the overall stability of the geotechnical body and ensuring the safety of the engineering structure. Currently, the commonly used steel bolts have disadvantages such as insufficient strength, excessive self-weight, easy corrosion, and relatively large creep and relaxation, and it is increasingly difficult to meet the construction requirements of high performance and long life for geotechnical anchoring in the new era infrastructure construction.

[0003] Manufacturing bolts using fiber composites can fundamentally solve the problems existing in steel bolts. The existing fiber composite bolts mainly include glass fiber composite bolts, basalt fiber composite bolts, and carbon fiber composite bolts. However, there are still some problems with the existing fiber composite bolts, which prevent them from being used on a large scale. Among them, glass fiber composite and basalt fiber composite bolts have relatively large creep, low creep fracture stress, and low shear and flexural strength; carbon fiber composite bolts have a relatively small fracture elongation rate, high price, and also low shear and flexural strength.

[0004] In addition, in the traditional production process of fiber composite bolts, producing fiber composite bolts that match the construction requirements is often based on empirical methods and trial-and-error methods, which not only have low production efficiency but also high costs. If fiber composite bolts higher than the construction requirements are directly selected, it will also bring the problem of high costs and cause waste of resources.

[0005] Therefore, it is necessary to explore a new production plan for fiber composite bolts to solve the above technical problems. Summary of the Invention

[0006] Aiming at the defects in the prior art, the present invention provides an intelligent design method and system for fiber composite bolts.

[0007] To achieve the above object, in a first aspect, the present invention provides an intelligent design method for fiber composite bolts. The method includes the following steps: making a carbon-glass hybrid bolt core using carbon fiber and glass fiber, and at the same time using an LCP fiber bundle to braid the surface of the carbon-glass hybrid bolt core to form a braided layer; after braiding the surface of the carbon-glass hybrid bolt core, using the LCP fiber bundle to wind around the surface of the braided layer to form an LCP fiber winding rib, thereby obtaining a fiber composite bolt; adjusting the adjustable parameters of the fiber composite bolt to produce multiple fiber composite bolts, and then constructing a production prediction data set for the fiber composite bolts; constructing a production parameter prediction model based on the production prediction data set for the fiber composite bolts, and using the production parameter prediction model to obtain the production parameter prediction value of the fiber composite bolt when given bolt performance parameters; producing the fiber composite bolt according to the production parameter prediction value. Using the present invention, fiber composite bolts with high mechanical properties and economy can be produced, and the production efficiency can be improved and the production cost can be reduced.

[0008] Optionally, the step of making a carbon-glass hybrid bolt core using carbon fiber and glass fiber, and at the same time using an LCP fiber bundle to braid the surface of the carbon-glass hybrid bolt core to form a braided layer includes the following steps:

[0009] Simultaneously extracting at least one carbon fiber and at least one glass fiber and straightening them, and then impregnating the carbon fiber and the glass fiber with a thermosetting resin to bond the fibers to each other;

[0010] After the carbon fiber and the glass fiber are impregnated with the thermosetting resin, a carbon-glass hybrid pultrusion technique is used to form the carbon-glass hybrid bolt core, and at the same time an LCP fiber bundle is used to braid the surface of the carbon-glass hybrid bolt core to form the braided layer.

[0011] Optionally, the step of, after braiding the surface of the carbon-glass hybrid bolt core, using the LCP fiber bundle to wind around the surface of the braided layer to form an LCP fiber winding rib, thereby obtaining a fiber composite bolt includes the following steps:

[0012] Impregnating the carbon-glass hybrid bolt core after braiding with a thermosetting resin, and then using an LCP fiber bundle to wind around the surface of the braided layer in a spiral shape to form an LCP fiber winding rib, thereby obtaining a semi-finished bolt;

[0013] Curing the semi-finished bolt under set temperature and pressure to tightly bond the carbon-glass hybrid bolt core, the braided layer and the LCP fiber winding rib, thereby obtaining the fiber composite bolt.

[0014] Optionally, the steps of adjusting the adjustable parameters of the fiber composite bolt to produce a plurality of the fiber composite bolts, and then constructing a production prediction data set for the fiber composite bolts are as follows:

[0015] Adjust the values of different adjustable parameters to generate a plurality of combinations of adjustable parameters, where the adjustable parameters include fiber hybridization ratio, braiding angle, braiding density, bolt diameter, and bolt length;

[0016] Produce corresponding fiber composite bolts according to the combinations of adjustable parameters and test their flexural strength, tensile strength, and shear strength;

[0017] Use all combinations of adjustable parameters and the flexural strength, tensile strength, and shear strength of the corresponding fiber composite bolts to construct a production prediction data set for the fiber composite bolts.

[0018] Optionally, the bolt performance parameters include bolt diameter, bolt length, flexural strength, tensile strength, and shear strength, and the production parameters of the fiber composite bolt include fiber hybridization ratio, braiding angle, and braiding density;

[0019] The steps of constructing a production parameter prediction model based on the production prediction data set for the fiber composite bolts and using the production parameter prediction model to obtain the production parameter prediction values of the fiber composite bolts when given bolt performance parameters are as follows:

[0020] Combine a multi-strategy improved genetic algorithm and an optimized BP neural network into a combined prediction model;

[0021] Use the production prediction data set for the fiber composite bolts to train and validate the combined prediction model to obtain the production parameter prediction model;

[0022] Input the given bolt performance parameters into the production parameter prediction model to obtain the production parameter prediction values of the fiber composite bolts.

[0023] Optionally, when the multi-strategy improved genetic algorithm runs, the following steps are executed:

[0024] Set algorithm parameters and generate chromosomes based on the initial weights and initial thresholds of the optimized BP neural network to initialize the population;

[0025] The genetic algorithm iteration starts, calculate the fitness values of each chromosome, then calculate the chromosome difference, and compare the calculated chromosome difference with the difference threshold;

[0026] When the chromosome difference is less than the difference threshold, use the chromosome change strategy to change the chromosome until the chromosome difference is not less than the difference threshold;

[0027] When the chromosome difference is not less than the difference threshold, crossover and mutation of the chromosome are performed according to the adaptive crossover probability model and the adaptive mutation probability model, and then new chromosomes are generated;

[0028] Determine whether the maximum number of iterations is reached, and output the best chromosome when the maximum number of iterations is reached, otherwise let the new chromosome obtained in the current iteration step enter the next iteration step.

[0029] Optionally, the chromosome difference, the adaptive crossover probability model, and the adaptive mutation probability model respectively satisfy the following relationships:

[0030]

[0031]

[0032]

[0033] Wherein, R is the chromosome difference, n is the number of chromosomes, is the fitness value of the jth chromosome, P is the crossover probability of the chromosome, is the starting value of the crossover probability, is the stable value of the crossover probability, is the larger fitness value among the fitness values of the two chromosomes participating in crossover, is the maximum fitness value of the chromosome, is the mutation probability of the chromosome, is the starting value of the mutation probability, is the stable value of the mutation probability, is the fitness value of the chromosome to be mutated.

[0034] Optionally, the chromosome change strategy includes the following steps:

[0035] Sort the chromosomes in descending order of the fitness value to obtain a chromosome sequence, and divide the chromosome sequence into a first sequence and a second sequence, where the first sequence is the first half of the chromosome sequence;

[0036] Extract a gene combination from each chromosome in the second sequence according to the set number of chromosome gene changes, and then construct an original gene combination sequence according to the sorting of the chromosomes;

[0037] Continuously adjust the sorting of the gene combinations in the original gene combination sequence to obtain a new gene combination sequence, and use the gene combination sequence with the largest difference from the original gene combination sequence as the final gene combination sequence;

[0038] According to the positions of genes on the chromosome in the original gene combination sequence, the genes on the chromosome in the second sequence are changed using the final gene combination sequence.

[0039] Optionally, combining the multi-strategy improved genetic algorithm and the optimized BP neural network into a combined prediction model includes the following steps:

[0040] An adaptive weight update model is introduced into the BP neural network to obtain the optimized BP neural network, and the adaptive weight update model satisfies the following relationship:

[0041]

[0042] Wherein, is the weight of the optimized BP neural network at the (i + 1)-th iteration step, is the weight of the optimized BP neural network at the i-th iteration step, 、 and are all algebraic expressions, , , MSE is the mean square error of the optimized BP neural network, is the weight of the optimized BP neural network at the (i - 1)-th iteration step, is the learning rate used by the optimized BP neural network at the (i - 1)-th iteration step, is a random number greater than 0 and less than 1;

[0043] The optimized BP neural network is used as the main part of the model of the combined prediction model to predict production parameters, and the multi-strategy improved genetic algorithm is used as the secondary part of the model of the combined prediction model to optimize the initial weights and initial thresholds of the optimized BP neural network.

[0044] In a second aspect, the present invention provides an intelligent design system for fiber composite bolts. The intelligent design system for fiber composite bolts uses the intelligent design method for fiber composite bolts provided by the present invention. The system includes: a fiber composite bolt production subsystem for manufacturing a carbon-glass hybrid bolt core using carbon fiber and glass fiber, and simultaneously weaving the surface of the carbon-glass hybrid bolt core with LCP fiber bundles to form a woven layer; after weaving the surface of the carbon-glass hybrid bolt core, winding the LCP fiber bundles on the surface of the woven layer to form LCP fiber winding ribs, thereby obtaining a fiber composite bolt; a bolt testing and data acquisition subsystem for adjusting the adjustable parameters of the fiber composite bolt to produce multiple fiber composite bolts, thereby constructing a production prediction data set for fiber composite bolts; and a production parameter judgment subsystem for constructing a production parameter prediction model based on the production prediction data set for fiber composite bolts, and using the production parameter prediction model to obtain the production parameter prediction value of the fiber composite bolt when given bolt performance parameters, thereby enabling the fiber composite bolt production subsystem to produce the fiber composite bolt according to the production parameter prediction value.

[0045] The present invention has at least the following beneficial effects:

[0046] 1. The carbon-glass hybrid bolt core of the present invention is formed by hybrid pultrusion of carbon fiber and glass fiber, which has high strength and corrosion resistance, effectively improves the mechanical properties of the bolt, and reduces the cost as much as possible, improving the economy of the bolt.

[0047] 2. The present invention weaves and winds the surface of the carbon-glass hybrid bolt core with LCP fiber, further enhancing the shear resistance and bending resistance of the bolt and improving the anchoring effect of the bolt.

[0048] 3. The fiber composite bolt produced by the present invention selects carbon fiber, glass fiber and LCP fiber as the main materials, has excellent corrosion resistance and creep resistance, overcomes the problems of easy rust and performance attenuation of traditional steel bolts, and enables the bolt to maintain long-term mechanical properties and structural stability in various harsh environments.

[0049] 4. The present invention constructs an accurate and reliable production parameter prediction model for optimizing the design of production parameters, can quickly obtain the optimal combination that meets specific performance requirements, not only improves the design and production efficiency of the bolt, but also eliminates the trial-and-error process in production, further reducing the production cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] To more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the accompanying drawings required for the embodiments. It should be understood that the following drawings only show some embodiments of the present application and should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0051] Figure 1 It is a schematic flow chart of an intelligent design method for a fiber composite bolt according to an embodiment of the present invention;

[0052] Figure 2 It is a schematic production flow chart of a fiber composite bolt according to an embodiment of the present invention;

[0053] Figure 3 It is a schematic diagram of a fiber composite bolt according to an embodiment of the present invention;

[0054] Figure 4 It is a schematic framework diagram of an intelligent design system for a fiber composite bolt according to an embodiment of the present invention.

[0055] Wherein: 1 - fiber yarn rack, 2 - calibration plate, 3 - main dipping tank, 4 - pre - mold, 5 - knitting machine, 6 - guiding hole, 7 - secondary dipping tank, 8 - winding machine, 9 - heating and curing mold, 10 - traction device, 11 - cutting saw, 12 - carbon - glass hybrid bolt core, 13 - knitting layer, 14 - LCP fiber winding rib. Specific Embodiments

[0056] The following will describe in detail the specific embodiments of the present invention. It should be noted that the embodiments described here are only for illustrative purposes and do not limit the present invention. In the following description, in order to provide a thorough understanding of the present invention, a large number of specific details are set forth. However, it is obvious to those of ordinary skill in the art that the present invention does not necessarily require these specific details. In other instances, well - known circuits, software, or methods are not specifically described to avoid obscuring the present invention.

[0057] Throughout the specification, the reference to "one embodiment", "an embodiment", "one example", or "an example" means that a particular feature, structure, or characteristic described in connection with that embodiment or example is included in at least one embodiment of the present invention. Thus, the phrases "in one embodiment", "in an embodiment", "one example", or "an example" appearing throughout the specification do not necessarily all refer to the same embodiment or example. In addition, the specific features, structures, or characteristics can be combined in any suitable combination and / or sub - combination in one or more embodiments or examples. Moreover, those of ordinary skill in the art should understand that the drawings provided herein are for illustrative purposes only and are not necessarily drawn to scale.

[0058] It should be noted in advance that in an optional embodiment, except for independent descriptions, the same symbols or letters appearing in all formulas have the same meanings and values.

[0059] In an optional embodiment, please refer to Figures 1 to 3 , the present invention provides an intelligent design method for fiber composite bolts, and the method includes the following steps:

[0060] S1. Use carbon fiber and glass fiber to make the core of a carbon-glass hybrid bolt, and at the same time use LCP fiber bundles to braid the surface of the carbon-glass hybrid bolt core to form a braided layer.

[0061] Among them, step S1 specifically includes the following steps:

[0062] S11. Simultaneously extract at least one carbon fiber and at least one glass fiber and straighten them, and then immerse the carbon fiber and the glass fiber in a thermosetting resin so that the fibers are bonded to each other.

[0063] Specifically, in this embodiment, at least one carbon fiber and at least one glass fiber are simultaneously extracted from the fiber yarn rack 1 by the traction device 10, and the carbon fiber and the glass fiber are straightened through the calibration plate 2 to ensure the uniform arrangement and neatness of the fiber bundle. After being straightened, the carbon fiber and the glass fiber enter the main impregnation tank 3 filled with thermosetting resin, so that the fiber bundle is immersed in the thermosetting resin, thereby bonding the fibers to each other, and at the same time providing the necessary corrosion resistance and high strength performance for the produced fiber composite bolt.

[0064] More specifically, carbon fiber and glass fiber need to be extracted from the fiber yarn rack 1 according to a preset fiber hybridization ratio, and the thermosetting resin in the main impregnation tank 3 is selected as epoxy resin or vinyl ester resin. In other optional embodiments, the thermosetting resin can also be selected according to different engineering requirements.

[0065] S12. After the carbon fiber and the glass fiber are immersed in the thermosetting resin, the carbon-glass hybrid bolt core is formed by using the carbon fiber and glass fiber hybrid pultrusion technology, and at the same time, the surface of the carbon-glass hybrid bolt core is braided with LCP fiber bundles to form the braided layer.

[0066] Specifically, in this embodiment, the fiber composite anchor rod formed by production is towed by the towing device 10, and then the fiber bundle infiltrated with thermosetting resin is pulled through the circular tubular pre-mold 4. The fiber bundle infiltrated with thermosetting resin is extruded in the pre-mold 4 to form the carbon-glass hybrid anchor rod core 12. At the same time, the knitting machine 5 will use 16 LCP fiber bundles to knit its surface to form a uniform knitting layer 13. LCP fiber, namely liquid crystal polyarylate fiber, has excellent strength and wear resistance, and can effectively improve the shear and bending resistance of the anchor rod.

[0067] Furthermore, the diameter of the pre-mold 4 will affect the diameter of the carbon-glass hybrid anchor rod core 12. Therefore, the diameter of the carbon-glass hybrid anchor rod core 12 can be changed by changing the diameter of the pre-mold 4, and then the diameter of the produced fiber composite anchor rod can be changed. In addition, when using LCP fiber bundles to fabricate the carbon-glass hybrid anchor rod core 12, the knitting angle and knitting density of the knitting machine 5 need to be set.

[0068] S2. After knitting the surface of the carbon-glass hybrid anchor rod core, use LCP fiber bundles to wind around the surface of the knitting layer to form LCP fiber winding ribs, and then obtain the fiber composite anchor rod.

[0069] Among them, step S2 specifically includes the following steps:

[0070] S21. Immerse the carbon-glass hybrid anchor rod core after knitting in thermosetting resin, and then use LCP fiber bundles to wind around the surface of the knitting layer in a spiral shape to form LCP fiber winding ribs, and then obtain a semi-finished anchor rod.

[0071] Specifically, in this embodiment, under the traction of the towing device 10, the carbon-glass hybrid anchor rod core 12 after knitting enters the secondary dipping tank 7 filled with thermosetting resin through the guide hole 6 for secondary dipping treatment to provide an attachment basis for winding the fiber. After the carbon-glass hybrid anchor rod core 12 after knitting undergoes secondary dipping treatment, it enters the winding machine 8. The winding machine 8 uses two LCP fiber bundles to wind around the knitting layer 13 in a spiral shape, and the distance between the two LCP fiber bundles remains fixed during the winding process to form the LCP fiber winding rib 14, and then obtain a semi-finished anchor rod.

[0072] Furthermore, the LCP fiber winding rib can enhance the anchoring effect of the fiber composite anchor rod in grouting. The winding angle of the LCP fiber bundle and the distance between the two LCP fiber bundles can be adjusted according to actual needs to change the mechanical biting force and embedding effect of the fiber composite anchor rod.

[0073] S22. Cure the semi-finished anchor rod under the set temperature and pressure to make the carbon-glass hybrid anchor rod core, the knitting layer and the LCP fiber winding rib tightly combined to obtain the fiber composite anchor rod.

[0074] Specifically, in this embodiment, the semi-finished anchor rod is fed into the heating and curing mold 9, and the temperature and pressure in the heating and curing mold 9 are adjusted to cure the thermosetting resin, so that the carbon and glass hybrid anchor rod core 12, the braided layer 13 and the LCP fiber winding rib 14 are tightly combined under the action of the thermosetting resin to form an integral body, and finally the fiber composite anchor rod is obtained.

[0075] Further, after the fiber composite anchor rod passes through the traction device 10, it is cut by the cutting saw 11, and fiber composite anchor rods of different lengths can be obtained. The produced fiber composite anchor rods are as Figure 3 shown.

[0076] S3. Adjust the adjustable parameters of the fiber composite anchor rod to produce a plurality of the fiber composite anchor rods, and then construct a production prediction data set of the fiber composite anchor rod.

[0077] Among them, step S3 specifically includes the following steps:

[0078] S31. Adjust the values of different adjustable parameters to generate a plurality of combinations of adjustable parameters. The adjustable parameters include the fiber hybrid ratio, the braiding angle, the braiding density, the anchor rod diameter, and the anchor rod length.

[0079] Specifically, in this embodiment, the adjustable parameters refer to the parameters that can be adjusted or designed by oneself before producing the fiber composite anchor rod.

[0080] Further, in other alternative embodiments, the adjustable parameters may further include fiber material properties, resin material properties, and fiber volume ratio, etc.

[0081] S32. Produce the corresponding fiber composite anchor rod according to the combination of adjustable parameters and test its flexural strength, tensile strength, and shear strength.

[0082] Specifically, in this embodiment, the tests of the flexural strength, tensile strength, and shear strength of the fiber composite anchor rod can all be realized by existing technical means.

[0083] S33. Use all combinations of adjustable parameters and the flexural strength, tensile strength, and shear strength of the corresponding fiber composite anchor rods to construct a production prediction data set of the fiber composite anchor rod.

[0084] S4. Construct a production parameter prediction model based on the production prediction data set of the fiber composite anchor rod, and use the production parameter prediction model to obtain the production parameter prediction value of the fiber composite anchor rod when given the anchor rod performance parameters.

[0085] Among them, the performance parameters of the anchor bolt include the diameter, length, tensile strength, and shear strength of the anchor bolt, and the production parameters of the anchor bolt include the fiber hybridization ratio, braiding angle, and braiding density. The reason for taking the length and diameter of the fiber composite anchor bolt as the data to be predicted in this embodiment is that during engineering construction, the length and diameter of the anchor bolt are restricted by its installation position. In order to produce an anchor bolt that adapts to the installation position when determining the installation position of the anchor bolt, the length and diameter of the anchor bolt also need to be used as the performance parameters of the fiber composite anchor bolt.

[0086] In other alternative embodiments, the performance parameters of the anchor bolt may further include the cost of the fiber composite anchor bolt per unit length, and the production parameters of the anchor bolt may further include fiber material properties, resin material properties, fiber volume ratio, etc.

[0087] Step S4 specifically includes the following steps:

[0088] S41. Combine the multi-strategy improved genetic algorithm and the optimized BP neural network into a combined prediction model.

[0089] Among them, step S41 specifically further includes the following steps:

[0090] S411. Introduce an adaptive weight update model into the BP neural network to obtain the optimized BP neural network.

[0091] Specifically, in this embodiment, the adaptive weight update model satisfies the following relationship:

[0092]

[0093] Among them, is the weight of the optimized BP neural network at the (i + 1)-th iteration step, is the weight of the optimized BP neural network at the i-th iteration step, , and are all algebraic expressions, , , , MSE is the mean square error of the optimized BP neural network, is the weight of the optimized BP neural network at the (i - 1)-th iteration step, is the learning rate used by the optimized BP neural network at the (i - 1)-th iteration step, is a random number greater than 0 and less than 1.

[0094] More specifically, the traditional BP neural network updates weights based on the gradient descent method. However, the weight update based on the gradient descent method has problems of slow convergence speed and being prone to falling into local minima. The adaptive weight update model provided by the present invention can limit the learning rate according to the change of the partial derivative of the weight with respect to the mean square error of the optimized BP neural network to dynamically update the weight, improve the weight update speed, and can avoid the problem that the traditional BP neural network is prone to falling into local minima when the gradient approaches 0, improve the training speed and prediction accuracy of the algorithm, and thus improve the accuracy of predicting production parameters.

[0095] S412. Use the optimized BP neural network as the main part of the model of the combined prediction model to predict production parameters, and use the multi-strategy improved genetic algorithm as the secondary part of the model of the combined prediction model to optimize the initial weights and initial thresholds of the optimized BP neural network.

[0096] Specifically, in this embodiment, when the multi-strategy improved genetic algorithm runs, it performs the following steps:

[0097] H1. Set algorithm parameters, and generate chromosomes according to the initial weights and initial thresholds of the optimized BP neural network to initialize the population.

[0098] Specifically, in this embodiment, the algorithm parameters refer to the parameters that need to be set in advance in the multi-strategy improved genetic algorithm, including population size, selection operator, crossover operator, mutation operator, starting value of crossover probability, stable value of crossover probability, starting value of mutation probability, and stable value of mutation probability. Among them, in this embodiment, the starting value of crossover probability, stable value of crossover probability, starting value of mutation probability, and stable value of mutation probability are set to 50, 0.7, 0.9, 0.01, and 0.1 in sequence, and the settings of the selection operator, crossover operator, and mutation operator can refer to the prior art.

[0099] Furthermore, the initial weights and initial thresholds of the optimized BP neural network are randomly generated. According to the initial weights and initial thresholds of the optimized BP neural network, chromosomes are generated in the form of real number coding to initialize the population. Each chromosome is an individual in the population, and each chromosome represents a possible combination of initial weights and initial thresholds. In addition, since the chromosomes are generated according to the initial weights and initial thresholds of the optimized BP neural network, the genes on the chromosomes can be divided into two types: weight genes and threshold genes. Each weight gene corresponds to a possible initial weight, and each threshold gene corresponds to a possible initial threshold.

[0100] H2. The genetic algorithm iteration starts. Calculate the fitness values of each chromosome, and then calculate the chromosome difference, and compare the calculated chromosome difference with the difference threshold.

[0101] Specifically, in this embodiment, the multi-strategy improved genetic algorithm is used to optimize the initial weights and initial thresholds of the BP neural network to improve the prediction accuracy of production parameters. The mean square error between the predicted data and the actual data of the optimized BP neural network can reflect the performance of the optimized BP neural network. Therefore, the reciprocal of the mean square error between the predicted data and the actual data of the optimized BP neural network is considered as the fitness function of the multi-strategy improved genetic algorithm. In addition, if the mean square error between the predicted data and the actual data of the optimized BP neural network is 0, the fitness value is directly taken as 100.

[0102] Furthermore, in the later stage of the operation of the genetic algorithm, the fitness of individuals to the environment increases significantly, and the chromosome difference decreases significantly, reducing the exploration ability of the algorithm in the later stage and making the algorithm prone to falling into local optimum. Therefore, in this embodiment, the chromosome difference is calculated and compared with the difference threshold to determine whether the chromosomes in the population tend to be consistent, providing data support for subsequent implementation of the chromosome change strategy to improve the exploration ability of the algorithm in the later stage. The chromosome difference satisfies the following relationship:

[0103]

[0104] where R is the chromosome difference, n is the number of chromosomes, is the fitness value of the j-th chromosome.

[0105] H3. When the chromosome difference is less than the difference threshold, use the chromosome change strategy to change the chromosomes until the chromosome difference is not less than the difference threshold.

[0106] Specifically, in this embodiment, the chromosome change strategy is used to increase the chromosome difference and avoid the algorithm falling into local optimum. Each time the chromosome change strategy is used to change the chromosomes, a new population will be obtained. After each new population is obtained, it is necessary to return to step H2 until the chromosome difference of the new population is not less than the difference threshold. The chromosome change strategy includes the following steps:

[0107] L1. Sort the chromosomes in descending order according to the fitness value to obtain a chromosome sequence, and divide the chromosome sequence into a first sequence and a second sequence. The first sequence is the first half of the chromosome sequence.

[0108] Specifically, in this embodiment, the first sequence and the second sequence are represented as and , is the j-th chromosome in the chromosome sequence, . If the number of chromosomes is even, then . If the number of chromosomes is odd, then or 。

[0109] L2. Extract a gene combination from each chromosome in the second sequence according to the set number of chromosomal gene alterations, and then construct an original gene combination sequence according to the chromosome sorting.

[0110] Specifically, in this embodiment, set the number of chromosomal gene alterations. Usually, the number of chromosomal gene alterations is set to an even number, with the weight genes and threshold genes each accounting for half. According to the set number of chromosomal gene alterations K, for any chromosome in the second sequence, randomly extract K genes from it to form the gene combination of this chromosome. The first half of the genes in each gene combination are weight genes, and the second half are threshold genes. According to the above description, an original gene combination sequence can be constructed according to the chromosome sorting in the second sequence. It is easy to know that the gene combinations in the original gene combination sequence are in one-to-one correspondence with the chromosomes in the second sequence.

[0111] Furthermore, the extracted gene combination can be expressed as , is the k-th gene extracted from the j-th chromosome, and is located at the k-th position on the j-th chromosome, , . It should be noted that when extracting genes to form gene combinations, for different chromosomes in the second sequence, the number of extracted genes is K, but the gene positions may be different, and the data corresponding to the genes at the same position in the gene combination are of the same type of data. For example, assume that the genes on each chromosome are numbered starting from 1 with positive integers. Then and both represent the same type of gene extracted from the first position on the chromosome. However, the first position on the (m + 1)-th chromosome may be the position where the 3rd gene on this chromosome is located, while the first position on the (m + 2)-th chromosome is the position where the 7th gene on this chromosome is located.

[0112] L3. Continuously adjust the sorting of the gene combinations in the original gene combination sequence to obtain a new gene combination sequence, and use the gene combination sequence with the largest difference from the original gene combination sequence as the final gene combination sequence.

[0113] Specifically, in this embodiment, continuously and randomly adjust the sorting of the gene combinations in the original gene combination sequence to obtain a new gene combination sequence. Each adjustment requires judging the difference between the new gene combination sequence and the original gene combination sequence. Finally, use the gene combination sequence with the largest difference from the original gene combination sequence as the final gene combination sequence.

[0114] More specifically, the difference between the new gene combination sequence and the original gene combination sequence is determined by the following relational expression:

[0115]

[0116] where y is the difference between the new gene combination sequence and the original gene combination sequence, is the initial weight value or initial threshold corresponding to the k-th gene in the t-th gene combination in the new gene combination sequence, is the initial weight value or initial threshold corresponding to the k-th gene in the t-th gene combination in the original gene combination sequence.

[0117] L4. According to the positions of the genes in the original gene combination sequence on the chromosome, use the final gene combination sequence to change the genes on the chromosome in the second sequence.

[0118] Specifically, in this embodiment, the genes in the gene combination in the original gene combination sequence are denoted as disrupted genes, and the gene combinations in the final gene combination sequence are sequentially corresponded to the chromosomes in the second sequence one by one. Next, for any chromosome in the second sequence, sequentially replace the disrupted genes with the genes of this chromosome in the final gene combination sequence, and the gene change of the chromosome in the second sequence can be achieved.

[0119] H4. When the chromosome difference is not less than the difference threshold, perform chromosome crossover and mutation according to the adaptive crossover probability model and the adaptive mutation probability model, and then generate new chromosomes.

[0120] Specifically, in this embodiment, the adaptive crossover probability model and the adaptive mutation probability model respectively satisfy the following relationships:

[0121]

[0122]

[0123] where P is the crossover probability of the chromosome, is the starting value of the crossover probability, is the stable value of the crossover probability, is the larger fitness value among the fitness values of the two chromosomes participating in the crossover, is the maximum fitness value of the chromosome, is the mutation probability of the chromosome, is the starting value of the mutation probability, is the stable value of the mutation probability, is the fitness value of the chromosome to be mutated.

[0124] Furthermore, the adaptive crossover probability model and the adaptive mutation probability model provided in this embodiment can automatically adjust the crossover probability and the mutation probability according to the fitness value of an individual and the overall fitness value of the population. In this way, while ensuring the global search ability of the algorithm, it can solve the problems of premature convergence and local optimal solutions existing in the existing BP neural network, and improve the prediction efficiency and accuracy of production parameters.

[0125] H5. Determine whether the maximum number of iterations is reached, and output the best chromosome when the maximum number of iterations is reached; otherwise, let the new chromosome obtained in the current iteration step enter the next iteration step.

[0126] Specifically, in this embodiment, the best chromosome output when the maximum number of iterations is reached is the combination of the best initial weight and the best initial threshold. If the maximum number of iterations is not reached, return to step H2.

[0127] S42. Use the fiber composite anchor bolt production prediction data set to train and verify the combined prediction model to obtain the production parameter prediction model.

[0128] S43. Input the given bolt performance parameters into the production parameter prediction model to obtain the production parameter prediction value of the fiber composite anchor bolt.

[0129] Specifically, in this embodiment, the production parameter prediction values include the fiber hybridization ratio prediction value, the braiding angle prediction value, and the braiding density prediction value.

[0130] S5. Produce the fiber composite anchor bolt according to the production parameter prediction value.

[0131] Specifically, in this embodiment, after obtaining the production parameter prediction value, carbon fiber and glass fiber can be drawn from the fiber yarn rack 1 according to the fiber hybridization ratio prediction value, and braiding can be performed on the core of the carbon-glass hybrid anchor bolt using the braiding machine 5 according to the braiding angle prediction value and the braiding density.

[0132] It should be noted that in some cases, the actions described in the specification can be executed in a different order and still achieve the desired result. In this embodiment, the given step order is only for making the embodiment look clearer and more convenient for explanation, rather than a limitation.

[0133] In an alternative embodiment, please refer to Figure 2 and Figure 4, the present invention also provides an intelligent design system for fiber composite anchor bolts. The intelligent design system for fiber composite anchor bolts uses the intelligent design method for fiber composite anchor bolts provided in this embodiment to improve the practicability of this method and the production efficiency of fiber composite anchor bolts. The system includes a fiber composite anchor bolt production subsystem A1, an anchor bolt test and data acquisition subsystem A2, and a production parameter judgment subsystem A3.

[0134] The fiber composite anchor bolt production subsystem A1 is used to make a carbon-glass hybrid anchor bolt core using carbon fiber and glass fiber, and at the same time use an LCP fiber bundle to braid the surface of the carbon-glass hybrid anchor bolt core to form a braided layer; after braiding the surface of the carbon-glass hybrid anchor bolt core, then use the LCP fiber bundle to wind around the surface of the braided layer to form an LCP fiber winding rib, thereby obtaining a fiber composite anchor bolt.

[0135] Specifically, in this embodiment, the fiber composite anchor bolt production subsystem A1 includes a fiber yarn rack 1, a calibration plate 2, a main impregnation tank 3, a pre-mold 4, a braiding machine 5, a guide hole 6, a secondary impregnation tank 7, a winding machine 8, a heating and curing mold 9, a traction device 10, and a cutting saw 11. The specific operation process of the fiber composite anchor bolt production subsystem A1 is shown in steps S1 and S2.

[0136] The anchor bolt test and data acquisition subsystem A2 is used to adjust the production parameters of the fiber composite anchor bolts to produce multiple fiber composite anchor bolts, and then construct a production prediction data set for fiber composite anchor bolts.

[0137] Specifically, in this embodiment, the anchor bolt test and data acquisition subsystem A2 executes the content described in step S3. The anchor bolt test and data acquisition subsystem A2 includes a parameter setting module, a data acquisition module, a data transmission module, and various instruments for testing the tensile strength and shear strength of fiber composite anchor bolts. First, the parameter setting module randomly generates multiple production parameter combinations, and then relevant personnel adjust the fiber composite anchor bolt production subsystem A1 according to the generated production parameter combinations to produce multiple fiber composite anchor bolts. After that, relevant personnel perform performance tests on all fiber composite anchor bolts and input the obtained data into the data acquisition module to generate a production prediction data set for fiber composite anchor bolts in combination with the production parameter combinations. Finally, the data transmission module transmits the production prediction data set for fiber composite anchor bolts to the production parameter judgment subsystem A3 through the Internet of Things.

[0138] The production parameter judgment subsystem A3 is used to construct a production parameter prediction model based on the production prediction data set for fiber composite anchor bolts, and when given the performance parameters of the anchor bolt, use the production parameter prediction model to obtain the production parameter prediction value of the fiber composite anchor bolt, so that the fiber composite anchor bolt production subsystem A1 produces the fiber composite anchor bolt according to the production parameter prediction value.

[0139] Specifically, in this embodiment, the production parameter judgment subsystem A3 specifically performs the content described in step S4. The production parameter judgment subsystem A3 includes a data input module, a data processing module and a data output module. The data input module receives the fiber composite anchor production prediction data set obtained by the anchor test and data acquisition subsystem A2 through the Internet of Things. The data processing module uses the production parameter prediction model to obtain the production parameter prediction value of the fiber composite anchor when the anchor performance parameters are given. The data output module outputs the production parameter prediction value.

[0140] After obtaining the predicted values ​​of the production parameters, the relevant personnel adjusted the fiber composite anchor production subsystem A1 so that it could produce the fiber composite anchor according to the predicted values ​​of the production parameters.

[0141] In summary, firstly, the present invention forms a carbon-glass hybrid anchor rod core with high strength and corrosion resistance through mixed pultrusion of carbon fiber and glass fiber, which effectively improves the mechanical properties of the anchor rod, reduces the cost as much as possible, and improves the economy of the anchor rod. Secondly, the present invention uses LCP fiber to weave and wind on the surface of the carbon-glass hybrid anchor rod core to form a woven layer and LCP fiber winding ribs, further enhancing the shear resistance and bending resistance of the anchor rod, and can improve the anchoring effect of the anchor rod. Thirdly, the fiber composite anchor rod produced by the present invention uses carbon fiber, glass fiber and LCP fiber as the main materials, has excellent corrosion resistance and creep resistance, overcomes the problem of easy rust and performance attenuation of traditional steel anchor rods, and enables the anchor rod to maintain long-term mechanical properties and structural stability in various harsh environments. Finally, the present invention constructs an accurate and reliable production parameter prediction model for optimizing the design of production parameters, which can quickly obtain the optimal combination that meets specific performance requirements, not only improving the design and production efficiency of the anchor rod, but also eliminating the trial and error process in production, and further reducing production costs.

[0142] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. These modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and specification of the present invention.

Claims

1. A fiber composite anchor intelligent design method, characterized in that: The steps include: Using carbon fiber and glass fiber to make a carbon-glass hybrid anchor rod core, and using LCP fiber bundles to weave the surface of the carbon-glass hybrid anchor rod core to form a braided layer; After the surface of the carbon-glass hybrid anchor core is braided, an LCP fiber bundle is wound on the surface of the braided layer to form an LCP fiber winding rib, thereby obtaining a fiber composite anchor; adjusting the adjustable parameters of the fiber composite anchor to produce a plurality of the fiber composite anchors, thereby constructing a fiber composite anchor production prediction data set; Combine the multi-strategy improved genetic algorithm and optimized BP neural network into a combined prediction model; Using the fiber composite anchor production prediction data set to train and verify the combined prediction model to obtain a production parameter prediction model; Inputting the given anchor performance parameters into the production parameter prediction model, thereby obtaining the production parameter prediction values ​​of the fiber composite anchor; The anchor performance parameters include anchor diameter, anchor length, bending strength, tensile strength and shear strength, and the production parameters of the fiber composite anchor include fiber mixing ratio, weaving angle and weaving density; The fiber composite anchor is produced according to the predicted production parameter values.

2. The fiber composite anchor intelligent design method according to claim 1, characterized in that: The method of using carbon fiber and glass fiber to make a carbon-glass hybrid anchor rod core and using LCP fiber bundles to weave the surface of the carbon-glass hybrid anchor rod core to form a braided layer comprises the following steps: At the same time, at least one carbon fiber and at least one glass fiber are extracted and straightened, and then the carbon fiber and the glass fiber are impregnated with a thermosetting resin so that the fibers are bonded to each other; After the carbon fiber and the glass fiber are impregnated with thermosetting resin, the carbon-glass hybrid anchor core is formed by hybrid pultrusion technology of the carbon fiber and the glass fiber, and the surface of the carbon-glass hybrid anchor core is woven with LCP fiber bundles to form the woven layer.

3. The intelligent design method for fiber composite anchor according to claim 1, characterized in that: After the surface of the carbon-glass hybrid anchor core is braided, LCP fiber bundles are wound on the surface of the braided layer to form LCP fiber winding ribs, thereby obtaining a fiber composite anchor, which includes the following steps: The braided carbon-glass hybrid anchor rod core is impregnated with a thermosetting resin, and then an LCP fiber bundle is wound around the surface of the braided layer in a threaded shape to form an LCP fiber-wound rib, thereby obtaining a semi-finished anchor rod; The anchor rod semi-finished product is cured at a set temperature and pressure, so that the carbon-glass hybrid anchor rod core, the braided layer and the LCP fiber winding rib are tightly combined to obtain the fiber composite anchor rod.

4. The fiber composite anchor intelligent design method according to claim 1, characterized in that: The step of adjusting the adjustable parameters of the fiber composite anchor to produce a plurality of the fiber composite anchors and then constructing a fiber composite anchor production prediction data set comprises the following steps: Adjusting the values ​​of different adjustable parameters to generate a plurality of adjustable parameter combinations, wherein the adjustable parameters include fiber mixing ratio, weaving angle, weaving density, anchor rod diameter and anchor rod length; Produce the corresponding fiber composite anchor rod according to the adjustable parameter combination and test its bending strength, tensile strength and shear strength; A fiber-reinforced anchor production prediction dataset is constructed using all adjustable parameter combinations and the corresponding flexural strength, tensile strength, and shear strength of the fiber-reinforced anchor.

5. The fiber composite anchor intelligent design method according to claim 1, characterized in that: The multi-strategy improved genetic algorithm performs the following steps when running: Setting algorithm parameters, and generating chromosomes according to the initial weights and initial thresholds of the optimized BP neural network to initialize the population; The genetic algorithm iteration starts, the fitness value of each chromosome is calculated, and then the chromosome difference is calculated, and the calculated chromosome difference is compared with the difference threshold; When the chromosome difference is less than the difference threshold, the chromosome is changed using a chromosome change strategy until the chromosome difference is not less than the difference threshold; When the chromosome difference is not less than the difference threshold, performing chromosome crossover and mutation according to an adaptive crossover probability model and an adaptive mutation probability model to generate a new chromosome; It is determined whether the maximum number of iterations has been reached, and when the maximum number of iterations has been reached, the best chromosome is output; otherwise, the new chromosome obtained in the current iteration step is entered into the next iteration step.

6. The fiber composite anchor intelligent design method according to claim 5, characterized in that: The chromosome difference, the adaptive crossover probability model and the adaptive mutation probability model respectively satisfy the following relationships: ， ， ， Wherein, R is the chromosome difference, n is the number of chromosomes, is the fitness value of the jth chromosome, P is the crossover probability of the chromosome, is the starting value of the crossover probability, is the stable value of crossover probability, is the larger fitness value of the two chromosomes involved in the crossover, is the maximum fitness value of the chromosome, is the probability of chromosome mutation, is the starting value of mutation probability, is the stable value of mutation probability, is the fitness value of the chromosome that needs to be mutated, is the larger fitness value of the two chromosomes undergoing mutation.

7. The fiber composite anchor intelligent design method according to claim 5, characterized in that: The chromosome alteration strategy comprises the following steps: Sort the chromosomes in descending order of the fitness values ​​to obtain a chromosome sequence, and divide the chromosome sequence into a first sequence and a second sequence, wherein the first sequence is the first half of the chromosome sequence; Extracting a gene combination from each chromosome in the second sequence according to the set number of chromosome gene changes, and then constructing an original gene combination sequence according to the chromosome order; Continuously adjusting the order of gene combinations in the original gene combination sequence to obtain a new gene combination sequence, and taking the gene combination sequence with the greatest difference from the original gene combination sequence as the final gene combination sequence; According to the location of the genes in the original gene combination sequence on the chromosome, the final gene combination sequence is used to change the genes on the chromosome in the second sequence.

8. The fiber composite anchor intelligent design method according to claim 1, characterized in that: The method of combining the multi-strategy improved genetic algorithm and the optimized BP neural network into a combined prediction model comprises the following steps: An adaptive weight update model is introduced into the BP neural network to obtain the optimized BP neural network, and the adaptive weight update model satisfies the following relationship: ， in, is the weight of the optimized BP neural network at the i+1th iteration step, is the weight of the optimized BP neural network at the i-th iteration step, , and are all algebraic expressions, , , , MSE is the mean square error of the optimized BP neural network, is the weight of the optimized BP neural network at the i-1th iteration step, is the learning rate used by the optimized BP neural network in the i-1th iteration step, is a random number greater than 0 and less than 1; The optimized BP neural network is used as the main part of the combined prediction model to predict production parameters, and the multi-strategy improved genetic algorithm is used as the secondary part of the combined prediction model to optimize the initial weight and initial threshold of the optimized BP neural network.

9. A fiber composite anchor intelligent design system, the fiber composite anchor intelligent design system using the fiber composite anchor intelligent design method according to any one of claims 1 to 8, characterized in that: include: A fiber composite anchor rod production subsystem, the fiber composite anchor rod production subsystem is used to use carbon fiber and glass fiber to make a carbon-glass hybrid anchor rod core, and use LCP fiber bundles to weave the surface of the carbon-glass hybrid anchor rod core to form a braided layer; after weaving the surface of the carbon-glass hybrid anchor rod core, the LCP fiber bundles are wound on the surface of the braided layer to form LCP fiber winding ribs, thereby obtaining a fiber composite anchor rod; An anchor rod testing and data acquisition subsystem, the anchor rod testing and data acquisition subsystem is used to adjust the adjustable parameters of the fiber composite anchor rod to produce a plurality of the fiber composite anchor rods, and then construct a fiber composite anchor rod production prediction data set; A production parameter judgment subsystem, the production parameter judgment subsystem is used to combine a multi-strategy improved genetic algorithm and an optimized BP neural network into a combined prediction model; use the fiber composite anchor production prediction data set to train and verify the combined prediction model to obtain a production parameter prediction model; input given anchor performance parameters into the production parameter prediction model to obtain production parameter prediction values ​​of the fiber composite anchor; the anchor performance parameters include anchor diameter, anchor length, bending strength, tensile strength and shear strength, and the production parameters of the fiber composite anchor include fiber mixing ratio, braiding angle and braiding density; use the fiber composite anchor production subsystem to produce the fiber composite anchor according to the production parameter prediction values.

Images