Machine learning based twin-tube pipe-roof construction control method and system

Abstract

The present application relates to the technical field of pipe curtain construction control, and in particular to a double-pipe pipe curtain construction control method and system based on machine learning. Before the double-pipe pipe curtain construction, the system analyzes various basic parameters of the construction soil body in depth, builds a soil body model covering various possibilities according to these parameters, comprehensively and deeply explores the construction quality, safety, efficiency and stability of each construction strategy under different soil body conditions, accurately determines the adaptation ability of each construction strategy, and selects the optimal construction strategy for the double-pipe pipe curtain construction. In the actual construction process of the double-pipe pipe curtain, it is dynamically analyzed whether the construction strategy needs to be intervened and controlled. After it is determined that intervention and control are needed, the most suitable strategy is quickly selected for control by virtue of an efficient and accurate screening mechanism, taking into comprehensive consideration of key factors such as rectification cost, compatibility with existing working conditions and actual performance effect.

Description

Technical Field

[0001] This invention relates to the field of pipe jacking construction control technology, and more specifically, to a machine learning-based method and system for controlling the construction of dual-pipe jackings. Background Technology

[0002] In the field of underground engineering, double-tube jacking construction is widely used in scenarios such as subway crossings and integrated utility tunnel construction due to its advantages such as trenchless construction and good support effect. However, traditional construction methods face significant technical bottlenecks: On the one hand, soil parameters (such as cohesion and internal friction angle) have natural variability and limitations in exploration, and traditional single-parameter modeling cannot cover the diversity of actual working conditions, resulting in significant differences between finite element simulation results and on-site conditions. On the other hand, the selection and adjustment of construction strategies are highly dependent on engineers' experience and lack a quantitative evaluation system. For example, the determination of key parameters such as jacking sequence and load mode is often based on analogy with historical cases, making it difficult to dynamically adapt to fluctuations in soil parameters. This leads to frequent problems such as exceeding jacking force limits and stress concentration in locking mechanisms during construction, requiring multiple work stoppages to adjust strategies, resulting in high project delay rates and overall costs.

[0003] In existing technologies, some studies have attempted to conduct construction simulations using finite element methods, but have not established a quantitative correlation between soil parameter variations and strategy adaptation. While a few intelligent systems have introduced machine learning, they only focus on optimizing a single indicator (such as settlement), lacking coordinated control of multi-dimensional construction objectives (safety, quality, and efficiency). Therefore, how to systematically solve core problems such as soil uncertainty and lag in strategy adjustments, and achieve intelligent and precise management of dual-pipe jacking construction, has become a pressing technical challenge for the industry.

[0004] Based on the above, this application discloses a machine learning-based construction control method and system for dual-tube duct curtains. Summary of the Invention

[0005] In view of the shortcomings of existing technologies, the purpose of this invention is to provide a construction control method and system for dual-tube curtain based on machine learning.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] The construction control method for a dual-tube pipe curtain based on machine learning consists of the following steps:

[0008] Step 1: Construct a soil dual-pipe construction simulation model for each set of soil variation parameters;

[0009] Step 2: Define the construction strategy;

[0010] Step 3: Select and implement the comprehensive strategy;

[0011] Step 4: Periodically determine whether the overall strategy needs to be adjusted;

[0012] Step 5: After the adjustment is determined, switch the target replacement strategy to the execution replacement strategy for dual-pipe curtain construction control.

[0013] Furthermore, the machine learning-based dual-tube duct jacking construction control system includes:

[0014] The construction soil model building unit acquires various basic parameters of the construction soil before the construction of the double-pipe curtain, further generates multiple sets of construction soil variation parameters, and then builds a soil double-pipe construction simulation model for each set of construction soil variation parameters.

[0015] The dual-tube jacking construction strategy adaptation selection unit defines multiple construction strategies, determines the basic construction parameters of the dual-tube jacking, further determines the strategy variation adaptation values ​​of various construction strategies, and then selects the comprehensive execution strategy and carries out the actual dual-tube jacking construction based on the comprehensive execution strategy.

[0016] The construction strategy optimization analysis unit periodically determines whether adjustments to the overall strategy are needed during the construction of the double-pipe jacking.

[0017] The construction strategy optimization control unit determines that an adjustment to the overall strategy is needed. It marks the overall strategy as the target replacement strategy and switches the target replacement strategy to the overall replacement strategy for dual-pipe curtain construction control.

[0018] Furthermore, the process of generating multiple sets of construction soil variation parameters is as follows: determine the parameter variation range of each basic parameter, further generate multiple variation parameters for each basic parameter, and randomly combine the variation parameters of each basic parameter to generate multiple sets of construction soil variation parameters.

[0019] Furthermore, the process for determining the strategy variation adaptation value of the construction strategy is as follows: Select a construction strategy, and import the construction control parameters and foundation construction parameters of the double-pipe jacking corresponding to this construction strategy into the soil double-pipe construction simulation model of each set of construction soil variation parameters. Further obtain the pipe jacking construction evaluation value of the soil double-pipe construction simulation model corresponding to each set of construction soil variation parameters. Accumulate and average the pipe jacking construction evaluation values ​​of the soil double-pipe construction simulation models corresponding to all sets of construction soil variation parameters to calculate the pipe jacking construction evaluation mean. Obtain the mean difference of variation parameter evaluation. Calculate the difference between the pipe jacking construction evaluation mean and the mean difference of variation parameter evaluation to calculate the strategy variation adaptation value of the construction strategy.

[0020] Furthermore, the process for obtaining the mean difference of the variation parameter estimation is as follows: the pipe jacking construction estimation values ​​of the soil double-pipe construction simulation model corresponding to all the soil variation parameter sets are compared pairwise, the absolute difference of each pair of pipe jacking construction estimation values ​​is calculated, the variation parameter estimation difference value is calculated, and all variation parameter estimation difference values ​​are accumulated and averaged to calculate the mean difference of the variation parameter estimation.

[0021] Furthermore, the estimated value of pipe jacking construction corresponding to each set of soil variation parameters in the soil dual-pipe construction simulation model is obtained as follows: The soil dual-pipe construction simulation model is controlled to simulate construction. During the simulation construction, various safety and quality monitoring data of the dual-pipe pipe jacking construction are collected in real time. After the simulation construction is completed, the simulation construction duration is determined. Simultaneously, the collected safety and quality monitoring data are used to extract features. The safety and quality monitoring features are integrated into a comprehensive safety and quality monitoring feature set. The comprehensive safety and quality monitoring feature set is imported into the safety and quality control machine learning model. The safety and quality control machine learning model derives a comprehensive safety and quality assessment value. The ratio of the comprehensive safety and quality assessment value to the simulation construction duration is calculated to obtain the estimated value of pipe jacking construction.

[0022] Furthermore, it is necessary to periodically determine whether adjustments to the overall strategy are required: collect various safety and quality monitoring data of the double-pipe jacking construction in real time, and update the jacking construction assessment value of the double-pipe jacking construction periodically. When the jacking construction assessment value of the double-pipe jacking construction is less than the construction assessment threshold, it is determined that adjustments to the overall strategy are required.

[0023] Furthermore, the process for selecting the replacement strategy is as follows: the remaining construction strategies are marked as control optimization strategies, the control substitution value of each control optimization strategy is further determined, and the control optimization strategy with the largest control substitution value is marked as the replacement strategy to be implemented.

[0024] Furthermore, the specific steps for determining the control substitution values ​​of the control optimization strategy are as follows: Select a control optimization strategy, obtain the various construction control parameters corresponding to the control optimization strategy, perform vector processing on each construction control parameter, and concatenate them in order into a one-dimensional vector A = (a1, a2, ..., a...). N N is the total number of construction control parameters. The process involves obtaining each construction control parameter corresponding to the target replacement strategy, performing vector processing on each parameter, and concatenating them sequentially into a one-dimensional vector B = (b1, b2, ..., b...). N ),pass Calculate the policy similarity, simultaneously obtain the policy variation fitness value of the control optimization policy, multiply the policy variation fitness value and the policy similarity to calculate the control substitution value of the control optimization policy.

[0025] Compared with the prior art, the present invention has the following beneficial effects:

[0026] Before constructing a double-tube jacking system, the system conducts an in-depth analysis of various basic parameters of the soil to be constructed. Based on these parameters, a soil model covering multiple possibilities is built. This allows for a comprehensive and in-depth exploration of the construction quality, safety, efficiency, and stability of various construction strategies under different soil conditions. The system accurately determines the adaptability of each construction strategy and selects the optimal strategy for double-tube jacking construction. During the actual construction of the double-tube jacking system, the system dynamically analyzes whether intervention and control of the construction strategy are necessary. Once intervention and control are determined, the system utilizes an efficient and accurate screening mechanism to quickly select the most suitable strategy for management by comprehensively considering key factors such as rectification costs, compatibility with existing working conditions, and actual performance effects.

[0027] The method of this invention effectively overcomes the problems of insufficient response to soil uncertainty and untimely strategy adjustment in traditional construction methods, significantly improving the overall quality, safety and construction efficiency of double-pipe jacking construction, and providing strong support for intelligent and precise management of underground engineering construction. Attached Figure Description

[0028] Figure 1 This is a flowchart of the method of the present invention;

[0029] Figure 2 This is a flowchart illustrating the determination of the strategy variation adaptation value for the construction strategy of this invention.

[0030] Figure 3 A flowchart for obtaining the estimated value of pipe jacking construction. Detailed Implementation

[0031] Example 1, as Figure 1 The construction control method for a dual-tube duct curtain based on machine learning has the following steps:

[0032] Step 1: Construct a soil dual-pipe construction simulation model for each set of soil variation parameters;

[0033] Step 2: Define the construction strategy;

[0034] Step 3: Select and implement the comprehensive strategy;

[0035] Step 4: Periodically determine whether the overall strategy needs to be adjusted;

[0036] Step 5: After the adjustment is determined, switch the target replacement strategy to the execution replacement strategy for dual-pipe curtain construction control.

[0037] The above methods effectively overcome the problems of insufficient response to soil uncertainty and untimely strategy adjustment in traditional construction methods, significantly improving the overall quality, safety and construction efficiency of double-pipe jacking construction, and providing strong support for intelligent and precise management of underground engineering construction.

[0038] Example 2, as Figures 2 to 3 The machine learning-based dual-tube jacking construction control system includes a construction soil model building unit, a dual-tube jacking construction strategy adaptation and selection unit, a construction strategy optimization and analysis unit, and a construction strategy optimization and control unit.

[0039] The construction soil model building unit acquires various basic parameters of the construction soil before the construction of the twin-tube jacking system. These basic parameters include, but are not limited to, density, cohesion, internal friction angle, elastic modulus, Poisson's ratio, natural moisture content, and compression modulus. These parameters are obtained through relevant experiments on the construction soil before construction. The unit also determines the variation range of each basic parameter. Each basic parameter corresponds to a specific variation range, which represents the reasonable fluctuation range of the corresponding basic parameter during the actual twin-tube jacking construction. This range is used to predict the potential changes in soil parameters during construction and addresses soil uncertainties. The core of this process is to transform the natural variability of soil parameters and the impact of construction disturbance into quantifiable ranges, providing a clear basis for design, construction, and risk control. This involves generating multiple variation parameters for each basic parameter (each variation parameter is unique and randomly selected based on its variation range), randomly combining these variation parameters to generate multiple sets of construction soil variation parameters (each set contains one variation parameter from each basic parameter), and finally building a dual-pipe construction simulation model for each set of construction soil variation parameters.

[0040] The process of building a soil-based dual-pipe construction simulation model with a set of soil variation parameters is as follows: A soil entity is built in general-purpose finite element software based on the set of soil variation parameters. The size of the soil entity is typically 5-8 times the pipe diameter (e.g., if the pipe depth H = 10m, the soil entity height is 50m). The properties of the soil entity are defined, and loads and boundary conditions are applied. Boundary conditions: fixed displacement at the bottom of the soil (Ux = Uy = 0), constrained horizontal displacement on both sides (Ux = 0), and the ground surface is a free boundary. Contact elements are set between the steel pipe and the soil, and the friction coefficient is defined (μ = 0.3-0.5, adjusted according to the soil type). Load application: Soil self-weight: Gravity loads are automatically generated by defining the soil density (ρ), thus building the soil-based dual-pipe construction simulation model.

[0041] The twin-tube jacking construction strategy selection unit defines multiple construction strategies (the strategies are manually defined, and each strategy differs in its construction control parameters, such as jacking load, jacking speed, and the order of twin-tube jacking). For example, strategy A uses simultaneous twin-tube jacking with a jacking load of 900kN applied simultaneously to both tubes at a speed of 8mm / min. Strategy B involves jacking the left tube first, followed by the right tube, with staged loading applied to the left tube with an initial load of 800kN. The load is gradually increased to 1300kN, and then reduced to 1100kN during the dual-pipe collaborative stage. The jacking speed is 10-15mm / min for the left pipe and 12mm / min for the right pipe. The basic construction parameters of the dual-pipe jacking are determined (the construction parameters of the dual-pipe jacking include the outer diameter, wall thickness, construction spacing, burial depth, and equivalent stiffness parameters of the dual-pipe interlocking connection). The strategy variation adaptation values ​​of various construction strategies are further determined. The construction strategy with the largest strategy variation adaptation value is marked as the comprehensive strategy, and the actual dual-pipe jacking construction is carried out according to the comprehensive strategy.

[0042] The process for determining the adaptation value of the construction strategy is as follows: A construction strategy is selected, and the construction control parameters corresponding to this strategy, as well as the foundation construction parameters of the double-pipe jacking, are imported into the soil double-pipe construction simulation model of each set of soil variation parameters. (After import, the soil double-pipe construction simulation model can perform double-pipe jacking construction based on the construction strategy, create double-pipe entities in the soil double-pipe construction simulation model according to the foundation construction parameters of the double-pipe jacking, define the material properties of the double-pipe entities, and perform construction on the double-pipe entities using the construction control parameters corresponding to the construction strategy.) Further, the soil double-pipe construction simulation model corresponding to each set of soil variation parameters is obtained. The estimated value of pipe jacking construction is calculated by summing and averaging the estimated values ​​of pipe jacking construction for the soil dual-pipe construction simulation model corresponding to all sets of soil variation parameters. The estimated values ​​of pipe jacking construction for the soil dual-pipe construction simulation model corresponding to all sets of soil variation parameters are compared pairwise, and the absolute difference between each pair of estimated values ​​is calculated to obtain the variation parameter estimation difference value. The variation parameter estimation difference values ​​are summed and averaged to obtain the variation parameter estimation difference mean value. The difference between the estimated value of pipe jacking construction and the variation parameter estimation difference mean value is calculated to obtain the strategy variation adaptation value of this construction strategy.

[0043] Each set of soil variation parameters corresponds to the pipe jacking construction evaluation value of the soil double-pipe construction simulation model. The acquisition process is as follows: The soil double-pipe construction simulation model is controlled to simulate construction. During the simulation, various safety and quality monitoring data of the double-pipe pipe jacking construction are collected in real time (the safety and quality monitoring data includes surface settlement data, interlocking shear stress, soil horizontal displacement data, axial force at the front end of the steel pipe, and mid-span bending moment; the purpose of collecting these safety and quality monitoring data is to control the safety and quality of the double-pipe pipe jacking construction). After the simulation is completed, the simulation construction duration is determined (the total duration from the start to the completion of the simulation construction), and the data is collected simultaneously. Feature extraction is performed on various safety and quality monitoring data to extract various safety and quality monitoring features (for example, for surface settlement data, features such as maximum settlement and settlement rate can be extracted; for interlocking shear stress, features such as average stress and stress fluctuation amplitude can be extracted). These various safety and quality monitoring features are integrated into a comprehensive safety and quality monitoring feature set. The comprehensive safety and quality monitoring feature set is then imported into a safety and quality control machine learning model. The safety and quality control machine learning model derives a comprehensive safety and quality assessment value. The ratio of the comprehensive safety and quality assessment value to the simulated construction time is calculated to obtain the pipe jacking construction assessment value.

[0044] The steps for building a machine learning model for safety and quality control are as follows: First, collect multiple comprehensive safety and quality monitoring feature sets and build a neural network model. Using these feature sets as the foundation, train the neural network model. During training, assign a comprehensive safety and quality assessment value to each feature set, with the value ranging from 1 to 100. The magnitude of the assessment value has a clear meaning; a larger value indicates better construction quality, higher safety, and higher efficiency in the double-pipe duct construction. Next, divide the multiple feature sets into training, validation, and test sets according to a specific ratio of 70%:15%:15%. Repeatedly train the neural network model using the training set and validate the performance during training using the validation set. Adjust the model parameters promptly based on the validation results, employing hyperparameter tuning, overfitting prevention, and training monitoring. Evaluate the final model using the test set (not used in training) to ensure the results do not rely on data snooping during training. Finally, the machine learning model for safety and quality control is successfully built.

[0045] The construction strategy optimization and analysis unit collects various safety and quality monitoring data in real time during the construction of the double-tube pipe jacking system, and updates the pipe jacking construction assessment value regularly (the corresponding periodic time interval is set according to the actual needs of the double-tube pipe jacking construction). When the pipe jacking construction assessment value is less than the construction assessment threshold (the construction assessment threshold is set comprehensively based on construction standards, risk tolerance, historical data experience, and regulatory requirements), it is determined that the overall strategy needs to be adjusted (otherwise, no adjustment to the overall strategy is required).

[0046] The construction strategy optimization control unit determines that an adjustment to the overall strategy is needed. It marks the overall strategy as the target replacement strategy and marks the other construction strategies as control optimization strategies. It further determines the control substitution value for each control optimization strategy, marks the control optimization strategy with the largest control substitution value as the execution replacement strategy, and switches the target replacement strategy to the execution replacement strategy for dual-pipe curtain construction control.

[0047] The specific steps for determining the control substitution values ​​of the control optimization strategy are as follows: Select a control optimization strategy, obtain the various construction control parameters corresponding to the control optimization strategy, perform vector processing on each construction control parameter, and concatenate them in order into a one-dimensional vector A = (a1, a2, ..., a...). N N represents the total number of construction control parameters (including but not limited to the following: jacking sequence, initial jacking force of the left pipe, maximum jacking force of the left pipe, initial jacking force of the right pipe, maximum jacking force of the right pipe, vector processing method; taking the jacking sequence as an example, the synchronous jacking vector value is 0, and the left-to-right jacking vector value is 1). Obtain the construction control parameters corresponding to the target replacement strategy, perform vector processing on each construction control parameter, and concatenate them in order into a one-dimensional vector B = (b1, b2, ..., b...). N ),pass Calculate the policy similarity, simultaneously obtain the policy variation fitness value of the control optimization policy, multiply the policy variation fitness value and the policy similarity to calculate the control substitution value of the control optimization policy.

[0048] Before constructing the dual-tube jacking system, the system conducts an in-depth analysis of various basic parameters of the soil. Based on these parameters, it builds a soil model covering multiple possibilities. This allows for a comprehensive and in-depth exploration of the construction quality, safety, efficiency, and stability of various construction strategies under different soil conditions. The system accurately determines the adaptability of each construction strategy and selects the optimal strategy for dual-tube jacking construction. During the actual construction of the dual-tube jacking system, the system dynamically analyzes whether intervention and control of the construction strategy are necessary. Once intervention and control are determined, the system utilizes an efficient and accurate screening mechanism to quickly select the most suitable strategy for management by comprehensively considering key factors such as rectification costs, compatibility with existing working conditions, and actual performance effects.

[0049] The above formulas are all dimensionless calculations, and the preset parameters in the formulas should be set by those skilled in the art according to the actual situation.

[0050] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0051] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0052] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0053] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0054] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0055] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0056] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for controlling a double pipe roof construction based on machine learning, characterized in that, The steps are as follows: Step one: before the double-tube pipe curtain construction, obtain the basic parameters of the construction soil, generate a plurality of construction soil variation parameter sets, and build a soil double-tube construction simulation model for each construction soil variation parameter set; The generation process of the plurality of construction soil variation parameter sets is as follows: determine the parameter variation interval of each basic parameter, further generate a plurality of variation parameters for each basic parameter, randomly combine the variation parameters of each basic parameter, and generate a plurality of construction soil variation parameter sets; Step two: define the construction strategy, determine the basic construction parameters of the double-tube pipe curtain, and further determine the strategy variation adaptation value of each construction strategy; The determination process of the strategy variation adaptation value of the construction strategy is as follows: select a construction strategy, import the construction control parameters corresponding to the construction strategy and the basic construction parameters of the double-tube pipe curtain into the soil double-tube construction simulation model of each construction soil variation parameter set, further obtain the pipe curtain construction evaluation value of the soil double-tube construction simulation model corresponding to each construction soil variation parameter set, calculate the average of the pipe curtain construction evaluation values of the soil double-tube construction simulation model corresponding to all construction soil variation parameter sets, calculate the average of the pipe curtain construction evaluation values, obtain the average of the variation parameter evaluation difference, and calculate the difference between the average of the pipe curtain construction evaluation values and the average of the variation parameter evaluation difference to obtain the strategy variation adaptation value of the construction strategy; The process of obtaining the average of the variation parameter evaluation difference is as follows: compare the pipe curtain construction evaluation values of the soil double-tube construction simulation model corresponding to all construction soil variation parameter sets two by two, calculate the absolute difference between each two compared pipe curtain construction evaluation values to obtain the variation parameter evaluation difference value, and calculate the average of all variation parameter evaluation difference values to obtain the average of the variation parameter evaluation difference; The process of obtaining the pipe curtain construction evaluation value of the soil double-tube construction simulation model corresponding to each construction soil variation parameter set is as follows: control the soil double-tube construction simulation model to simulate construction, real-time collect various quality monitoring data of the double-tube pipe curtain during the simulation construction, determine the simulation construction time after the simulation construction is completed, and simultaneously extract the features of the collected various quality monitoring data to obtain various quality monitoring features, integrate the various quality monitoring features into a comprehensive quality monitoring feature set, import the comprehensive quality monitoring feature set into a safety quality control machine learning model, and export a comprehensive quality monitoring evaluation value from the safety quality control machine learning model; calculate the ratio of the comprehensive quality monitoring evaluation value to the simulation construction time to obtain the pipe curtain construction evaluation value; Step three: select the execution comprehensive strategy and execute; Step four: determine whether to adjust the execution comprehensive strategy regularly; Step five: after adjustment, replace the target strategy with the execution replacement strategy for double-tube pipe curtain construction control.

2. The double pipe roof construction control system based on machine learning according to claim 1, applied to the double pipe roof construction control method based on machine learning, characterized in that, It includes: The construction soil model building unit obtains the basic parameters of the construction soil before the double-tube pipe curtain construction, further generates a plurality of construction soil variation parameter sets, and builds a soil double-tube construction simulation model for each construction soil variation parameter set; The double-pipe pipe curtain construction strategy adaptive selection unit defines multiple construction strategies, determines the basic construction parameters of the double-pipe pipe curtain, further determines the strategy anomaly adaptive values of various construction strategies, and then selects an execution comprehensive strategy, and actual double-pipe pipe curtain construction is performed according to the execution comprehensive strategy; The construction strategy optimization analysis unit periodically determines whether the execution comprehensive strategy needs to be adjusted during the double-pipe pipe curtain construction process. The construction strategy optimization control unit marks the execution comprehensive strategy as a target replacement strategy when it is determined that the execution comprehensive strategy needs to be adjusted, and switches the target replacement strategy to an execution replacement strategy for double-pipe pipe curtain construction control.

3. The machine learning based twin-tube pipe-roof construction control system according to claim 2, characterized by, Periodically determine whether the execution comprehensive strategy needs to be adjusted: real-time acquisition of various quality monitoring data of the double-pipe pipe curtain construction, and periodic updating of the pipe curtain construction evaluation value of the double-pipe pipe curtain construction. When the pipe curtain construction evaluation value of the double-pipe pipe curtain construction is less than the construction evaluation threshold, it is determined that the execution comprehensive strategy needs to be adjusted.

4. The machine learning based twin-tube pipe-roof construction control system according to claim 2, characterized by, The selection process of the execution replacement strategy is as follows: mark the remaining construction strategies as control optimization strategies, further determine the control replacement values of each control optimization strategy, and mark the control optimization strategy with the largest control replacement value as the execution replacement strategy.

5. The machine learning based twin-tube pipe-roof construction control system according to claim 2, characterized by, The determination step of the control substitution value of the control optimization strategy is specifically as follows: selecting a control optimization strategy, obtaining each construction control parameter corresponding to the control optimization strategy, performing vector processing on each construction control parameter, and sequentially splicing into a one-dimensional vector A=(a1, a2,..., a N ), N is the total number of construction control parameters, obtaining each construction control parameter corresponding to the target substitution strategy, performing vector processing on each construction control parameter, and sequentially splicing into a one-dimensional vector B=(b1, b2,..., b N ), calculating the strategy similarity through Sim(A, B)= , synchronously obtaining the strategy anomaly adaptation value of the control optimization strategy, multiplying the strategy anomaly adaptation value and the strategy similarity to calculate the control substitution value of the control optimization strategy.

Images