Machine learning-based shield synchronous grouting material proportioning management method and system

Abstract

This invention discloses a machine learning-based method and system for managing the proportion of materials used in shield tunneling synchronous grouting, relating to the field of material preparation and proportion management. This invention enhances the generalization ability and prediction accuracy of the model by combining simplified model data expansion with standard model construction. An iterative adjustment mechanism is introduced to automatically correct substandard proportion schemes until a set of candidates meeting all performance requirements is obtained. During the proportion selection stage, raw material storage limitations and economic factors are comprehensively considered, prioritizing schemes with sufficient inventory. When inventory is insufficient, a cost optimization strategy is adopted, effectively balancing technical feasibility and economic rationality. Finally, multi-objective optimization is achieved through performance index weighting, ensuring that the selected proportion scheme meets engineering performance requirements while possessing the best overall benefits. The entire process is highly automated, reducing human trial-and-error costs and improving the efficiency of proportion design and the level of engineering quality control.

Description

Technical Field

[0001] This invention belongs to the field of material preparation and proportion management, and specifically relates to a method and system for managing the proportion of materials for synchronous grouting in tunnel boring machines based on machine learning. Background Technology

[0002] In existing technologies, the traditional shield tunneling synchronous grouting material ratio does not establish a precise mapping relationship between ratio factors and performance indicators, resulting in insufficient model generalization ability and limited prediction accuracy. The ratio adjustment process lacks an automated mechanism, making it unable to quickly respond to deviations in performance indicators and requiring repeated manual testing for verification. Furthermore, it does not comprehensively consider raw material storage constraints and economic factors, which can easily lead to inventory backlog or excessively high procurement costs. It is impossible to achieve synergistic optimization of technical feasibility and economic benefits, making it difficult to maximize comprehensive benefits while meeting engineering performance requirements. Summary of the Invention

[0003] To address the problems in related technologies, this invention proposes a method and system for managing the proportion of grouting materials for tunnel boring machines based on machine learning, in order to overcome the aforementioned technical problems existing in the current related technologies.

[0004] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution: This invention relates to a machine learning-based method for managing the proportion of materials used in shield tunneling synchronous grouting, comprising the following steps: S1. Set several key proportioning factors and corresponding grout performance index types, then conduct grouting material proportioning experiments and obtain the corresponding key proportioning factors and performance index data. S2. Based on the data obtained in S1, construct a simplified mapping model for the performance index data of the grouting slurry; then use the simplified mapping model to expand the data obtained in S1. S3. Construct a standard mapping model for grout performance index data based on the data expanded from S2; S4. Input the current initial key ratio factor data into the standard mapping model of S3 for mapping; when the mapping result does not meet the interval requirements, repeatedly adjust the initial key ratio factor data and remap until several sets of alternative key ratio factor data that meet the interval requirements of all performance indicators are obtained. S5. Calculate the raw material requirement based on the data obtained in S4. If the actual storage quantity meets the raw material requirement, filter out the corresponding ratio factor data as secondary alternative ratio factors. Otherwise, proceed to S6. Select the maximum index data from the performance indicators corresponding to the secondary alternative ratio factors according to their weights. S6. Calculate the total cost of raw material purchase for each alternative ratio scheme and select the scheme with the lowest cost for raw material procurement; and use the corresponding alternative key ratio factor data as the final key ratio factor dataset.

[0005] Preferably, step S1 includes the following steps: S11. Set several key proportioning factors for the shield tunnel synchronous grouting material to obtain a set of key proportioning factors for the grouting material; the set of key proportioning factors for the grouting material includes water-cement ratio, powder-ash ratio, asphalt content and rubber content, etc.; then set several performance index types for the grouting slurry to obtain a set of grouting slurry performance index types. S12. Based on the set of key proportioning factors of grouting materials and the set of performance index types of grouting slurry, grouting material proportioning experiments are conducted using different key proportioning factors. After the proportioning is completed, the corresponding performance index data of various types of grouting slurry are used to obtain the experimental dataset of grouting material proportioning factors and the experimental dataset of grouting slurry performance index. By setting a set of key proportioning factors and a set of grout performance index types for shield tunneling synchronous grouting materials, a complete experimental design framework was constructed, which can comprehensively cover the core parameters of material composition. Through experimental verification of the synergistic effect of multiple factors, a standardized performance index system was established, providing a high-quality and structured data foundation for subsequent machine learning model training.

[0006] Preferably, step S2 includes the following steps: S21. Based on the experimental dataset of grouting material ratio factors and the experimental dataset of grouting slurry performance indicators, a simplified mapping model is constructed with grouting material ratio factor data as input and grouting slurry performance indicator data as output, to obtain the final simplified mapping model of grouting slurry performance indicators. S22. Based on the set of key proportioning factors of the grouting material and the actual proportioning requirements, set the value range of each key proportioning factor of the grouting material and use the Latin cube sampling method to generate multiple sets of key proportioning factor data combinations within the corresponding set key proportioning factor range to obtain an extended dataset of grouting material proportioning factors; input each set of grouting material proportioning factor data in the extended dataset of grouting material proportioning factors into the simplified mapping model of the final grouting slurry performance index for mapping to obtain an extended dataset of grouting slurry performance index; The generation of extended datasets enables machine learning models to learn a wider range of proportion-performance mapping relationships, enhancing the model's predictive stability under extreme proportion conditions and boundary situations, and providing a reliable theoretical basis and data guarantee for material selection and performance prediction in engineering applications.

[0007] Preferably, step S3 includes the following steps: S31. Based on the extended dataset of grouting material proportioning factors and the extended dataset of grouting slurry performance indicators, construct a standard mapping model with grouting material proportioning factor data as input and grouting slurry performance indicator data as output, and obtain the final standard mapping model of grouting slurry performance indicators. The model is trained on an extended dataset and makes full use of the diverse ratio combinations generated by Latin cube sampling, which significantly improves the model's generalization ability and prediction stability. By deep learning the complex nonlinear relationship between various ratio factors and performance indicators, the model can accurately capture the comprehensive influence mechanism of key parameters such as water-cement ratio, powder-ash ratio, asphalt content, and rubber content on performance indicators such as fluidity, consistency, bleeding rate, setting time, and compressive strength.

[0008] Preferably, step S4 includes the following steps: S41. Based on the set of key proportioning factors for the grouting material, set the predetermined data of various key proportioning factors when preparing the grouting material to obtain the current initial key proportioning factor dataset; input the current initial key proportioning factor dataset into the final grouting slurry performance index standard mapping model for mapping to obtain the current initial slurry performance index dataset; based on the actual grouting material preparation requirements, obtain the specified value range of various performance indicators corresponding to the currently prepared grouting material to obtain the current performance index range set; S42. In conjunction with the current performance index interval set, if any performance index data in the current initial grout performance index dataset does not fall within the corresponding performance index interval, adjust the current initial key ratio factor dataset according to the value interval of each key ratio factor corresponding to the current grouting material preparation. The adjusted dataset is then input back into the final grouting slurry performance index standard mapping model for mapping, resulting in the current adjusted grout performance index dataset. Otherwise, no adjustment is required. S43. Repeat S42 to obtain the current adjusted key ratio factor dataset corresponding to the data of each performance index of several groups being located in the corresponding interval, and obtain the current candidate key ratio factor dataset. Through multiple rounds of iterative optimization, the mixing ratio parameters can be intelligently identified and corrected to ensure that the performance of the final grouting material fully meets the requirements of engineering specifications. This significantly improves the efficiency and accuracy of the mixing ratio optimization and avoids the blindness and high cost of the traditional trial-and-error method.

[0009] Preferably, step S5 includes the following steps: S51. Based on the current demand for grouting materials and the current dataset of key ratio factors for alternatives, calculate the demand for each type of raw material in the current process of preparing grouting materials, and obtain the current set of raw material demand. S52. Obtain the weight values ​​of various performance indicators of the grouting slurry according to the actual preparation requirements, and obtain the current performance indicator weight set; obtain the actual storage quantity of each preparation raw material, and obtain the current raw material actual storage quantity dataset; compare the current raw material actual storage quantity dataset with each group of preparation demand data in the current preparation raw material demand quantity set. If the actual storage amount of each type of raw material in the current raw material actual storage amount dataset is greater than or equal to the corresponding raw material demand amount, the key ratio factor data corresponding to the group of preparation demand data is selected according to the current candidate key ratio factor dataset; after the selection is completed, the current secondary candidate key ratio factor dataset is obtained. Otherwise, execute S6; S53. Based on the current secondary candidate key ratio factor dataset, obtain the corresponding slurry performance index data to obtain the current secondary candidate performance index dataset. Based on the current performance indicator weight set, the dataset corresponding to the largest performance indicator data is selected from the current secondary candidate performance indicator dataset in descending order of weight to obtain the current final performance indicator dataset; the key ratio factor data corresponding to the current final performance indicator dataset is used as the current final selection key ratio factor dataset. By calculating the demand for various raw materials based on alternative formulation schemes and then dynamically comparing them with actual inventory, the feasibility of the preparation scheme is ensured, effectively avoiding the risk of production interruption due to raw material shortages. The optimal formulation combination can be selected from multiple alternative schemes, achieving efficient resource utilization and optimal control of product quality.

[0010] Preferably, in the process of calculating the required amount of various raw materials in the current preparation of grouting materials as described in S51, the actual loss and utilization rate of various raw materials in the preparation process should be taken into account. By fully considering the quality loss factors throughout the entire process from raw material procurement to final product preparation, including multiple influencing factors such as storage loss, transportation loss, weighing error, and physicochemical loss during stirring, it not only ensures the sufficiency and accuracy of raw material supply in the actual preparation process, effectively avoiding production interruptions due to insufficient estimation or resource waste caused by over-purchasing, but also provides a reliable data foundation for subsequent cost accounting and inventory management. Preferably, step S6 includes the following steps: S61. Based on the current raw material demand set and the current raw material actual storage set, calculate the total cost required to purchase raw materials corresponding to each set of raw material demand, and obtain the current total purchase cost set. S62. Select the minimum cost group in the current total purchase cost set and purchase the required amount of raw materials. After the purchase is completed, based on the current candidate key ratio factor dataset, use the key ratio factor data corresponding to the preparation requirement data of this group as the current final key ratio factor dataset. By comprehensively considering raw material procurement costs while meeting performance requirements, and by accurately calculating the total purchase cost of each formulation, it provides decision-makers with an intuitive basis for cost comparison; it can automatically identify and select the most cost-effective formulation, effectively reducing material procurement costs and production costs.

[0011] The shield tunnel synchronous grouting material ratio management system based on machine learning includes an experimental data acquisition module, a simplified mapping model construction module, a standard mapping model construction module, a ratio factor adjustment module, a ratio factor screening module, and a raw material procurement determination module. The experimental data acquisition module is used to acquire key proportioning factors and performance index data in the grouting material proportioning experiment. The simplified mapping model construction module is used to construct a simplified mapping model for grouting slurry performance index data. The standard mapping model construction module is used to construct a standard mapping model for grouting slurry performance index data. The proportioning factor adjustment module is used to repeatedly adjust the initial key proportioning factor data; The ratio factor filtering module is used to filter the acquired key ratio factor data based on the storage capacity; The raw material procurement decision module is used to calculate the total cost of raw material procurement for each alternative ratio scheme and select the scheme with the lowest cost for raw material procurement.

[0012] The present invention has the following beneficial effects: 1. This invention establishes a complete mapping relationship between key proportioning factors and performance indicators, ensuring that proportioning design is based on sound principles. It enhances the model's generalization ability and prediction accuracy by combining simplified model data expansion with standard model construction. An iterative adjustment mechanism is introduced to automatically correct substandard proportioning schemes until a set of candidates meeting all performance requirements is obtained. During the proportioning screening stage, raw material storage limitations and economic factors are comprehensively considered, prioritizing schemes with sufficient inventory. When inventory is insufficient, a cost optimization strategy is adopted, effectively balancing technical feasibility and economic rationality. Finally, multi-objective optimization is achieved through performance indicator weight ranking, ensuring that the selected proportioning scheme meets engineering performance requirements while possessing the best overall benefits. The entire process is highly automated, reducing human trial-and-error costs and improving proportioning design efficiency and engineering quality control.

[0013] 2. In this invention, the performance of the initial mix design is predicted by a standard mapping model, and then the performance index range required by the actual project is accurately verified. When the prediction result deviates from the target range, the dynamic adjustment program of the mix factor is automatically started. Through multiple rounds of iterative optimization, the system can intelligently identify and correct the mix parameters to ensure that the performance of the final output grouting material fully meets the requirements of the engineering specifications.

[0014] 3. In this invention, the required quantities of various raw materials are calculated based on alternative formulation schemes, and then dynamically compared with the actual inventory to ensure the feasibility of the preparation scheme and effectively avoid the risk of production interruption due to raw material shortages. By establishing a performance index weighting system, the system can select the optimal formulation combination from multiple alternative schemes to achieve efficient resource utilization and optimal control of product quality.

[0015] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating the shield tunnel synchronous grouting material ratio management method based on machine learning according to the present invention. Figure 2 A schematic diagram of the process for constructing a simplified mapping model of the final grout performance index for this invention; Figure 3 This is a schematic diagram of the novel synchronous grouting material obtained by combining several sets of proportioning factors in this invention; Figure 4 A schematic diagram of the process for constructing the final grout performance index standard mapping model for this invention; Figure 5 This is a schematic diagram of the process for determining and adjusting key grouting ratio factors according to the present invention. Figure 6 This is a schematic diagram of the process for screening current key proportioning factors according to the present invention; Figure 7 This is a schematic diagram of the process for determining the purchase of current raw materials according to the present invention; Figure 8 This is a schematic diagram of the module of the shield tunnel synchronous grouting material proportioning management system based on machine learning of the present invention. Detailed Implementation

[0018] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0019] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0020] Example 1 Please see Figure 1 This embodiment describes a machine learning-based method for managing the proportion of materials used in shield tunneling synchronous grouting, comprising the following steps: S1. Set several key proportioning factors and corresponding grout performance index types, then conduct grouting material proportioning experiments and obtain the corresponding key proportioning factors and performance index data. S1 includes the following steps: S11. Several key proportioning factors of the shield tunnel synchronous grouting material are set to obtain a set of key proportioning factors of the grouting material; the set of key proportioning factors of the grouting material includes water-cement ratio, powder-ash ratio, asphalt content and rubber content, etc.; then, several performance index types of the grout are set to obtain a set of grout performance index types; the set of grout performance index types includes fluidity, consistency, bleeding rate, initial setting time, final setting time and 28-day compressive strength, etc. S12. Based on the set of key proportioning factors of grouting materials and the set of performance index types of grouting slurry, grouting material proportioning experiments are conducted using different key proportioning factors. After the proportioning is completed, the corresponding performance index data of various types of grouting slurry are used to obtain the experimental dataset of grouting material proportioning factors and the experimental dataset of grouting slurry performance index. For example, the contents of the experimental dataset of grouting material proportioning factors and the experimental dataset of grouting slurry performance indicators are shown in Table 1 below: Table 1. Schematic diagram of experimental data for grouting material proportioning.

[0021] By systematically defining the key proportioning factors and grout performance index types for shield tunneling synchronous grouting materials, a complete experimental design framework was constructed, providing a theoretical basis and practical guidance for the scientific proportioning of grouting materials. This scheme comprehensively covers the core parameters of material composition, ensuring the systematic nature and representativeness of the experimental design. Through experimental verification of the synergistic effects of multiple factors, the contribution and mutual influence of each proportioning factor on grout performance were effectively identified. Simultaneously, by establishing a standardized performance index system, quantitative evaluation and precise control of grouting material quality were achieved, providing a high-quality, structured data foundation for subsequent machine learning model training. This systematic experimental method not only improves R&D efficiency and reduces trial-and-error costs but also provides reliable data support for the optimized design and engineering application of grouting materials, significantly improving the accuracy and reliability of grouting material performance prediction, and providing important guarantees for shield tunneling construction quality control and engineering safety. S2. Based on the data obtained in S1, construct a simplified mapping model for the performance index data of the grouting slurry; then use the simplified mapping model to expand the data obtained in S1. Please see Figure 2 , Figure 3 S2 includes the following steps: S21. Based on the experimental dataset of grouting material ratio factors and the experimental dataset of grouting slurry performance indicators, a simplified mapping model is constructed with grouting material ratio factor data as input and grouting slurry performance indicator data as output, to obtain the final simplified mapping model of grouting slurry performance indicators. The structure of the simplified mapping model for the final grout performance indicators can be referenced in Table 2 below: Table 2. Simplified Mapping Model Structure for Grouting Slurry Performance Indicators

[0022] For continuous mix proportioning factors (such as water-cement ratio, powder-ash ratio, etc.) and performance index outputs (such as fluidity, compressive strength, etc.), the model achieves numerical prediction through a voting mechanism of multiple decision trees, effectively avoiding the problem of overfitting easily with a single decision tree. The model also sets up a feature importance evaluation mechanism, which can quantify the influence of each mix proportioning factor such as water-cement ratio, powder-ash ratio, asphalt content, and rubber content on grout performance indexes, providing intuitive guidance for optimizing grouting material formulations. Furthermore, the random forest regression model is suitable for small sample data and is applicable to small-scale data obtained through actual experiments. S22. Based on the set of key proportioning factors of the grouting material and the actual proportioning requirements, set the value range of each key proportioning factor of the grouting material and use the Latin Hypercube Sampling (LHS) method (an efficient multidimensional random sampling technique that ensures uniform distribution of samples in the parameter space through stratified design, particularly suitable for high-dimensional uncertainty analysis and optimization problems) to generate multiple sets of key proportioning factor data combinations within the corresponding set key proportioning factor ranges, thus obtaining an extended dataset of grouting material proportioning factors; input each set of grouting material proportioning factor data in the extended dataset of grouting material proportioning factors into the simplified mapping model of the final grouting slurry performance index for mapping, thus obtaining an extended dataset of grouting slurry performance index; By generating an extended dataset within a defined range of key proportioning factors using the Latin cube sampling method, the coverage and accuracy of grouting material performance prediction are significantly improved. This method ensures the uniform distribution of each proportioning factor across the entire value range, avoiding data clustering or blank areas that may occur with traditional random sampling, effectively improving the representativeness of the samples and the model's generalization ability. By inputting the generated multiple sets of proportioning factor data into the performance index mapping model, performance prediction of unverified proportioning schemes is achieved, significantly reducing the number and cost of actual proportioning experiments. This not only ensures the rationality of data distribution but also captures the interaction effects between proportioning factors, providing more comprehensive data support for the optimized design of grouting materials. Furthermore, the generation of the extended dataset enables the machine learning model to learn a wider range of proportioning-performance mapping relationships, enhancing the model's predictive stability under extreme proportioning conditions and boundary situations, and providing a reliable theoretical basis and data guarantee for material selection and performance prediction in engineering applications. S3. Construct a standard mapping model for grout performance index data based on the data expanded from S2; Please see Figure 4 S3 includes the following steps: S31. Based on the extended dataset of grouting material proportioning factors and the extended dataset of grouting slurry performance indicators, construct a standard mapping model with grouting material proportioning factor data as input and grouting slurry performance indicator data as output, and obtain the final standard mapping model of grouting slurry performance indicators. S31 includes the following steps: S311. Construct an initial standard mapping model for grouting slurry performance indicators and set a first training data ratio (e.g., 8:2 or 7:3, which can be adjusted adaptively according to the actual training situation); divide the extended dataset of grouting material proportioning factors and the extended dataset of grouting slurry performance indicators according to the first training data ratio to obtain the first training dataset and the first test dataset. S312. Set a first training error threshold (10%~15%, which can be adjusted adaptively according to the actual training situation); input the first training dataset into the initial grouting slurry performance index standard mapping model for training; during the training process, if the training error is less than the first training error threshold, stop training and obtain the trained grouting slurry performance index standard mapping model; otherwise, continue training until the training error is less than the first training error threshold. S313. Set a first test accuracy threshold (90%~95%, which can be adjusted adaptively according to the actual test situation); input the first test dataset into the trained grouting slurry performance index standard mapping model for testing; after the test is completed, obtain the first test accuracy data; if the first test accuracy data is greater than or equal to the first test accuracy threshold, use the trained grouting slurry performance index standard mapping model as the final grouting slurry performance index standard mapping model; otherwise, return to S312 to continue training the trained grouting slurry performance index standard mapping model and repeat S313 until the first test accuracy data is greater than or equal to the first test accuracy threshold. The structure of the initial grouting slurry performance index standard mapping model can be referenced in Table 3 below: Table 3. Schematic diagram of the standard mapping model for grouting slurry performance indicators.

[0023] By constructing a standard mapping model between grout material proportioning factors and grout performance indicators, high-precision prediction and intelligent control of grout material performance are achieved. This model, trained on an extended dataset, fully utilizes diverse proportioning combinations generated by Latin cube sampling, significantly improving its generalization ability and predictive stability. Through deep learning of the complex nonlinear relationships between various proportioning factors and performance indicators, the model accurately captures the comprehensive influence mechanism of key parameters such as water-cement ratio, powder-ash ratio, asphalt content, and rubber content on performance indicators such as fluidity, consistency, bleeding rate, setting time, and compressive strength. This standard mapping model not only possesses powerful predictive capabilities but also identifies key influencing factors through feature importance analysis, providing scientific guidance for grout material formulation optimization. It effectively solves the problems of blindness and inefficiency inherent in traditional empirical proportioning methods, providing a reliable digital decision support tool for shield tunneling synchronous grouting projects. S4. Input the current initial key ratio factor data into the standard mapping model of S3 for mapping; when the mapping result does not meet the interval requirements, repeatedly adjust the initial key ratio factor data and remap until several sets of alternative key ratio factor data that meet the interval requirements of all performance indicators are obtained. Please see Figure 5S4 includes the following steps: S41. Based on the set of key proportioning factors for the grouting material, set the predetermined data of various key proportioning factors when preparing the grouting material to obtain the current initial key proportioning factor dataset; input the current initial key proportioning factor dataset into the final grouting slurry performance index standard mapping model for mapping to obtain the current initial slurry performance index dataset; based on the actual grouting material preparation requirements, obtain the specified value range of various performance indicators corresponding to the currently prepared grouting material to obtain the current performance index range set; S42. In conjunction with the current performance index interval set, if any performance index data in the current initial grout performance index dataset does not fall within the corresponding performance index interval, adjust the current initial key ratio factor dataset according to the value interval of each key ratio factor corresponding to the current grouting material preparation. The adjusted dataset is then input back into the final grouting slurry performance index standard mapping model for mapping, resulting in the current adjusted grout performance index dataset. Otherwise, no adjustment is required. S43. Repeat S42 (the number of repetitions can be adaptively set according to the actual situation) to obtain the current adjusted key ratio factor dataset corresponding to the data of each performance index of several groups that are located in the corresponding interval, and obtain the current candidate key ratio factor dataset. By constructing a closed-loop optimization iteration mechanism, intelligent adjustment and optimization of grouting material proportions are achieved. This method first predicts the performance of the initial proportion scheme based on a standard mapping model, then performs precise verification by combining the performance index range required by the actual engineering project. When the prediction result deviates from the target range, a dynamic adjustment program for the proportion factors is automatically initiated. Through multiple rounds of iterative optimization, the system can intelligently identify and correct the proportion parameters, ensuring that the performance of the final output grouting material fully meets the engineering specifications. This closed-loop control strategy based on model prediction-verification-adjustment significantly improves the efficiency and accuracy of proportion optimization, avoiding the blindness and high cost problems of traditional trial-and-error methods. Simultaneously, this method possesses good adaptability, enabling rapid adjustment of the optimization strategy according to different engineering needs, providing reliable intelligent decision support for the standardized production and engineering application of grouting materials, and effectively ensuring the quality and safety of tunnel boring machine construction. S5. Calculate the raw material requirement based on the data obtained in S4. If the actual storage quantity meets the raw material requirement, filter out the corresponding ratio factor data as secondary alternative ratio factors. Otherwise, proceed to S6. Select the maximum index data from the performance indicators corresponding to the secondary alternative ratio factors according to their weights. Please see Figure 6 S5 includes the following steps: S51. Based on the current demand for grouting materials and the current dataset of key ratio factors for alternatives, calculate the demand for each type of raw material in the current process of preparing grouting materials, and obtain the current set of raw material demand. In calculating the required quantities of various raw materials during the current preparation of grouting materials, as described in S51, the actual loss and utilization rate of various raw materials during the preparation process should be considered (which can be evaluated and calculated based on historical preparation experience). By introducing actual loss and utilization parameters during the preparation process, a more accurate raw material demand calculation model was constructed, significantly improving the practicality and economy of grouting material proportioning decisions. This method fully considers quality loss factors throughout the entire process from raw material procurement to final product preparation, including multiple influencing factors such as storage loss, transportation loss, weighing error, and physicochemical losses during stirring. Through quantitative analysis of historical preparation experience data, a scientific loss rate and utilization rate evaluation system was established. This calculation method not only ensures the sufficiency and accuracy of raw material supply during actual preparation, effectively avoiding production interruptions due to insufficient estimation or resource waste due to over-purchasing, but also provides a reliable data foundation for subsequent cost accounting and inventory management. At the same time, this method enhances the feasibility of the proportioning scheme, enabling theoretical proportions to be better translated into actual production operations, significantly improving the efficiency and quality control level of grouting material preparation, and providing strong support for the smooth implementation of engineering projects. S52. Obtain the weight values ​​of various performance indicators of the grouting slurry according to the actual preparation requirements, and obtain the current performance indicator weight set; obtain the actual storage quantity of each preparation raw material, and obtain the current raw material actual storage quantity dataset; compare the current raw material actual storage quantity dataset with each group of preparation demand data in the current preparation raw material demand quantity set. If the actual storage amount of each type of raw material in the current raw material actual storage amount dataset is greater than or equal to the corresponding raw material demand amount, the key ratio factor data corresponding to the group of preparation demand data is selected according to the current candidate key ratio factor dataset; after the selection is completed, the current secondary candidate key ratio factor dataset is obtained. Otherwise, execute S6; S53. Based on the current secondary candidate key ratio factor dataset, obtain the corresponding slurry performance index data to obtain the current secondary candidate performance index dataset. Based on the current performance indicator weight set, the dataset corresponding to the largest performance indicator data is selected from the current secondary candidate performance indicator dataset in descending order of weight to obtain the current final performance indicator dataset; the key ratio factor data corresponding to the current final performance indicator dataset is used as the current final selection key ratio factor dataset. For example, as follows: 1. Initial proportion setting: The key proportioning factors for grouting materials were set as follows: water-cement ratio (0.95), powder-ash ratio (1.6), asphalt content (11%), and rubber content (8%). The initial key proportioning factor dataset was: water-cement ratio = 0.95, powder-ash ratio = 1.6, asphalt content = 11%, and rubber content = 8%. This proportioning was input into the standard mapping model for grout performance indicators to obtain the initial grout performance indicator dataset: fluidity = 215 mm, consistency = 175 mm, and bleeding rate = 8%. Water content = 2.3%, initial setting time = 5.5h, final setting time = 8.1h, 28-day compressive strength = 3.0MPa; Based on actual preparation requirements, the performance index range set is set as follows: fluidity [200, 250]mm, consistency [160, 200]mm, bleeding rate [1.0, 3.0]%, initial setting time [4.0, 7.0]h, final setting time [6.0, 9.0]h, 28-day compressive strength [2.8, 3.5]MPa; 2. First adjustment and optimization: Since the initial fluidity of 215 mm falls within the range of [200, 250] mm, the bleeding rate of 2.3% and the 28-day compressive strength of 3.0 MPa are also within this range, as are the initial setting time of 5.5 h and the final setting time of 8.1 h. However, considering the stability of the project, the water-cement ratio was adjusted to 0.90, the powder-ash ratio to 1.7, the asphalt content to 12%, and the rubber content to 7%. The adjusted mix proportions were input into the model, resulting in the following dataset of slurry performance indicators: fluidity = 228 mm, consistency = 188 mm, bleeding rate = 1.8%, initial setting time = 5.2 h, final setting time = 7.8 h, and 28-day compressive strength = 3.2 MPa. All performance indicators meet the range requirements. 3. Secondary adjustment and verification: Further optimization was performed, adjusting the water-cement ratio to 0.92, the powder-ash ratio to 1.65, the asphalt content to 11.5%, and the rubber content to 7.5%. The model was then re-entered, yielding the following results: fluidity = 222 mm, consistency = 182 mm, bleeding rate = 1.9%, initial setting time = 5.0 h, final setting time = 7.5 h, and 28-day compressive strength = 3.1 MPa. All indicators were within the specified range, forming a dataset of candidate key mix proportions. 4. Calculation of raw material requirements: Based on the alternative mixing ratios, the raw material requirements for preparing 100m³ of grouting material are calculated as follows: cement requirement = 45000kg, fly ash requirement = 28000kg, water requirement = 22000kg, asphalt requirement = 11000kg, and rubber requirement = 7500kg. Considering the actual loss rates (cement loss rate 5%, fly ash loss rate 3%, water loss rate 2%, asphalt loss rate 4%, and rubber loss rate 2%), the actual procurement requirements are as follows: cement = 47250kg, fly ash = 28840kg, water = 22440kg, asphalt = 11440kg, and rubber = 7650kg. 5. Inventory matching and cost optimization: The current raw material inventory dataset is as follows: cement = 50,000 kg, fly ash = 30,000 kg, water = 25,000 kg, asphalt = 12,000 kg, and rubber = 8,000 kg. Comparison shows that the inventory of all raw materials exceeds the actual demand, meeting the preparation conditions. Based on the performance index weight set (flowability weight 0.25, consistency weight 0.20, bleeding rate weight 0.15, initial setting time weight 0.15, final setting time weight 0.10, 28-day compressive strength weight 0.15), the proportioning scheme corresponding to the maximum value of each index is selected from the secondary candidate performance index dataset. The final selected key proportioning factor dataset is determined as follows: water-cement ratio = 0.92, fly ash ratio = 1.65, asphalt content = 11.5%, and rubber content = 7.5%. By constructing an intelligent decision-making mechanism that calculates raw material demand and matches inventory, precise management of the grouting material preparation process is achieved. Specifically, the demand for various raw materials is first calculated based on alternative proportions, and then dynamically compared with actual inventory to ensure the feasibility of the preparation plan and effectively avoid the risk of production interruption due to raw material shortages. By establishing a performance index weighting system, the system can select the optimal proportion combination from multiple alternatives, achieving efficient resource utilization and optimal control of product quality. This decision-making process not only considers the constraints of raw material supply but also fully integrates engineering performance requirements. The weighted ranking algorithm ensures that key performance indicators are prioritized, significantly improving the scientificity and practicality of grouting material proportioning decisions. At the same time, this method has good scalability and can adapt to changes in raw material demand for projects of different scales, providing a complete solution for standardized production and intelligent control of grouting materials, significantly reducing production costs and improving engineering efficiency. S6. Calculate the total cost of raw material purchase for each alternative ratio scheme and select the scheme with the lowest cost for raw material procurement; and use the corresponding alternative key ratio factor data as the final key ratio factor dataset; Please see Figure 7 S6 includes the following steps: S61. Based on the current raw material demand set and the current raw material actual storage set, calculate the total cost required to purchase raw materials corresponding to each set of raw material demand, and obtain the current total purchase cost set. S62. Select the minimum cost group in the current total purchase cost set and purchase the required amount of raw materials. After the purchase is completed, based on the current candidate key ratio factor dataset, use the key ratio factor data corresponding to the preparation requirement data of this group as the current final key ratio factor dataset. By constructing a cost optimization decision-making mechanism, the economic efficiency of the grouting material preparation process is maximized. This method, while meeting performance requirements, comprehensively considers raw material procurement costs and provides decision-makers with an intuitive cost comparison basis by accurately calculating the total purchase cost of each formulation. It can automatically identify and select the most cost-effective formulation, effectively reducing material procurement and production costs. This cost-benefit analysis-based decision-making process not only ensures that the performance indicators of the grouting material meet engineering requirements but also achieves optimal resource allocation through intelligent cost accounting. This method is particularly suitable for large-scale engineering applications, significantly improving the economic efficiency of grouting material preparation while ensuring product quality stability and reliability, providing a scientific decision support tool for cost control and resource management in engineering projects.

[0024] Example 2 Please see Figure 8 This embodiment discloses a shield tunnel synchronous grouting material proportion management system based on machine learning. The system can implement the methods of the above embodiments, including an experimental data acquisition module, a simplified mapping model construction module, a standard mapping model construction module, a proportioning factor adjustment module, a proportioning factor screening module, and a raw material procurement determination module. The experimental data acquisition module is used to acquire key proportioning factors and performance index data in the grouting material proportioning experiment. The simplified mapping model construction module is used to construct a simplified mapping model for grouting slurry performance index data. The standard mapping model construction module is used to construct a standard mapping model for grouting slurry performance index data. The proportioning factor adjustment module is used to repeatedly adjust the initial key proportioning factor data; The ratio factor filtering module is used to filter the acquired key ratio factor data based on the storage capacity; The raw material procurement decision module is used to calculate the total cost of raw material procurement for each alternative ratio scheme and select the scheme with the lowest cost for raw material procurement.

[0025] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0026] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to the present invention.

Claims

1. A machine learning-based method for managing the proportion of materials used in shield tunneling synchronous grouting, characterized in that, Includes the following steps: S1. Set several key proportioning factors and corresponding grout performance index types, then conduct grouting material proportioning experiments and obtain the corresponding key proportioning factors and performance index data. S2. Based on the data obtained in S1, construct a simplified mapping model for the performance index data of the grouting slurry; then use the simplified mapping model to expand the data obtained in S1. S3. Construct a standard mapping model for grout performance index data based on the data expanded from S2; S4. Input the current initial key ratio factor data into the standard mapping model of S3 for mapping; when the mapping result does not meet the interval requirements, repeatedly adjust the initial key ratio factor data and remap until several sets of alternative key ratio factor data that meet the interval requirements of all performance indicators are obtained. S5. Calculate the raw material requirement based on the data obtained in S4. If the actual storage quantity meets the raw material requirement, filter out the corresponding ratio factor data as secondary alternative ratio factors. Otherwise, proceed to S6. Select the maximum index data from the performance indicators corresponding to the secondary alternative ratio factors according to their weights. S6. Calculate the total cost of raw material purchase for each alternative ratio scheme and select the scheme with the lowest cost for raw material procurement; and use the corresponding alternative key ratio factor data as the final key ratio factor dataset.

2. The method for managing the proportion of grouting materials for tunnel boring machines based on machine learning as described in claim 1, characterized in that: The key proportioning factors mentioned in S1 include the water-cement ratio, powder-ash ratio, asphalt content, and rubber content.

3. The method for managing the proportion of shield tunneling synchronous grouting materials based on machine learning according to claim 2, characterized in that: The simplified mapping model for the final grout performance index described in S2 takes grout material proportioning factor data as input and outputs grout performance index data as output.

4. The method for managing the proportion of shield tunneling synchronous grouting materials based on machine learning according to claim 3, characterized in that: The input of the final grout performance index standard mapping model described in S3 is the grout material proportioning factor data, and the output is the grout performance index data.

5. The method for managing the proportion of shield tunneling synchronous grouting materials based on machine learning according to claim 4, characterized in that, S4 includes the following steps: S41. Set the data of various key proportioning factors predetermined when preparing the grouting material and input them into the final grouting slurry performance index standard mapping model for mapping; S42. If the performance index data in the mapping result of S41 is not located within the corresponding performance index range, adjust the corresponding initial key ratio factor dataset, and then input it again into the final grouting slurry performance index standard mapping model for mapping to obtain the current adjusted slurry performance index dataset; otherwise, no adjustment is required.

6. The method for managing the proportion of shield tunneling synchronous grouting materials based on machine learning according to claim 5, characterized in that, S4 further includes: S43. Repeat S42 to obtain the current adjusted key ratio factor data corresponding to several groups of performance index data that are all within the corresponding specified intervals, and obtain the current candidate key ratio factor dataset.

7. The method for managing the proportion of shield tunneling synchronous grouting materials based on machine learning according to claim 6, characterized in that, S5 includes the following steps: S51. Based on the current demand for grouting materials and the current dataset of key ratio factors for alternatives, calculate the demand for each type of raw material in the current process of preparing grouting materials, and obtain the current set of raw material demand. S52. Obtain the weight values ​​of various performance indicators of the grouting slurry and the actual storage quantity of each preparation raw material. Then compare the actual storage quantity data of the current raw materials with the preparation demand data of each group of the current preparation raw material demand. If the actual storage quantity of each type of raw material is greater than or equal to the corresponding raw material demand quantity, the key proportioning factor data of the group corresponding to the group of preparation demand quantity data is selected; after the selection is completed, the current secondary candidate key proportioning factor dataset is obtained. Otherwise, execute S6; S53. Based on the current secondary candidate key ratio factor dataset, obtain the corresponding slurry performance index data to obtain the current secondary candidate performance index dataset; select the dataset corresponding to the largest performance index data from the current secondary candidate performance index dataset in descending order of weight to obtain the current final performance index dataset; use the key ratio factor data corresponding to the current final performance index dataset as the current final key ratio factor dataset.

8. The shield tunneling synchronous grouting material ratio management method based on machine learning according to claim 7, characterized in that, S6 includes the following steps: S61. Based on the current raw material demand set and the current raw material actual storage data set, calculate the total cost required to purchase raw materials corresponding to each set of raw material demand, and obtain the current total purchase cost set.

9. The method for managing the proportion of shield tunneling synchronous grouting materials based on machine learning according to claim 8, characterized in that, S6 further includes the following steps: S62. Select the minimum cost group in the current total purchase cost set and purchase the required amount of raw materials. After the purchase is completed, use the key ratio factor data corresponding to the preparation demand data of this group as the current final key ratio factor dataset.

10. A system for implementing the machine learning-based shield tunneling synchronous grouting material ratio management method as described in any one of claims 1-9, characterized in that: It includes an experimental data acquisition module, a simplified mapping model construction module, a standard mapping model construction module, a ratio factor adjustment module, a ratio factor screening module, and a raw material procurement determination module; The experimental data acquisition module is used to acquire key proportioning factors and performance index data in the grouting material proportioning experiment. The simplified mapping model construction module is used to construct a simplified mapping model for grouting slurry performance index data. The standard mapping model construction module is used to construct a standard mapping model for grouting slurry performance index data. The proportioning factor adjustment module is used to repeatedly adjust the initial key proportioning factor data; The ratio factor filtering module is used to filter the acquired key ratio factor data based on the storage capacity; The raw material procurement decision module is used to calculate the total cost of raw material procurement for each alternative ratio scheme and select the scheme with the lowest cost for raw material procurement.

Images