Method for optimizing performance of desert sand concrete u-shaped channel based on intelligent algorithm

Abstract

The application relates to the technical field of desert sand concrete proportioning optimization, and discloses a desert sand concrete U-shaped channel performance optimization method based on an intelligent algorithm, which comprises the following steps: generating an initial proportioning set based on a component list through a multi-objective genetic algorithm; the component list comprises a list of components used for constructing a desert sand concrete U-shaped channel; performing the k+1th round of iteration optimization on the kth optimization proportioning set based on the kth objective function and a constraint condition through the multi-objective genetic algorithm to obtain the k+1th optimization proportioning set; the kth objective function is obtained by adjusting the weight of the k-1th objective function based on the kth difference degree; the constraint condition is at least associated with the climate state of Ningxia; k is an integer greater than 1; and if the k+1th difference degree is less than a degree threshold value, the k+1th optimization proportioning set is determined as the target proportioning set corresponding to the component list. The scheme can improve the ability of the desert sand concrete U-shaped channel to withstand climate erosion in Ningxia.

Description

Technical Field

[0001] This application relates to the field of desert sand concrete mix design optimization technology, specifically to, but not limited to, a method for optimizing the performance of U-shaped channels in desert sand concrete based on intelligent algorithms. Background Technology

[0002] The Ningxia Plain is located in the inland northwest of China and is a typical arid and semi-arid climate zone with an average annual precipitation of less than 200 mm. Therefore, in this region, U-shaped canal lining, as a core component of farmland water conservancy infrastructure, plays an important role in water conveyance and irrigation. At the same time, the region experiences cold winters, which causes the canal foundation soil to undergo frequent freeze-thaw cycles with the changing seasons.

[0003] In practical applications, traditional orthogonal experimental methods and polynomial damage models are mainly used to optimize concrete mix design and predict freeze-thaw damage for U-shaped channels. However, orthogonal experimental methods rely on manually designed variable combinations, cannot globally search for the optimal solution, and the number of experiments increases exponentially with the number of variables. Polynomial damage models rely on linear relationships, but actual freeze-thaw damage exhibits a phased nonlinear growth. As a result, the component mix design of U-shaped channels optimized by the above methods cannot cope with the complex environmental changes in Ningxia. Summary of the Invention

[0004] Based on the above technical problems, this application provides a method for optimizing the performance of desert sand concrete U-shaped channels based on intelligent algorithms. This method can specifically improve the performance of desert sand concrete U-shaped channels, thereby enabling them to resist the erosion caused by the complex environmental changes in Ningxia and improving the stability, robustness, and load-bearing capacity of the desert sand concrete U-shaped channels in Ningxia.

[0005] The technical solution provided in this application is as follows: This application provides a method for optimizing the performance of U-shaped channels made of desert sand concrete based on intelligent algorithms, including: An initial mix design set is generated based on a component list using a multi-objective genetic algorithm; wherein, the component list includes a list of components used to construct the desert sand concrete U-shaped channel; and the initial mix design set includes a set of initialized component ratios among the components in the component list. The multi-objective genetic algorithm, based on the k-th objective function and constraints, performs the (k+1)-th round of iterative optimization on the k-th optimal ratio set to obtain the (k+1)-th optimal ratio set. The k-th objective function is obtained by adjusting the weights of the (k-1)-th objective function based on the k-th degree of difference. The k-th degree of difference includes the difference between the k-th predicted score set and the k-th measured score set corresponding to the k-th optimal ratio set. The k-th degree of difference is greater than or equal to a threshold. The k-th optimal ratio set is associated with the initial ratio set. The constraints are at least associated with the climate conditions of Ningxia region. k is an integer greater than 1. If the degree of difference at the (k+1)th point is less than the degree threshold, the (k+1)th optimized ratio set is determined as the target ratio set corresponding to the component list; wherein, the degree of difference at the (k+1)th point includes the difference between the (k+1)th predicted score set and the (k+1)th measured score set corresponding to the (k+1)th optimized ratio set.

[0006] The method for optimizing the performance of U-shaped channels in desert sand concrete based on intelligent algorithms provided in this application has at least the following beneficial effects: The performance optimization method for U-shaped channels of desert sand concrete based on intelligent algorithms provided in this application embodiment generates an initial mix design set based on a component list using a multi-objective genetic algorithm. The component list includes a list of components used to construct the U-shaped channel of desert sand concrete, and the initial mix design set includes the initial mix design sets between each component in the component list. This improves the correlation between the initial mix design set and the component list of the U-shaped channel of desert sand concrete. Furthermore, generating the initial mix design set using a multi-objective genetic algorithm enhances the randomness and diversity of the component ratios in the initial mix design set. The k-th objective function of the multi-objective genetic algorithm is obtained by adjusting the weights of the (k-1)-th objective function based on the k-th degree of difference. The k-th degree of difference includes the difference between the k-th predicted score set corresponding to the k-th optimized mix design set and the k-th measured score set. The k-th optimized mix design set is associated with the initial mix design set. Thus, through the k-th degree of difference indirectly associated with the initial mix design set, targeted and precise weight adjustment of the k-th objective function of the multi-objective genetic algorithm can be achieved, ensuring that the k-th objective function satisfies the requirements of the multi-objective genetic algorithm. The process of iterative optimization involves the selection and optimization of proportions. Based on this, a multi-objective genetic algorithm is used to perform the (k+1)th round of iterative optimization on the k-th optimized proportion set, based on the k-th objective function and constraints, to obtain the (k+1)-th optimized proportion set. The constraints are at least related to the climate conditions of Ningxia, enabling targeted optimization of the k-th optimized proportion set. This improves the component proportions in the (k+1)-th optimized proportion set and enhances its ability to resist erosion corresponding to the climate conditions of Ningxia. On the other hand, if the degree of difference in the (k+1)-th set is less than a threshold, [the process will proceed as planned]. The (k+1)th optimized mix ratio set is determined as the target mix ratio set corresponding to the component list, and the (k+1)th difference degree includes the difference between the (k+1)th predicted score set and the (k+1)th measured score set corresponding to the (k+1)th optimized mix ratio set. In this way, the performance of the desert sand concrete U-shaped channel corresponding to the component ratio in the (k+1)th optimized mix ratio set can be improved in a targeted manner, so that the desert sand concrete U-shaped channel can resist the erosion of complex environmental changes in Ningxia, thereby improving the stability, robustness and bearing capacity of the desert sand concrete U-shaped channel in Ningxia. Attached Figure Description

[0007] Figure 1 A flowchart illustrating the performance optimization method for U-shaped channels in desert sand concrete based on intelligent algorithms provided in this application embodiment; Figure 2 A schematic diagram of the architecture of the desert sand concrete U-shaped channel performance optimization method provided in the embodiments of this application. Detailed Implementation

[0008] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0009] It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit this application.

[0010] The Ningxia Plain, with its flat terrain along both banks of the Yellow River, has historically formed a large area of ​​gravity-fed irrigation, and remains an important commodity grain base and specialty agricultural base in my country. At the same time, the Ningxia Plain is located in the inland northwest of China, a typical arid and semi-arid climate zone with an average annual precipitation of less than 200 millimeters. Therefore, in this region, U-shaped canal linings, as a core component of farmland water conservancy infrastructure, play a vital role in water conveyance and irrigation.

[0011] However, the lining concrete of U-shaped channels in Ningxia faces severe performance challenges: on the one hand, Ningxia is located in a seasonally frozen soil region, and the canal foundation soil experiences a freezing period of 80 to 100 days in winter, leading to severe frost heave damage to the U-shaped channel lining slabs; on the other hand, the soil salinity is high in some areas of Ningxia, which also exacerbates the deterioration rate and extent of the U-shaped channel concrete; for some U-shaped channels with excessive desert sand replacement rates, it also leads to a decrease in the interfacial bond between aggregates and cementitious materials, resulting in increased porosity and decreased impermeability; at the same time, desert sand concrete is significantly more brittle than ordinary concrete, which causes the tensile strength and fracture toughness of the U-shaped channel to decrease by more than 38% after freeze-thaw cycles.

[0012] All of the above factors combined have led to short service life and prominent leakage problems in the concrete lining of U-shaped channels in Ningxia, which has restricted the agricultural water supply guarantee capacity of the irrigation area.

[0013] To address the aforementioned technical challenges, traditional orthogonal experimental design and polynomial damage models have been developed to optimize concrete mix design and predict freeze-thaw damage in U-shaped channels. However, the orthogonal experimental design relies on manually designed variable combinations, making it impossible to globally search for the optimal solution, and the number of experiments increases exponentially with the number of variables. The polynomial damage model relies on linear relationships, but actual freeze-thaw damage exhibits a phased, non-linear growth, resulting in the composition ratio of U-shaped channels optimized using these methods being unable to cope with the complex environmental changes in Ningxia.

[0014] Based on the above technical problems, this application provides a method for optimizing the performance of U-shaped channels made of desert sand concrete based on intelligent algorithms. Figure 1 A flowchart illustrating the performance optimization method for U-shaped channels in desert sand concrete based on intelligent algorithms provided in this application embodiment is shown below. Figure 1 As shown, the method may include the following steps: Step 101: Generate an initial set of proportions based on the component list using a multi-objective genetic algorithm.

[0015] The component list includes a list of components used to construct the desert sand concrete U-shaped channel; the initial mix design set includes a set of initial component proportions among the components in the component list.

[0016] In some embodiments, the component list may include components such as desert sand, concrete, adhesive, and water; for example, the types and / or quantities of components in the component list may be adjusted according to the specific needs of constructing a desert sand concrete U-shaped channel.

[0017] In some embodiments, the initial set of proportions may include an initialized set of proportions generated by a multi-objective genetic algorithm.

[0018] In some embodiments, a multi-objective genetic algorithm may include a non-dominated sorting genetic algorithm II (NSGA-II); its optimization objective may include the proportion of each component in the component list.

[0019] In some embodiments, the initial set of proportions can be generated in the following ways: Set parameters such as population size, maximum number of iterations, crossover probability, and mutation probability. Then, input the above parameters and component list into the multi-objective genetic algorithm. The multi-objective genetic algorithm processes the above parameters and component list through fast non-dominated sorting and crowding calculation to generate an initial matching set. The population size can be 100, the maximum number of generations can be 60, the crossover probability can be 0.7, and the mutation probability can be 0.03.

[0020] Step 102: Using a multi-objective genetic algorithm based on the k-th objective function and constraints, perform the (k+1)-th round of iterative optimization on the k-th optimal ratio set to obtain the (k+1)-th optimal ratio set.

[0021] The k-th objective function is obtained by adjusting the weights of the (k-1)-th objective function based on the k-th degree of difference; the k-th degree of difference includes the degree of difference between the k-th predicted score set corresponding to the k-th optimized allocation set and the k-th measured score set; the k-th degree of difference is greater than or equal to the degree threshold; the k-th optimized allocation set is associated with the initial allocation set; the constraints are at least associated with the climate state of Ningxia region; k is an integer greater than 1.

[0022] In some embodiments, the set of k-th predicted scores may include a set of performance scores obtained by predicting the performance of virtual channels corresponding to the set of k-th optimized allocation ratios; correspondingly, the set of k-th predicted scores can be determined in the following ways: The environmental data of the desert sand concrete U-shaped channel is determined, and then the environmental data is input into the state prediction model so that the state prediction model can predict the performance of the virtual channel corresponding to the component ratio in the k-th optimal mix set based on the environmental data, and obtain the k-th prediction score set. The virtual channel can include the virtualized desert sand concrete U-shaped channel constructed by the state prediction model based on the component ratio in the k-th optimal mix set, and the environmental data can include temperature, humidity, light, wind speed, and the duration of the above data.

[0023] In some embodiments, the set of k-th measured scores may include a set of performance scores obtained by testing the performance of the physical channel corresponding to the set of k-th optimized mix proportions; wherein, the physical channel may include a model of an actual desert sand concrete U-shaped channel constructed according to the component proportions in the set of k-th optimized mix proportions.

[0024] Accordingly, the set of k-th measured scores can be obtained in the following way: A test environment is constructed based on environmental parameters, and the performance parameters of the physical channel are continuously tested multiple times in the test environment to obtain the set of k-th measured scores.

[0025] In some embodiments, the set of k-th predicted scores and the set of k-th measured scores may each include performance scores of multiple types or dimensions.

[0026] In some embodiments, the k-th degree of difference may include M difference scores; where M can characterize the number of types or dimensions of performance scores contained in the k-th predicted score set and the k-th measured score set, and M is an integer greater than 1; correspondingly, the k-th degree of difference can be obtained in the following way: The difference between the predicted score of type m in the set of predicted scores of k and the measured score of type m in the set of measured scores of k is determined as the m-th difference score after the k-th round of iterative optimization; when m takes values ​​from 1 to M, the first difference score to the M-th difference score after the k-th round of iterative optimization are statistically analyzed to obtain the k-th difference degree; where m is an integer greater than or equal to 1 and less than or equal to M.

[0027] In some embodiments, the climate status may include the status of extreme weather in Ningxia; for example, extreme weather may include at least one of the following: duration of cold winter weather, wind speed, temperature, and freeze-thaw cycle frequency.

[0028] In some embodiments, the constraints may also include the salinity, pH, and moisture content of the soil used for setting up desert sand concrete U-shaped channels in Ningxia.

[0029] In some embodiments, the constraints may include at least one condition constraining the frost resistance, compressive strength, impermeability, and maintenance cost of desert sand concrete U-shaped channels under the climate conditions corresponding to the climate state; for example, the compressive strength may include a compressive strength greater than or equal to the C25 standard, the frost resistance may include a frost resistance grade greater than or equal to F150, and the impermeability may include an impermeability grade greater than or equal to P6.

[0030] In some embodiments, when the multi-objective genetic algorithm performs the first round of iterative optimization, its input data can be an initial ratio set, an initial objective function, and constraints, and its output data can be a first optimized ratio set. At this time, by adjusting the weights of the initial objective function, the first objective function can be obtained. For example, the first optimized ratio set can be processed through the first objective function to determine the superiority or degree of improvement of various ratio sets in the first optimized ratio set. Then, based on the above superiority or degree of improvement, the first generation of non-dominated sorting is performed on the first optimized ratio set, so as to perform the second round of iterative optimization on the result of the non-dominated sorting through the multi-objective genetic algorithm.

[0031] In some embodiments, the k-th objective function can be obtained in the following way: If the degree of difference of the kth rank is greater than the degree threshold, then at least some of the weights of the (k-1)th objective function are adjusted based on the degree of difference of the kth rank to obtain the kth objective function.

[0032] In some embodiments, the (k+1)th optimal ratio set can be obtained in the following way: By using the two-layer evaluation structure corresponding to the Pareto dominance relationship associated by the multi-objective genetic algorithm, the k-th optimal ratio set is processed based on the constraints to obtain the Pareto optimal solution set for this round of iteration optimization, and the Pareto optimal solution set is determined as the k+1-th optimal ratio set.

[0033] Specifically, through a two-layer evaluation structure, the group allocation ratios in the k-th optimal allocation set can first be sorted using a fast non-dominated sorting method. The priority of convergence can be established based on the dominance relationship between various group allocation ratios, thereby driving the population to approach the Pareto front. Secondly, crowding distance calculation is introduced within the same front layer as a measure of the diversity of various group allocation ratios relative to the constraints. Then, the group allocation ratios in the k-th optimal allocation set are traversed based on the binary criterion of front layer first and then crowding distance, thus obtaining the k+1-th optimal allocation set.

[0034] It should be noted that each round of iteration optimization in a multi-objective genetic algorithm may include multiple iteration optimization processes.

[0035] Step 103: If the degree of difference at the (k+1)th step is less than the degree threshold, the (k+1)th optimized ratio set is determined as the target ratio set corresponding to the component list.

[0036] The degree of difference at k+1 includes the difference between the predicted score set at k+1 corresponding to the optimized ratio set at k+1 and the measured score set at k+1.

[0037] Accordingly, if the degree of difference at the (k+1)th step is greater than or equal to the degree threshold, the operation of determining the (k+1)th optimized ratio set as the target ratio set can be omitted.

[0038] In some embodiments, if the degree of difference at the (k+1)th time is greater than or equal to the degree threshold, the (k+2)th round of iterative optimization can be performed using a multi-objective genetic algorithm based on the (k+1)th optimized ratio set.

[0039] In some embodiments, if the degree of difference at the (k+1)th rank is less than the degree threshold, the iterative optimization process performed by the target genetic algorithm can be stopped.

[0040] In some embodiments, the performance indicators represented by the target mix can resist the frequent freeze-thaw cycles of erosion in the Ningxia region.

[0041] As can be seen from the above, in the performance optimization method for U-shaped desert sand concrete channels based on intelligent algorithms provided in this application embodiment, an initial mix design set is generated based on a component list using a multi-objective genetic algorithm. The component list includes a list of components used to construct the U-shaped desert sand concrete channel, and the initial mix design set includes an initial mix design set between each component in the component list. This improves the correlation between the initial mix design set and the component list of the U-shaped desert sand concrete channel. Moreover, generating the initial mix design set using a multi-objective genetic algorithm can improve the randomness and diversity of the component ratios in the initial mix design set. Furthermore, the k-th objective function of the multi-objective genetic algorithm is obtained by adjusting the weights of the (k-1)-th objective function based on the k-th degree of difference. The k-th degree of difference includes the degree of difference between the k-th predicted score set corresponding to the k-th optimized mix design set and the k-th measured score set. The k-th optimized mix design set is associated with the initial mix design set. Thus, through the k-th degree of difference indirectly associated with the initial mix design set, targeted and precise weight adjustment of the k-th objective function of the multi-objective genetic algorithm can be achieved, thereby enabling the k-th objective function to satisfy multiple objectives. The algorithm addresses the requirements for ratio selection and optimization during the iterative optimization process of a standard genetic algorithm. Based on this, a multi-objective genetic algorithm, using the k-th objective function and constraints, performs the (k+1)-th round of iterative optimization on the k-th optimized ratio set, obtaining the (k+1)-th optimized ratio set. The constraints are at least related to the climate conditions of Ningxia, enabling targeted optimization of the k-th optimized ratio set. This improves the component ratios in the (k+1)-th optimized ratio set and enhances its ability to resist erosion corresponding to Ningxia's climate conditions. Furthermore, if the (k+1)-th difference is less than a threshold value... The (k+1)th optimized mix ratio set is determined as the target mix ratio set corresponding to the component list, and the (k+1)th difference degree includes the difference between the (k+1)th predicted score set and the (k+1)th measured score set corresponding to the (k+1)th optimized mix ratio set. In this way, the performance of desert sand concrete U-shaped channels corresponding to the component ratios in the (k+1)th optimized mix ratio set can be improved in a targeted manner, thereby enhancing the erosion resistance of desert sand concrete U-shaped channels in Ningxia against complex environmental changes, and further improving the stability, robustness and bearing capacity of the performance of desert sand concrete U-shaped channels in Ningxia.

[0042] Based on the foregoing embodiments, in the performance optimization method for desert sand concrete U-shaped channels based on intelligent algorithms provided in this application embodiment, the (k+1)th predicted score set includes the (k+1)th predicted frost resistance score, the (k+1)th predicted permeability score, and the (k+1)th predicted cost score; correspondingly, the above method can also perform the following steps: Step A1: Determine the (k+1)th predicted data set corresponding to the (k+1)th optimized ratio set.

[0043] The (k+1)th prediction dataset includes the (k+1)th predicted mass loss rate, the (k+1)th predicted dynamic modulus loss rate, the (k+1)th predicted ultrasonic loss rate, the (k+1)th predicted seepage height, the (k+1)th predicted porosity, the (k+1)th predicted cement content, and the (k+1)th predicted admixture content.

[0044] In some embodiments, the (k+1)th predicted mass loss rate can be obtained by simulation or modeling under virtual climate conditions corresponding to the climate state of the summer region, based on the component ratios in the (k+1)th optimized ratio set of the virtual channel, representing the percentage of mass loss of the virtual channel per unit time.

[0045] In some embodiments, the (k+1)th predicted dynamic modulus loss rate can characterize the ratio of the predicted dynamic modulus before and after virtual channel damage under virtual climate conditions; for example, the (k+1)th predicted dynamic modulus loss rate can quantify the degree of crack propagation and structural degradation inside the virtual channel under virtual climate conditions.

[0046] In some embodiments, the (k+1)th predicted ultrasonic loss rate can characterize the loss difference rate of ultrasonic transmission in a virtual channel; for example, the loss difference rate can correspond to the component differences at different locations in the channel.

[0047] In some embodiments, the k+1th predicted seepage height may include the extent and intensity of water leakage from the virtual channel during irrigation.

[0048] In some embodiments, the (k+1)th predicted porosity may include the proportion of voids in the virtual channel to the total volume of the virtual channel, which is closely related to water leakage in the virtual channel.

[0049] In some embodiments, the (k+1)th predicted cement content may include the cement content in the virtual channel corresponding to the component proportions in the (k+1)th optimized mix proportion set.

[0050] In some embodiments, the (k+1)th predicted admixture content may include the admixture content in the virtual channel corresponding to the component ratio in the (k+1)th optimized ratio set.

[0051] In some embodiments, the (k+1)th prediction dataset can be determined in the following way: A simulation environment is constructed based on the climate conditions and the (k+1)th optimized mix proportion set, and a virtual channel is built. Then, static predictions are performed on the virtual channel in the simulation environment to obtain the (k+1)th predicted ultrasonic loss rate, the (k+1)th predicted seepage height, the (k+1)th predicted porosity, the (k+1)th predicted cement content, and the (k+1)th predicted admixture content. At the same time, dynamic predictions are performed on the virtual channel in the simulation environment to obtain the (k+1)th predicted mass loss rate and the (k+1)th predicted dynamic modulus of elasticity loss rate.

[0052] Step A2: Process the (k+1)th predicted mass loss rate, the (k+1)th predicted dynamic modulus loss rate, and the (k+1)th predicted ultrasonic loss rate to obtain the (k+1)th predicted antifreeze score.

[0053] In some embodiments, the (k+1)th predicted antifreeze score may include the ability or degree to which a virtual channel maintains its channel body stability during a freezing-to-thawing cycle under extreme cold weather conditions.

[0054] In some embodiments, the (k+1)th predicted antifreeze score can be obtained in the following way: The predicted mass loss rate, predicted dynamic modulus loss rate, and predicted ultrasonic loss rate of the (k+1)th prediction are statistically analyzed to obtain the predicted antifreeze score of the (k+1)th prediction.

[0055] Step A3: Process the (k+1)th predicted seepage height and the (k+1)th predicted porosity to obtain the (k+1)th predicted permeability score.

[0056] In some embodiments, the k+1th predicted permeability score can characterize the range and / or extent of leakage in the virtual channel in a simulated or analog environment.

[0057] In some embodiments, the (k+1)th predicted penetration score can be obtained in the following way: The predicted seepage height of the (k+1)th predicted porosity is corrected to obtain the predicted permeability score of the (k+1)th predicted porosity.

[0058] Step A4: Process the (k+1)th predicted cement content and the (k+1)th predicted admixture content to obtain the (k+1)th predicted cost score.

[0059] In some embodiments, the (k+1)th predicted cost score can characterize the economic costs of cement and admixtures incurred for constructing and / or maintaining the virtual channel.

[0060] In some embodiments, the (k+1)th predicted cost score can be obtained in the following way: Determine the first cost per unit mass of cement and the second cost per unit weight of admixture. Then, based on the first cost and the (k+1)th predicted cement content, determine the (k+1)th predicted cement cost. Based on the second cost and the (k+1)th predicted admixture content, determine the (k+1)th predicted admixture cost. Finally, the sum of the (k+1)th predicted cement cost and the (k+1)th predicted admixture cost is determined as the (k+1)th predicted cost score.

[0061] As can be seen from the above, in the performance optimization method for U-shaped channels of desert sand concrete based on intelligent algorithms provided in this application embodiment, the (k+1)th predicted data set corresponding to the (k+1)th optimized mix proportion set is determined. The (k+1)th predicted data set includes the (k+1)th predicted mass loss rate, the (k+1)th predicted dynamic modulus of elasticity loss rate, the (k+1)th predicted ultrasonic loss rate, the (k+1)th predicted seepage height, the (k+1)th predicted porosity, the (k+1)th predicted cement content, and the (k+1)th predicted admixture content. This improves the comprehensiveness and richness of the (k+1)th predicted data set. Based on this, the (k+1)th predicted mass loss rate, the (k+1)th predicted dynamic modulus of elasticity loss rate, and the (k+1)th predicted ultrasonic loss rate are processed to obtain the (k+1)th predicted frost resistance score, which can improve the (k+1)th predicted... The accuracy of the frost resistance score is improved; furthermore, processing the (k+1)th predicted seepage height and (k+1)th predicted porosity to obtain the (k+1)th predicted permeability score can improve the accuracy of the (k+1)th predicted permeability score; simultaneously, processing the (k+1)th predicted cement content and (k+1)th predicted admixture content to obtain the (k+1)th predicted cost score can improve the accuracy of the (k+1)th predicted cost score; on the other hand, the (k+1)th predicted score set includes the (k+1)th predicted frost resistance score, the (k+1)th predicted permeability score, and the (k+1)th predicted cost score, thus improving the completeness and comprehensiveness of the scores in the (k+1)th predicted score set, and enabling the (k+1)th predicted score set to fully reflect the performance level corresponding to the component proportions in the (k+1)th optimized mix proportion set.

[0062] Based on the foregoing embodiments, the performance optimization method for desert sand concrete U-shaped channels based on intelligent algorithms provided in this application can also perform the following operations: If the degree of difference at the (k+1)th digit is greater than or equal to the degree threshold, based on the predicted antifreeze score at the (k+1)th digit and the degree of difference at the (k+1)th digit, update the first weight in the objective function related to antifreeze performance. Based on the predicted permeability score at the (k+1)th digit and the degree of difference at the (k+1)th digit, update the second weight in the objective function related to permeability performance. Based on the predicted cost score at the (k+1)th digit and the degree of difference at the (k+1)th digit, update the third weight in the objective function related to cost performance, thus obtaining the objective function at the (k+1)th digit.

[0063] Accordingly, if the degree of difference at the (k+1)th point is less than the degree threshold, then the operation of updating any weight of the kth objective function can be omitted, and the target matching set can be determined by the method provided in the aforementioned embodiments.

[0064] In some embodiments, the first weight can be used to weight the predicted frost resistance score obtained after each round of iterative optimization; correspondingly, by adjusting the first weight, the uncertainty of the predicted frost resistance score obtained after each round of iterative optimization can be adjusted to reduce the uncertainty of the predicted frost resistance score.

[0065] In some embodiments, the second weight can be used to weight the predicted penetration score obtained after each round of iterative optimization; accordingly, by adjusting the second weight, the uncertainty of the predicted penetration score obtained after each round of iterative optimization can be adjusted to reduce the uncertainty of the above predicted penetration score.

[0066] In some embodiments, the third weight can be used to weight the predicted cost score obtained after each round of iterative optimization; accordingly, by adjusting the third weight, the uncertainty of the predicted cost score obtained after each round of iterative optimization can be adjusted to reduce the uncertainty of the above predicted cost score.

[0067] For example, the degree of difference at the (k+1)th digit can be determined in the following way: The root mean square error between the predicted score of type m in the (k+1)th predicted score set and the measured score of type m in the (k+1)th measured score set is determined as the m-th difference score after the (k+1)th iteration optimization. With m ranging from 1 to M, the first difference score to the M-th difference score after the (k+1)th iteration optimization are integrated to obtain the (k+1)th difference degree. Specifically, this can be shown in equation (1): (1) in, For the m-th difference score, For the predicted score of type m in the (k+1)th predicted score set, It represents the measured score of type m in the (k+1)th measured score set.

[0068] Specifically, updating any weight in the k-th objective function can be achieved in the following way: Based on the ratio of the reciprocal of the m-th difference score corresponding to the predicted score of the m-th type in the (k+1)-th difference degree to the (k+1)-th difference degree, the m-th correction coefficient is determined. Then, the m-th weight in the k-th objective function is corrected based on the m-th correction coefficient to obtain the corrected m-th weight. The m-th weight in the k-th objective function is then updated according to the corrected m-th weight. When the value of m traverses from 1 to M, the update of all weights of the k-th objective function can be realized. Specifically, it can be shown in Equation (2): (2) in, For the m-th correction factor, Let m be the weight in the k-th objective function. This is the corrected weight of the m-th weight.

[0069] It should be noted that after obtaining the (k+1)th objective function, a multi-objective genetic algorithm can be used to perform the (k+2)th round of iterative optimization on the (k+1)th optimized ratio set based on the (k+1)th objective function, thereby obtaining the (k+2)th optimized ratio set. The new (k+2)th difference degree is then determined by the method in the aforementioned embodiment until the difference degree corresponding to the new round of iterative optimization is less than or equal to the degree threshold.

[0070] As can be seen from the above, in the performance optimization method for U-shaped channels in desert sand concrete based on intelligent algorithms provided in this application embodiment, if the (k+1)th degree of difference is greater than or equal to the degree threshold, based on the (k+1)th predicted frost resistance score and the (k+1)th degree of difference, the first weight associated with frost resistance in the k-th objective function is updated; based on the (k+1)th predicted permeability score and the (k+1)th degree of difference, the second weight associated with permeability in the k-th objective function is updated; and based on the (k+1)th predicted cost score and the (k+1)th degree of difference, the third weight associated with cost performance in the k-th objective function is updated, thus obtaining the (k+1)th objective function. Through the above operations, the updating of each weight in the k-th objective function is controlled, enabling targeted updates of each weight in the k-th objective function, thereby increasing the probability of matching the (k+1)th objective function with the iterative optimization process of the multi-objective genetic algorithm.

[0071] Based on the foregoing embodiments, the method for optimizing the performance of U-shaped channels in desert sand concrete based on intelligent algorithms provided in this application can be achieved by determining the (k+1)th predicted data set corresponding to the (k+1)th optimized mix proportion set through the following steps: Step B1: Analyze the (k+1)th optimized ratio set using a Long Short-Term Memory (LSTM) network to obtain the (k+1)th predicted mass loss rate and the (k+1)th predicted dynamic modulus loss rate.

[0072] In some embodiments, the (k+1)th predicted mass loss rate and the (k+1)th predicted motion model loss rate can be implemented in the following ways: The number of freeze-thaw cycles and environmental data sequences corresponding to climatic conditions are determined. Then, LSTM is used to predict the data related to macroscopic performance degradation in the (k+1)th optimized ratio set based on the environmental data sequences, so as to obtain the (k+1)th predicted mass loss rate and the (k+1)th predicted dynamic elastic modulus loss rate. The environmental data sequences may include temperature sequences, salinity sequences and humidity sequences, and the macroscopic performance may include mass loss and the performance of the virtual channel against elastic deformation.

[0073] In some embodiments, the LSTM can employ three bidirectional LSTM layers (Bi-LSTM), each with 128 LSTM cells, and the Dropout can be set to 0.3.

[0074] It should be noted that before the LSTM analyzes any optimized mix set, the initial LSTM can be trained based on the sample data. The optimizer used during training can be Adam, the learning rate can be 0.001, and the loss function can be the mean square error function. The sample data can include the changes in the mass loss rate and dynamic elastic modulus loss rate of desert sand concrete U-shaped channels with various weather conditions.

[0075] Step B2: Analyze the (k+1)th optimized mix ratio set using a convolutional neural network (CNN) to obtain the (k+1)th predicted seepage height and the (k+1)th predicted porosity.

[0076] In some embodiments, the (k+1)th predicted seepage height and the (k+1)th predicted porosity can be obtained as follows: By using CNN based on environmental data sequences, the proportions related to microscopic performance degradation in the (k+1)th optimized proportion set are predicted, resulting in the (k+1)th predicted seepage height and the (k+1)th predicted porosity. The microscopic performance can be characterized by data such as pore area, pore perimeter, and fractal dimension.

[0077] It should be noted that the LSTM and CNN in this embodiment can be configured in parallel.

[0078] In some embodiments, the (k+1)th predicted data set can be obtained by splicing together the (k+1)th predicted mass loss rate, the (k+1)th predicted dynamic modulus loss rate, the (k+1)th predicted seepage height, and the (k+1)th predicted porosity.

[0079] As can be seen from the above, in the performance optimization method for desert sand concrete U-shaped channels based on intelligent algorithms provided in this application embodiment, LSTM is used to analyze the (k+1)th optimized mix proportion set to obtain the (k+1)th predicted mass loss rate and the (k+1)th predicted dynamic modulus loss rate. CNN is used to analyze the (k+1)th optimized mix proportion set to obtain the (k+1)th predicted seepage height and the (k+1)th predicted porosity. Thus, by using LSTM and CNN to analyze the (k+1)th optimized mix proportion set respectively, the efficiency of obtaining the (k+1)th predicted data set can be improved; furthermore, by leveraging the advantages of LSTM and CNN in data analysis dimensions, the accuracy of the (k+1)th predicted data set can be improved.

[0080] Based on the foregoing embodiments, the method for optimizing the performance of U-shaped channels for desert sand concrete based on intelligent algorithms provided in this application involves processing the (k+1)th predicted mass loss rate, the (k+1)th predicted dynamic modulus loss rate, and the (k+1)th predicted ultrasonic loss rate to obtain the (k+1)th predicted frost resistance score. This can be achieved in the following way: The (k+1)th predicted mass loss rate, (k+1)th predicted dynamic modulus loss rate, and (k+1)th predicted ultrasonic loss rate are processed by the first function to obtain the (k+1)th predicted antifreeze score.

[0081] In some embodiments, the first function may include a first linear function, which may linearly weight the (k+1)th predicted mass loss rate, the (k+1)th predicted dynamic modulus loss rate, and the (k+1)th predicted ultrasonic loss rate based on a first weight set to obtain the (k+1)th predicted antifreeze score; specifically, the first function may be as shown in equation (3): (3) in, The antifreeze score can be predicted for the (k+1)th digit. We can predict the quality loss rate for the (k+1)th digit. The dynamic modulus loss rate can be predicted for the (k+1)th digit. The ultrasonic loss rate can be predicted for the (k+1)th wave. , as well as This can form a first weight set, where the values ​​of the first weight set can be (0.4, 0.35, 0.25).

[0082] Accordingly, the (k+1)th predicted seepage height and the (k+1)th predicted porosity are processed to obtain the (k+1)th predicted permeability score, which can be achieved in the following way: The (k+1)th predicted seepage height and (k+1)th predicted porosity are processed by the second function to obtain the (k+1)th predicted permeability score.

[0083] In some embodiments, the second function may include a second linear function, which may linearly weight the (k+1)th predicted seepage height and the (k+1)th predicted porosity based on a second weight set to obtain the (k+1)th predicted permeability score; specifically, the second function may be as shown in equation (4): (4) in, The seepage height can be predicted for the (k+1)th water level. Porosity can be predicted for the (k+1)th term. The penetration score can be predicted for the (k+1)th node. and This can form a second weight set, and the values ​​of the second weight set can be (0.6, 0.4).

[0084] Accordingly, the (k+1)th predicted cement content and the (k+1)th predicted admixture content are processed to obtain the (k+1)th predicted cost score, which can be achieved in the following way: The predicted cement content and the predicted admixture content of the (k+1)th time are processed by the third function to obtain the predicted cost score of the (k+1)th time.

[0085] In some embodiments, the third function may include a third linear function, which may linearly weight the (k+1)th predicted cement content and the (k+1)th predicted admixture content based on a third weight set to obtain the (k+1)th predicted cost score; specifically, the third function may be as shown in equation (5): (5) in, The cement content can be predicted for the (k+1)th digit. The admixture content can be predicted for the (k+1)th admixture. The cost score can be predicted for the (k+1)th term. and This can form a third weight set, where the weights can characterize the unit cost of cement and admixtures.

[0086] Accordingly, the technical solutions provided in the embodiments of this application can also perform the following operations: The first, second, and third functions are processed based on the initial weight set to obtain the initial objective function of the multi-objective genetic algorithm.

[0087] The first objective function is obtained by adjusting the weights of the initial objective function.

[0088] In some embodiments, the initial objective function can be obtained in the following way: The first, second, and third functions are linearly weighted based on the initial weight set to obtain the initial objective function; for example, the initial objective function can be shown in equation (6): (6) in, , ,as well as This can form the initial weight set.

[0089] In some embodiments, the first objective function can be obtained by updating at least a portion of the weights in the initial weight set of the initial objective function using the method provided in the foregoing embodiments.

[0090] As can be seen from the above, in the performance optimization method for desert sand concrete U-shaped channels based on intelligent algorithms provided in this application embodiment, the data in the k-th prediction data set is processed by the first function, the second function, and the third function respectively, to obtain the k+1-th predicted frost resistance score, the k+1-th predicted permeability score, and the k+1-th predicted cost score. In this way, targeted processing of the data in the k+1-th prediction data set is achieved. Furthermore, the first function, the second function, and the third function are processed based on the initial weight set to obtain the initial objective function of the multi-objective genetic algorithm. Thus, through the above process, the initial objective function of the multi-objective genetic algorithm and the first to third functions of the prediction data set corresponding to the optimized ratio set obtained by the iterative optimization of the multi-objective genetic algorithm can be associated, thereby improving the consistency between the initial objective function and the optimized ratio set processed by the multi-objective genetic algorithm. On this basis, the first objective function is obtained by adjusting the weights of the initial objective function, realizing the continuous updating of the objective function of the multi-objective genetic algorithm.

[0091] Based on the foregoing embodiments, the performance optimization method for desert sand concrete U-shaped channels based on intelligent algorithms provided in this application can also perform the following operations: Select the (k+1)th candidate ratio set from the (k+1)th optimized ratio set, and construct the (k+1)th channel set based on the (k+1)th candidate ratio set; obtain the (k+1)th measured score set by performing performance tests on the (k+1)th channel set.

[0092] In some embodiments, the (k+1)th candidate ratio set may include a subset of the group ratios that satisfy the threshold set in the (k+1)th optimized ratio set; for example, the (k+1)th candidate ratio set can be selected in the following manner: Based on the relationship between the proportions of each component in the (k+1)th candidate ratio set and the thresholds in the threshold set, the proportions of the components in the (k+1)th candidate ratio set are screened to obtain the (k+1)th candidate ratio set.

[0093] In some embodiments, the (k+1)th channel set can be constructed as follows: Based on the component ratios in the (k+1)th candidate mix design set, desert sand concrete U-shaped channels are constructed respectively to obtain the (k+1)th channel set; for example, the channels in the (k+1)th channel set can be the physical channels in the aforementioned embodiments.

[0094] In some embodiments, the set of measured scores for the (k+1)th time can be obtained in the following way: An experimental environment is constructed based on the environmental conditions, and each channel in the (k+1)th channel set is set in the experimental environment. During the continuous experimental conditions corresponding to the experimental environment, the macroscopic and microscopic performance parameters of each channel in the (k+1)th channel set are statistically analyzed to obtain the (k+1)th measured score set.

[0095] As can be seen from the above, in the performance optimization method for U-shaped channels in desert sand concrete based on intelligent algorithms provided in this application embodiment, the (k+1)th candidate mix design set is selected from the (k+1)th optimized mix design set, and the (k+1)th channel set is constructed based on the (k+1)th candidate mix design set. Then, performance testing is performed on the (k+1)th channel set to obtain the (k+1)th measured score set. This improves the correlation between the channels in the (k+1)th channel set and the (k+1)th candidate mix design set, and reduces the number of channels in the (k+1)th channel set, thereby reducing the amount of data for the (k+1)th measured scores. Simultaneously, through the above operations, the effectiveness of the (k+1)th measured score set is improved.

[0096] Based on the foregoing embodiments, the performance optimization method for desert sand concrete U-shaped channels based on intelligent algorithms provided in this application can also perform the following operations: The k+1 candidate matching set is processed by a dual-channel hybrid model to obtain the k+1 predicted score set.

[0097] The dual-channel hybrid model includes LSTM and CNN configured in parallel.

[0098] In some embodiments, the number of freeze-thaw cycles and the environmental data sequence corresponding to the weather conditions can be determined first. Then, LSTM can be used to predict the data related to macroscopic performance degradation in the (k+1)th candidate mix set based on the environmental data sequence to obtain the (k+1)th predicted mass loss rate and the (k+1)th predicted dynamic modulus loss rate.

[0099] In some embodiments, a CNN can be used to predict the proportions related to microscopic performance degradation in the (k+1)th candidate proportion set based on environmental data sequences, thereby obtaining the (k+1)th predicted seepage height and the (k+1)th predicted porosity.

[0100] As can be seen from the above, in the performance optimization method for desert sand concrete U-shaped channels based on intelligent algorithms provided in this application embodiment, the (k+1)th candidate mix design set is processed by a dual-channel hybrid model to obtain the (k+1)th predicted score set. The dual-channel hybrid model includes LSTM and CNN set in parallel. In this way, the efficiency of processing the (k+1)th candidate mix design set can be improved; and, by leveraging the advantages of LSTM and CNN in data processing dimensions, the accuracy of the (k+1)th predicted score set can be improved.

[0101] Based on the foregoing embodiments, in the performance optimization method for desert sand concrete U-shaped channels based on intelligent algorithms provided in this application, if the (k+1)th difference degree is less than the degree threshold, the (k+1)th optimized mix proportion set is determined as the target mix proportion set corresponding to the component list, which can be achieved through the following steps: Step C1: Determine the (k+1)th performance optimization score corresponding to the (k+1)th optimized ratio set.

[0102] In some embodiments, the (k+1)th performance optimization score can intuitively quantify the degree of improvement in the performance of the (k+1)th optimized mix proportion set in the environment characterized by the environmental state, and the performance of the desert sand concrete U-shaped channel.

[0103] In some embodiments, the k+1th performance optimization score can be determined in the following way: The predicted score set corresponding to the (k+1)th candidate ratio set in the (k+1)th optimized ratio set is processed by the (k+1)th objective function to obtain the (k+1)th performance optimization score.

[0104] Step C2: If the degree of difference of the (k+1)th component is less than the degree threshold, the difference between the (k+1)th performance optimization score and the kth performance optimization score is less than or equal to the target threshold, and the difference between the kth performance optimization score and the (k-1)th performance optimization score is less than or equal to the target threshold, then the (k+1)th optimized ratio set is determined as the target ratio set corresponding to the component list.

[0105] Accordingly, if the degree of difference of the (k+1)th optimization set is greater than the degree threshold, or the difference between the (k+1)th performance optimization score and the kth performance optimization score is greater than the target threshold, or the difference between the kth performance optimization score and the (k-1)th performance optimization score is greater than the target threshold, then the operation of determining the (k+1)th optimized ratio set as the target ratio set can be omitted.

[0106] In some embodiments, the k-th performance optimization score can characterize the degree of performance improvement of the k-th candidate ratio set corresponding to the k-th optimized ratio set; for example, the k-th performance optimization score can be calculated in the following manner: The k-th performance optimization score is obtained by processing the predicted score set corresponding to the k-th candidate ratio set through the k-th objective function.

[0107] Accordingly, the (k-1)th performance optimization score can characterize the degree of performance improvement of the (k-1)th candidate ratio set corresponding to the (k-1)th optimized ratio set; for example, the (k-1)th performance optimization score can be calculated in the following way: The (k-1)th performance optimization score is obtained by calculating the predicted score set corresponding to the (k-1)th candidate ratio set using the (k-1)th objective function.

[0108] In some embodiments, the (k+1)th candidate ratio set corresponding to the (k+1)th optimized ratio set can be determined as the target ratio set.

[0109] As can be seen from the above, in the performance optimization method for desert sand concrete U-shaped channels based on intelligent algorithms provided in this application embodiment, the performance optimization score corresponding to the (k+1)th optimized configuration set is determined, thereby realizing the performance quantification of the (k+1)th optimized configuration set. Furthermore, if the degree of difference of the (k+1)th set is less than the degree threshold, the performance optimization score of the (k+1)th set, the score difference between the (k+1)th and the performance optimization score of the (k)th set is less than or equal to the target threshold, and the score difference between the performance optimization score of the (k)th set and the performance optimization score of the (k-1)th set is less than or equal to the target threshold, the (k+1)th optimized mix ratio set is determined as the target mix ratio set. In this way, strict condition control is achieved for the operation of determining the target mix ratio set, and the convergence of the performance optimization score of the target mix ratio set is consistent with that of the optimized mix ratio set output by the multi-objective genetic algorithm is also achieved.

[0110] Figure 2 This is a schematic diagram of the architecture of the desert sand concrete U-shaped channel performance optimization method provided in the embodiments of this application, as shown below. Figure 2 As shown, the architecture may include: The preprocessing module 201 is used to collect climate data and mix proportion data of desert sand concrete U-shaped channels in Ningxia. For example, the climate data may include data such as temperature, humidity, and duration of different temperatures and humidity, and the mix proportion data may include the desert sand replacement rate, water-cement ratio, sand ratio, cement dosage, admixture dosage, and other component proportions.

[0111] The collaborative optimization module 202 is used to construct the climate state based on the climate data collected by the preprocessing module, and at the same time construct the constraints and the initial objective function. It generates the initial ratio set based on the ratio data collected by the preprocessing module, and performs iterative optimization through a multi-objective genetic algorithm to obtain the (k+1)th optimized ratio set, and finally obtains the target ratio set.

[0112] Specifically, the collaborative optimization module 202 may include an iterative optimization module 203, a performance prediction module 204, and a difference calculation module 205; wherein: The iterative optimization module 203 is used to perform iterative optimization based on the initial ratio set and the initial objective function under constraints using a multi-objective genetic algorithm to obtain the (k+1)th optimized ratio set.

[0113] The performance prediction module 204 is used to predict the (k+1)th optimized mix set using LSTM to obtain the (k+1)th predicted mass loss rate and the (k+1)th predicted dynamic modulus loss rate. At the same time, it uses CNN to predict the (k+1)th optimized mix set to obtain the (k+1)th predicted seepage height and the (k+1)th predicted porosity. The module integrates the (k+1)th predicted mass loss rate, the (k+1)th predicted dynamic modulus loss rate, the (k+1)th predicted seepage height, and the (k+1)th predicted porosity to obtain the (k+1)th predicted data set.

[0114] The difference calculation module 205 is used to determine the degree of difference between the (k+1)th predicted score set and the (k+1)th measured score set. When the degree of difference is greater than or equal to the degree threshold, the iterative optimization module is triggered to perform the next iterative optimization. When the degree of difference is less than the degree threshold, the (k+1)th candidate ratio set corresponding to the (k+1)th optimized ratio set is determined as the target ratio set.

[0115] Output module 206 is used to output the target ratio set.

[0116] As can be seen from the above, the technical solution provided in this application embodiment can achieve iterative optimization of the component ratio corresponding to the desert sand concrete U-shaped channel by adopting a multi-objective genetic algorithm for iterative optimization. In the above iterative optimization process, by tracking and comparing the degree of difference between the predicted data and the measured data corresponding to the component ratio obtained by iterative optimization, the convergence of the multi-objective genetic algorithm can be accelerated, thereby improving the efficiency of determining the target ratio set. At the same time, with the help of the constraints and iterative objective function of the multi-objective genetic algorithm, the accuracy and pertinence of the target ratio set can be improved, so that the desert sand concrete U-shaped channel constructed based on the target ratio set can resist the corrosion of the climate in Ningxia and improve the stability and robustness of the U-shaped channel performance.

[0117] The description of the various embodiments above tends to emphasize the differences between the various embodiments. The similarities or similarities between them can be referred to, and for the sake of brevity, they will not be repeated here.

[0118] The methods disclosed in the various method embodiments provided in this application can be arbitrarily combined to obtain new method embodiments without conflict.

[0119] The features disclosed in the various product embodiments provided in this application can be arbitrarily combined without conflict to obtain new product embodiments.

[0120] The features disclosed in the various method or device embodiments provided in this application can be arbitrarily combined without conflict to obtain new method or device embodiments.

[0121] It should be noted that the aforementioned computer-readable storage media can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), magnetic random access memory (FRAM), flash memory, magnetic surface memory, optical disc, or compact disc read-only memory (CD-ROM), etc.; or it can be various electronic devices including one or any combination of the above-mentioned memories, such as mobile phones, computers, tablet devices, personal digital assistants, etc.

[0122] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0123] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0124] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware nodes. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0125] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0126] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0127] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0128] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method for performance optimization of U-shaped channels made of desert sand concrete based on intelligent algorithms, characterized in that, The method includes: An initial mix design set is generated based on a component list using a multi-objective genetic algorithm; wherein, the component list includes a list of components used to construct the desert sand concrete U-shaped channel; and the initial mix design set includes an initialized set of component ratios among the components in the component list. The multi-objective genetic algorithm, based on the k-th objective function and constraints, performs the (k+1)-th round of iterative optimization on the k-th optimal ratio set to obtain the (k+1)-th optimal ratio set. The k-th objective function is obtained by adjusting the weights of the (k-1)-th objective function based on the k-th degree of difference. The k-th degree of difference includes the difference between the k-th predicted score set and the k-th measured score set corresponding to the k-th optimal ratio set. The k-th degree of difference is greater than or equal to a threshold. The k-th optimal ratio set is associated with the initial ratio set. The constraints are at least associated with the climate conditions of Ningxia region. k is an integer greater than 1. If the degree of difference at the (k+1)th point is less than the degree threshold, the (k+1)th optimized ratio set is determined as the target ratio set corresponding to the component list; wherein, the degree of difference at the (k+1)th point includes the difference between the (k+1)th predicted score set and the (k+1)th measured score set corresponding to the (k+1)th optimized ratio set.

2. The method according to claim 1, characterized in that, The (k+1)th predicted score set includes the (k+1)th predicted frost resistance score, the (k+1)th predicted permeability score, and the (k+1)th predicted cost score; the method further includes: Determine the (k+1)th predicted data set corresponding to the (k+1)th optimized mix proportion set; wherein, the (k+1)th predicted data set includes the (k+1)th predicted mass loss rate, the (k+1)th predicted dynamic modulus loss rate, the (k+1)th predicted ultrasonic loss rate, the (k+1)th predicted water seepage height, the (k+1)th predicted porosity, the (k+1)th predicted cement content, and the (k+1)th predicted admixture content; The (k+1)th predicted mass loss rate, the (k+1)th predicted dynamic modulus loss rate, and the (k+1)th predicted ultrasonic loss rate are processed to obtain the (k+1)th predicted antifreeze score. The (k+1)th predicted seepage height and the (k+1)th predicted porosity are processed to obtain the (k+1)th predicted permeability score. The predicted cement content and the predicted admixture content of the (k+1)th time are processed to obtain the predicted cost score of the (k+1)th time.

3. The method according to claim 2, characterized in that, The method further includes: If the degree of difference at the (k+1)th rank is greater than or equal to the degree threshold, based on the (k+1)th rank predicted antifreeze score and the (k+1)th rank difference, the first weight associated with antifreeze performance in the k-th objective function is updated; based on the (k+1)th rank predicted permeability score and the (k+1)th rank difference, the second weight associated with permeability performance in the k-th objective function is updated; based on the (k+1)th rank predicted cost score and the (k+1)th rank difference, the third weight associated with cost performance in the k-th objective function is updated, thus obtaining the (k+1)th objective function.

4. The method according to claim 2, characterized in that, The determination of the (k+1)th predicted data set corresponding to the (k+1)th optimized allocation set includes: The k+1th optimized ratio set is analyzed by a Long Short-Term Memory (LSTM) network to obtain the k+1th predicted mass loss rate and the k+1th predicted dynamic modulus loss rate. The (k+1)th optimized mix ratio set is analyzed by a convolutional neural network (CNN) to obtain the (k+1)th predicted seepage height and the (k+1)th predicted porosity.

5. The method according to claim 2, characterized in that, The process of processing the (k+1)th predicted mass loss rate, the (k+1)th predicted elastic modulus loss rate, and the (k+1)th predicted ultrasonic loss rate to obtain the (k+1)th predicted freeze resistance score includes: The (k+1)th predicted mass loss rate, the (k+1)th predicted dynamic modulus loss rate, and the (k+1)th predicted ultrasonic loss rate are processed by the first function to obtain the (k+1)th predicted antifreeze score. The (k+1)th predicted seepage height and the (k+1)th predicted porosity are processed to obtain the (k+1)th predicted permeability score. The (k+1)th predicted seepage height and the (k+1)th predicted porosity are processed by the second function to obtain the (k+1)th predicted permeability score. The process of processing the (k+1)th predicted cement content and the (k+1)th predicted admixture content to obtain the (k+1)th predicted cost score includes: The predicted cement content and the predicted admixture content of the (k+1)th time are processed by the third function to obtain the predicted cost score of the (k+1)th time. The method further includes: The first function, the second function, and the third function are processed based on the initial weight set to obtain the initial objective function of the multi-objective genetic algorithm; wherein the first objective function is obtained by adjusting the weights of the initial objective function.

6. The method according to claim 1, characterized in that, The method further includes: Select the (k+1)th candidate ratio set from the (k+1)th optimized ratio set; Construct the (k+1)th channel set based on the (k+1)th candidate allocation set; The k+1th channel set is obtained by performing performance tests on the k+1th channel set.

7. The method according to claim 6, characterized in that, The method further includes: The (k+1)th candidate matching set is processed by a dual-channel hybrid model to obtain the (k+1)th predicted score set; wherein, the dual-channel hybrid model includes LSTM and CNN set in parallel.

8. The method according to claim 1, characterized in that, If the degree of difference at the (k+1)th stage is less than the degree threshold, the (k+1)th optimized ratio set is determined as the target ratio set corresponding to the component list, including: Determine the (k+1)th performance optimization score corresponding to the (k+1)th optimized ratio set; If the degree of difference of the (k+1)th time is less than the degree threshold, the difference between the (k+1)th performance optimization score and the kth performance optimization score is less than or equal to the target threshold, and the difference between the kth performance optimization score and the (k-1)th performance optimization score is less than or equal to the target threshold, then the (k+1)th optimized ratio set is determined as the target ratio set corresponding to the component list.

Images