Multi-objective optimization method and system for shield tunneling parameters
The multi-objective optimization method using a random forest and NSGA-II algorithm optimizes shield tunneling parameters, addressing safety, efficiency, and cost challenges by predicting excavation speed and minimizing energy consumption, enhancing the NSGA-II algorithm's performance.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- SHANDONG UNIV
- Filing Date
- 2024-06-17
- Publication Date
- 2026-07-02
AI Technical Summary
Current shield tunneling methods face challenges in ensuring construction safety, efficiency, and cost reduction, particularly in large-diameter tunnels with varying strata and increased mud pressure differences, leading to difficulties in controlling excavation speed and energy consumption.
A multi-objective optimization method combining a random forest algorithm and a non-dominated sorting genetic algorithm (NSGA-II) is employed to predict excavation speed and minimize energy consumption, using parameters like total thrust and cutter head rotation speed, with data preprocessing and an arithmetic crossover operator to enhance optimization.
This approach achieves safe and efficient shield tunneling by optimizing parameters, balancing excavation speed and energy consumption, aligning with practical engineering needs and improving the NSGA-II algorithm's performance.
Smart Images
Figure 2026521962000001_ABST
Abstract
Description
Technical Field
[0001] (Cross - reference to Related Applications) This invention claims the priority of a Chinese patent application with the application number 202310826020.3 and the invention title "Multi - objective Optimization Method and System for Shield Construction Parameters", which was filed with the China National Intellectual Property Administration on July 6, 2023. All of its contents are incorporated into this invention by reference for all purposes and constitute a part of this invention.
[0002] This invention belongs to the technical field of multi - objective optimization of shield parameters, and particularly relates to a multi - objective optimization method and system for shield construction parameters.
Background Art
[0003] The description of this part only shows the information of the background art related to this invention and does not necessarily constitute the prior art.
[0004] The shield method is the first - choice method for urban underground tunnel construction, which has characteristics such as being environmentally friendly, safe, and highly efficient, and is widely used in urban tunnel construction in China. The shield method is to propel the shield machine along the designed axis in the stratum and support the surrounding rock mass by the shield shell and segments to prevent the tunnel from collapsing. Compared with conventional tunnels of about 6 meters, in large - diameter shield tunnels, the difference in the mud pressure between the top and bottom of the cutter head further increases, and the compatibility with the water pressure and soil pressure of the stratum decreases, so the effective support for the excavation surface decreases. Also, with the increase in the shield diameter, the probability that the shield crosses composite strata also increases significantly, so the requirements for the adaptability of the shield structure and the corresponding construction technology increase.
[0005] In current underground space development, safe and efficient excavation using shields is an urgent issue. Maintaining real-time face stability during shield excavation requires sufficient bearing pressure within the mud chamber, but the excavation effect varies depending on the combination of excavation parameters. One particularly noteworthy issue is the energy consumption of the shield machine during excavation, which determines the cost of shield construction. Furthermore, the excavation speed of the shield machine is a crucial parameter for safe and efficient shield excavation, clearly guiding engineering scheduling and cost management. If the excavation speed is too high, shield attitude control becomes difficult, leading to greater surface undulation; however, if the excavation speed is too low, it affects construction efficiency.
[0006] Current research has shown that while the optimization goal for shield tunneling parameters is relatively simple, in actual engineering projects, it is impossible to ensure construction safety while simultaneously considering construction efficiency and cost reduction. [Overview of the Initiative]
[0007] To solve the technical challenges present in the background technology described above, the present invention provides a multi-objective optimization method and system for shield tunneling parameters. By combining a random forest algorithm and a multi-objective optimization algorithm, the excavation speed predicted based on the random forest is used as the first fitness function of the multi-objective optimization algorithm, and the shield tunneling energy consumption ratio is used as the second fitness function, thereby achieving multi-objective optimization of shield tunneling parameters, while ensuring construction safety and considering construction efficiency and cost reduction.
[0008] To achieve the above objective, the present invention employs the following technical solutions.
[0009] A first aspect of the present invention provides a multi-objective optimization method for shield tunneling parameters.
[0010] A multi-objective optimization method for shield construction parameters, Steps to obtain initial shield construction parameters, The method includes the steps of optimizing shield construction parameters using a multi-objective optimization algorithm, predicting the excavation speed using a random forest based on the shield construction parameters, calculating the shield excavation energy consumption ratio based on the excavation speed and shield construction parameters, and using the maximization of the excavation speed and the minimization of the shield excavation energy consumption ratio as the two fitness functions of the multi-objective optimization algorithm.
[0011] Furthermore, the Spearman correlation coefficient is used to identify multiple shield construction parameters that have the highest correlation with the target excavation speed, and the excavation speed is predicted using a random forest based on the identified shield construction parameters.
[0012] Furthermore, the identified shield construction parameters include total thrust, cutter head rotation speed, cutter head torque, cutter head compression force, face pressure, penetration degree, tunnel burial depth, cohesion, and internal friction angle.
[0013] Furthermore, the shield tunneling energy consumption ratio is related to the tunneling speed, the total thrust, cutter head rotation speed, and cutter head torque among the shield construction parameters, and the shield machine cutter head radius.
[0014] Furthermore, the training of the aforementioned random forest is The steps include obtaining shield construction parameters and excavation speed, dividing the section into a blank section, an ascent section, a stabilization section, and a ring construction section, and then selecting the shield construction parameters and excavation speed for the stabilization section and the ring construction section. The selected data is subjected to the following steps: outlier detection using an isolated forest, replacement of detected outliers with missing values, filling in missing data using the mean method, and then normalization. This includes the step of training a random forest using normalized data.
[0015] Furthermore, a non-controlling genetic algorithm is used as the multi-objective optimization algorithm, and this non-controlling genetic algorithm uses an arithmetic crossover operator when crossing over populations.
[0016] Furthermore, the non-dominant genetic algorithm introduces a cumulative ranking fitness strategy during Pareto ranking.
[0017] A second aspect of the present invention provides a multi-objective optimization system for shield tunneling parameters.
[0018] A multi-objective optimization system for shield construction parameters, A data acquisition module configured to obtain initial shield construction parameters, The system includes an optimization module configured to optimize shield tunneling parameters using a multi-objective optimization algorithm, predict the tunneling speed using a random forest based on the shield tunneling parameters, calculate the shield tunneling energy consumption ratio based on the tunneling speed and shield tunneling parameters, and set the maximization of the tunneling speed and the minimization of the shield tunneling energy consumption ratio as the two fitness functions of the multi-objective optimization algorithm.
[0019] A third aspect of the present invention provides a computer-readable storage medium.
[0020] A computer-readable storage medium in which a computer program is stored, wherein when the program is executed by a processor, the steps of the multi-objective optimization method for shield construction parameters described in the first embodiment are realized.
[0021] A fourth aspect of the present invention provides a computer device.
[0022] A computer device comprising a memory, a processor, and a computer program stored in the memory and executable by the processor, wherein when the program is executed by the processor, the steps of the multi-objective optimization method for shield construction parameters according to the first aspect are realized.
[0023] Compared with the prior art, the beneficial effects of the present invention are as follows. The present invention uses a multi-objective optimization algorithm for shield tunneling parameters. During shield construction, not only construction efficiency but also construction cost needs to be considered. Since there are many shield construction parameters, the RF algorithm and NSGA-II are combined. The tunneling speed predicted based on RF is used as the first fitness function of NSGA-II, and the shield tunneling energy consumption ratio is used as the second fitness function to realize the multi-objective optimization of shield construction parameters, provide an optimal parameter control range while meeting multiple goals, consider construction efficiency and construction cost reduction while ensuring construction safety, and is more in line with the practical significance of engineering compared with single-objective optimization.
[0024] The present invention improves the NSGA-II algorithm. Instead of the SBX crossover operator in NSGA-II, an arithmetic crossover operator is used to improve the search performance of the algorithm. At the same time, a cumulative ranking fitness strategy is introduced to improve the global search ability of the algorithm and the uniformity of the Pareto front, and finally an optimal solution set for multiple objectives is obtained. Based on the optimal solution, an optimized range of tunneling parameters is obtained to assist shield construction, thereby enabling safe and efficient shield tunneling.
[0025] The present invention uses the Spearman correlation coefficient to select the parameters required for the model from the data, and improves the quality and availability of the data through data preprocessing.
Brief Description of the Drawings
[0026] The drawings in the specification, which constitute part of the present invention, are for the purpose of further understanding the present invention, and the exemplary embodiments and descriptions thereof are for the purpose of interpretation of the present invention and are not intended to improperly limit the present invention. [Figure 1] This is a flowchart of the multi-objective optimization method for shield construction parameters shown in Embodiment 1 of the present invention. [Figure 2] This diagram illustrates the effect of predicting the excavation speed using the RF model shown in Example 1 of the present invention. [Figure 3] This is a flowchart illustrating the improvement of the NSGA-II algorithm model shown in Embodiment 1 of the present invention. [Modes for carrying out the invention]
[0027] The present invention will be further described below with reference to drawings and embodiments.
[0028] It should be noted that the following detailed description is illustrative and intended to further illustrate the invention. Unless otherwise indicated, all technical and scientific terms used herein have the same meaning as those generally understood by those skilled in the art to which the invention pertains.
[0029] It should be noted that the terms used herein are merely for describing specific embodiments and are not intended to limit the exemplary embodiments of the present invention. For example, unless otherwise specified in the context, the singular form used herein is intended to include plural forms, and it should also be understood that when the terms “equip” and / or “include” are used herein, it indicates the presence of features, steps, operations, devices, assemblies and / or combinations thereof.
[0030] Note that the flowcharts and block diagrams in the drawings illustrate the feasible system architectures, functions, and operations of the methods and systems of various embodiments of this disclosure. Note that each block in a flowchart or block diagram may represent a module, program segment, or portion of code, and such module, program segment, or portion of code may contain one or more executable instructions for implementing the logical function specified in each embodiment. Note that in some alternative implementations, the functions represented in a block may be implemented in a different order than that shown in the drawings. For example, two consecutive blocks may be executed substantially simultaneously, or, depending on the function, may be executed in reverse order. Note that each block in a flowchart and / or block diagram, and combinations of blocks in a flowchart and / or block diagram, may be implemented by a dedicated system based on hardware that performs the specified function or operation, or by a combination of dedicated hardware and computer instructions.
[0031] Example 1 This embodiment provides a multi-objective optimization method for shield tunneling parameters.
[0032] The multi-objective optimization method for shield construction parameters provided in this embodiment includes the steps of: obtaining initial shield construction parameters; optimizing the shield construction parameters using a multi-objective optimization algorithm; predicting the excavation speed using a random forest based on the shield construction parameters in the optimization of the shield construction parameters; calculating the shield excavation energy consumption ratio based on the excavation speed and the shield construction parameters; and using the maximization of the excavation speed and the minimization of the shield excavation energy consumption ratio as the two fitness functions of the multi-objective optimization algorithm.
[0033] Here, the initial shield construction parameters may be generated randomly or set manually.
[0034] The multi-objective optimization method for shield tunneling parameters provided in this embodiment selects the parameters required for the model from the data and improves the quality and availability of the data through data preprocessing. Prediction is performed using a random forest (RF), and its hyperparameters are adjusted to improve prediction accuracy, which can then be used as the fitness function for multi-objective optimization. The tunneling velocity predicted based on the RF is used as the first fitness function of NSGA-II (Non-dominated Sorting Genetic Algorithms, NSGA), and the shield tunneling energy consumption ratio is used as the second fitness function to obtain a multi-objective solution set. By introducing an arithmetic crossover operator strategy and a cumulative ranking fitness strategy into the NSGA-II algorithm, faster convergence and preservation of population diversity are achieved. Finally, an optimal solution set for multiple objectives is obtained based on a Pareto front, providing a reasonable range of parameter control, thereby supporting shield tunneling.
[0035] The multi-objective optimization method for shield construction parameters provided in this embodiment specifically includes the following steps:
[0036] Step 1: Obtain shield construction parameters and excavation speed, then split the data. The data for each ring acquired from the sensors mainly consists of four parts: the empty section, the rising section, the stabilization section, and the ring construction section. The values for the empty section and the ring construction section are almost zero, making analysis meaningless. We select and analyze the data for the rising section and the stabilization section of 10 rings, that is, we select the shield construction parameters and excavation speed for the stabilization section and the ring construction section.
[0037] Here, the shield excavation process involves numerous parameters, including shield excavation parameters, tunnel geometry parameters, and geological parameters. While sensors continuously record data, it is meaningless to analyze data recorded when the shield is shut down, only selecting data recorded during operation.
[0038] Step 2: Outlier Handling For the selected data, outliers are detected using an isolated forest, the anomalies detected by the isolated forest are replaced with missing values, and then the missing data is filled in using the mean method.
[0039] The isolated forest is an unsupervised learning algorithm based on decision trees. The algorithm's idea is to identify samples distributed outside a high-density population as anomalies. That is, it extracts some samples and places them at the root node of a single tree, sets a cut value, and divides the tree until it can no longer be divided or reaches a set height, and identifies the first isolated group of samples as anomaly samples. The above operation is performed on each tree in the forest, the division results for the entire forest are aggregated, and the anomaly score for each sample is calculated based on the following formula.
number
number
[0040] Sudden shutdowns or anomalies in sensor monitoring data may cause some outliers, but these outliers are removed so as not to affect subsequent model predictions.
[0041] Step 3: Data Normalization Formula
number
[0042] Step 4 The correlation between shield parameters and target excavation speed in the dataset is identified using the Spearman correlation coefficient, and the Spearman correlation coefficient is calculated as follows.
number
number
number
[0043] Step 5 Through correlation analysis, variables related to the target excavation speed are obtained and used as key parameters for rationally controlling the shield tunneling operation. Total thrust, cutter head rotation speed, cutter head torque, cutter head compression force, face pressure, penetration degree, tunnel burial depth, cohesion, and internal friction angle are used as input parameters, and the excavation speed is predicted as the output parameter.
[0044] Step 6 When inputting the data processed in steps 1-3 into the RF regression model and training the samples with the RF algorithm, it is necessary to adjust the model's key hyperparameters. Improving the model's regression fitting performance can be achieved by adjusting the number of decision trees, maximum number of features, and maximum depth.
[0045] The RF hyperparameter settings, which directly affect the model's prediction accuracy, include the number of decision trees (n_estimators), the maximum number of features (max_features), and the maximum depth (max_depth). The prediction effect on the test set is shown in Figure 2. The model's quantitative metrics include the coefficient of determination (R) of the training set. 2 The value is 0.9057, and the test set R 2 It was found that the value is 0.8991, demonstrating that the model has very high generalization ability.
[0046] To verify the predictive performance of the RF model, we used the coefficient of determination R, which is a quantitative metric. 2 Select this option. The formula for calculating the performance indicator is as follows:
number
number
number
[0047] Step 7 The specific energy consumption ratio of shield tunneling is the energy consumed by the shield machine per unit of excavated volume, and is also called the specific energy (SE) of shield tunneling. A portion of this is consumed by the cutter during soil excavation. T The remaining portion is consumed by frictional force between the shield shell and the soil mass, and the formula for this is as follows:
number
[0048] From an RF-NSGA-II optimization perspective, the soil and geometric parameters of a particular engineering project are fixed. To achieve the goals of minimizing the specific energy of excavation and maximizing the excavation speed in shield tunneling, the construction parameters need to be optimized.
[0049] Step 8 After training with RF, the relationship between shield tunneling parameters and tunneling speed is obtained, and a first objective function, maxF2, which determines the tunneling speed, is established. Subsequently, a second objective function, minF1, which determines the specific energy of shield tunneling, is established.
[0050] Specifically, RF is used to accurately predict the drilling speed, and the first fitness function of NSGA-II is constructed using the nonlinear relationship between the shield parameters and the target.
[0051] Specifically, the specific energy of shield tunneling, calculated empirically, is used as the second fitness function. Constraints are then set on each variable during the plan generation process to ensure the generated plan can be rationally implemented, thus forming the constraints for the variables. This is to prevent large adjustment ranges and potential safety issues arising from significant differences between optimized shield machine parameters and actual shield parameters.
[0052] Step 9 When optimizing construction parameters, it is necessary to define the decision-making range of the population and constrain its variables so that the initial population has actual meaning.
[0053] Step 10 Execute the NSGA-II algorithm.
[0054] In research on multi-objective optimization in other fields, various algorithms such as MODE and MOGA are used. The NSGA-II algorithm is relatively widely used and can generate optimal solutions and support decision-makers' decision-making, but the search performance of the crossover operator used is low, and there are certain limitations in terms of population diversity and convergence speed.
[0055] In this embodiment, to make the solutions obtained by the NSGA-II algorithm more uniform, an arithmetic crossover operator is used instead of the SBX crossover operator in NSGA-II during population crossover. This preserves the superior characteristics of parent individuals, improves the quality of individuals after crossover, and enhances the search performance of the algorithm. A cumulative ranking fitness strategy is introduced during Pareto ranking to prevent dilution of the fitness values of superior individuals, thereby achieving faster convergence and maintaining population diversity, ultimately obtaining a set of optimal solutions for the Pareto front across multiple objectives. The algorithm flowchart is shown in Figure 3, where the population size, tournament size, crossover probability, and number of evolutionary generations are set according to the complexity of the objectives to obtain the Pareto front.
[0056] Here, the arithmetic crossover operator is the corresponding gene of the offspring obtained by weighting the corresponding genes of the two parent individuals. Specifically, assuming that certain genes of the two parent individuals P1 and P2 are g(1) and g(2), respectively, then the corresponding gene of the offspring is
number
[0057] Step 11 The NSGA-II algorithm obtains the optimal solution, and by controlling the shield tunneling parameters within the optimization range while satisfying two objectives, the safety and efficiency of shield tunneling are enhanced.
[0058] Example 2 This embodiment provides a multi-objective optimization system for shield tunneling parameters.
[0059] A multi-objective optimization system for shield construction parameters, A data acquisition module configured to obtain initial shield construction parameters, The system includes an optimization module configured to optimize shield tunneling parameters using a multi-objective optimization algorithm, predict the tunneling speed using a random forest based on the shield tunneling parameters, calculate the shield tunneling energy consumption ratio based on the tunneling speed and shield tunneling parameters, and set the maximization of the tunneling speed and the minimization of the shield tunneling energy consumption ratio as the two fitness functions of the multi-objective optimization algorithm.
[0060] It should be noted that the above module implements the same examples and application scenarios as in Example 1, but is not limited to what is disclosed in Example 1. It should also be noted that the module may run as part of a system on a computer system, such as a computer executable instruction set.
[0061] Example 3 This embodiment provides a computer-readable storage medium in which a computer program is stored, and when the program is executed by a processor, the steps of the multi-objective optimization method for shield construction parameters described in Embodiment 1 are realized.
[0062] Example 4 This embodiment provides a computer device comprising memory, a processor, and a computer program stored in memory and executable by the processor, wherein when the program is executed by the processor, the steps of the multi-objective optimization method for shield construction parameters described in Embodiment 1 are realized.
[0063] The foregoing description represents only preferred embodiments of the present invention and is not intended to limit the invention. Those skilled in the art will know that the present invention can be modified and altered in various ways. Any modifications, equivalent substitutions, improvements, etc., made without departing from the spirit and principles of the present invention shall be within the scope of protection of the present invention.
Claims
1. Steps to obtain initial shield construction parameters, A method for multi-objective optimization of shield tunneling parameters, characterized by comprising the steps of optimizing shield tunneling parameters using a multi-objective optimization algorithm, predicting the excavation speed using a random forest based on the shield tunneling parameters, calculating the shield tunneling energy consumption ratio based on the excavation speed and shield tunneling parameters, and using the maximization of the excavation speed and the minimization of the shield tunneling energy consumption ratio as the two fitness functions of the multi-objective optimization algorithm.
2. A multi-objective optimization method for shield construction parameters according to claim 1, characterized by identifying multiple shield construction parameters that have the highest correlation with the target excavation speed using the Spearman correlation coefficient, and predicting the excavation speed using a random forest based on the identified shield construction parameters.
3. The multi-purpose optimization method for shield construction parameters according to claim 2, characterized in that the identified shield construction parameters include total thrust, cutter head rotation speed, cutter head torque, cutter head compression force, face pressure, penetration degree, tunnel burial depth, cohesion, and internal friction angle.
4. The shield tunneling energy consumption ratio is related to the tunneling speed and to the total thrust, cutter head rotation speed, cutter head torque, and shield machine cutter head radius among the shield tunneling parameters, as described in claim 1, for a multi-purpose optimization method for shield tunneling parameters.
5. The training of the aforementioned random forest is The steps include obtaining shield construction parameters and excavation speed, dividing the section into a blank section, an ascent section, a stabilization section, and a ring construction section, and then selecting the shield construction parameters and excavation speed for the stabilization section and the ring construction section. The selected data is subjected to the following steps: outlier detection using an isolated forest, replacement of detected outliers with missing values, filling in missing data using the mean method, and then normalization. A multi-objective optimization method for shield construction parameters according to claim 1, characterized by comprising the step of training a random forest using normalized data.
6. The multi-objective optimization method for shield construction parameters according to claim 1, characterized in that a non-controlling genetic algorithm is used as the multi-objective optimization algorithm, and the non-controlling genetic algorithm uses an arithmetic crossover operator when crossing over populations.
7. The multi-objective optimization method for shield construction parameters according to claim 6, characterized in that the non-inhibitory genetic algorithm introduces a cumulative ranking fitness strategy during Pareto ranking.
8. A data acquisition module configured to obtain initial shield construction parameters, A multi-objective optimization system for shield tunneling parameters, comprising: an optimization module configured to optimize shield tunneling parameters using a multi-objective optimization algorithm, predict the tunneling speed using a random forest based on the shield tunneling parameters in the optimization of the shield tunneling parameters, calculate the shield tunneling energy consumption ratio based on the tunneling speed and the shield tunneling parameters, and set the maximization of the tunneling speed and the minimization of the shield tunneling energy consumption ratio as two fitness functions of the multi-objective optimization algorithm.
9. A computer-readable storage medium in which a computer program is stored, characterized in that when the program is executed by a processor, the steps of the multi-objective optimization method for shield construction parameters described in any one of claims 1 to 7 are realized.
10. A computer device comprising memory, a processor, and a computer program stored in memory and executable by the processor, wherein when the program is executed by the processor, the steps of the multi-objective optimization method for shield construction parameters described in any one of claims 1 to 7 are realized.