A shield machine excavation error compensation control method and system

Abstract

The present invention relates to a shield machine excavation error compensation control method and system, which relates to the field of tunnel excavation operations, including: performing time domain alignment and cutting area mapping alignment on the predicted soil characteristic distribution sequence and the predicted tool wear distribution sequence, integrating to obtain the predicted operation status data sequence; based on the predicted operation status data sequence, taking the minimum error of the overall excavation operation of the cutter head as the expected goal, iteratively optimizing and analyzing the excavation control parameters of the shield machine until the predetermined optimization scale is met, outputting the optimal excavation parameter sequence, and controlling the shield machine to perform tunnel construction in the future time zone. The present invention can solve the technical problems that the traditional method lacks a dynamic adjustment mechanism, cannot quickly and specifically adjust the excavation control parameters according to soil changes and tool wear conditions, resulting in insufficient excavation accuracy and low operating efficiency; by effectively reducing the excavation operation error, the accuracy and operating efficiency of the shield machine in the excavation process can be significantly improved.

Description

Technical Field

[0001] The present invention relates to the field of tunnel excavation operations, and in particular to a shield machine excavation error compensation control method and system. Background Art

[0002] Tunnel excavation is an important part of modern urban construction, subway construction, railway construction and other underground projects. In tunnel excavation, the shield machine, as the main construction equipment, undertakes complex tasks such as tunnel excavation, support and lining.

[0003] At present, traditional methods usually use fixed excavation parameters to control the operation of shield machines. However, factors such as the hardness, humidity and distribution of rock layers of different soil layers often cause the shield machine to be unable to adapt to soil changes during the excavation process. At the same time, changes in soil quality directly affect the wear of the cutter, and traditional excavation methods cannot adjust the working parameters in real time to cope with these changes. Therefore, when encountering soil changes, shield machines are prone to excavation errors, which in turn affects the accuracy and efficiency of tunnel construction. Summary of the invention

[0004] The present invention aims to solve the technical problem that traditional shield machine excavation control methods lack a dynamic adjustment mechanism and are unable to quickly and specifically adjust excavation control parameters according to soil changes and tool wear, resulting in insufficient excavation accuracy and low operating efficiency. A shield machine excavation error compensation control method and system are provided to solve the problem.

[0005] The technical solution of the present invention to solve the above technical problems is as follows:

[0006] In the first aspect, the present invention provides a shield machine excavation error compensation control method, comprising: during the shield machine excavation construction process, using an acoustic survey instrument to collect and record soil characteristic data in the tunnel at a fixed point, obtain the soil characteristic distribution sequence of the historical time zone, and predict the predicted soil characteristic distribution sequence of the future time zone based on the soil characteristic distribution sequence; using a sensor to monitor and collect the tool status information of several tools in the shield machine cutter head, obtain the tool wear distribution sequence of the historical time zone, and predict the predicted tool wear distribution sequence of the future time zone based on the tool wear distribution sequence; performing time domain alignment and cutting area mapping alignment on the predicted soil characteristic distribution sequence and the predicted tool wear distribution sequence, and integrating them to obtain a predicted operation status data sequence, wherein the cutting area mapping alignment is performed according to the interval between the tool and the center of the cutter head; based on the predicted operation status data sequence, with the minimum error of the overall excavation operation of the cutter head as the expected goal, the excavation control parameters of the shield machine are iteratively optimized and analyzed until the predetermined optimization scale is met, and the optimal excavation parameter sequence is output to control the shield machine to perform tunnel construction in the future time zone.

[0007] In the second aspect, the present invention provides a shield machine excavation error compensation control system, including: a predicted soil characteristic distribution sequence acquisition unit, which is used to collect and record soil characteristic data in the tunnel through an acoustic survey instrument at a fixed point during the shield machine excavation construction process, obtain the soil characteristic distribution sequence of the historical time zone, and predict the predicted soil characteristic distribution sequence of the future time zone based on the soil characteristic distribution sequence; a predicted tool wear distribution sequence acquisition unit, which is used to collect tool status information of several tools in the shield machine cutter disc through a sensor fixed-point monitoring, obtain the tool wear distribution sequence of the historical time zone, and predict the tool wear distribution sequence of the future time zone based on the tool wear distribution sequence Predicted tool wear distribution sequence; predicted operation status data sequence obtaining unit, used for time domain alignment and cutting area mapping alignment of the predicted soil characteristic distribution sequence and the predicted tool wear distribution sequence, and integrating them to obtain the predicted operation status data sequence, wherein the cutting area mapping alignment is performed according to the interval between the tool and the center of the cutter disc; tunneling control parameter iterative optimization unit, used for iterative optimization analysis of the tunneling control parameters of the shield machine based on the predicted operation status data sequence, with the minimum error of the overall tunneling operation of the cutter disc as the expected goal, until the predetermined optimization scale is met, and the optimal tunneling parameter sequence is output to control the shield machine to execute tunnel construction in the future time zone.

[0008] The beneficial effects of the present invention are as follows: during the tunneling construction of the shield machine, the soil characteristic data in the tunnel is collected and recorded at fixed points by an acoustic survey instrument to obtain the soil characteristic distribution sequence of the historical time zone, and the predicted soil characteristic distribution sequence of the future time zone is predicted based on the soil characteristic distribution sequence; on the other hand, the tool status information of several tools in the cutter head of the shield machine is collected by fixed-point monitoring by a sensor to obtain the tool wear distribution sequence of the historical time zone, and the predicted tool wear distribution sequence of the future time zone is predicted based on the tool wear distribution sequence; then the predicted soil characteristic distribution sequence and the predicted tool wear distribution sequence are aligned in the time domain and aligned in the cutting area mapping, and integrated to obtain the predicted operation status data sequence, wherein the tool state data is obtained according to the distance between the tool and the cutter. The cutting area mapping is aligned with the interval of the cutter head center; further based on the predicted operation status data sequence, the tunneling control parameters of the shield machine are iteratively optimized and analyzed with the minimum error of the overall tunneling operation of the cutter head as the expected goal, until the predetermined optimization scale is met, and the optimal tunneling parameter sequence is output; finally, the shield machine is controlled to execute tunnel construction in the future time zone according to the optimal tunneling parameter sequence; that is to say, by introducing a dynamic adjustment mechanism based on real-time monitoring and prediction, the tunneling control parameters can be quickly and accurately optimized according to future soil changes and tool wear, and the automation and intelligence of the tunneling control of the shield machine can be improved, thereby effectively reducing the error of tunneling operations, significantly improving the accuracy and operating efficiency of the shield machine during tunneling, and realizing efficient and accurate tunnel construction. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 A schematic flow chart of a shield machine excavation error compensation control method provided by the present invention;

[0010] Figure 2 A schematic structural diagram of a shield machine excavation error compensation control system provided by the present invention.

[0011] In the accompanying drawings, the components represented by the reference numerals are described as follows:

[0012] Prediction of soil characteristics distribution sequence acquisition unit 01, prediction of tool wear distribution sequence acquisition unit 02, prediction of operation status data sequence acquisition unit 03, excavation control parameter iterative optimization unit 04. DETAILED DESCRIPTION

[0013] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0014] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0015] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or explanation". Any embodiment described as "for example" in the present invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present invention.

[0016] Embodiment 1, as Figure 1 As shown, an embodiment of the present invention provides a shield machine excavation error compensation control method, comprising the following steps:

[0017] A100: During the tunneling construction process of the shield machine, the soil characteristic data in the tunnel is collected and recorded at fixed points through an acoustic survey instrument to obtain the soil characteristic distribution sequence of the historical time zone, and the predicted soil characteristic distribution sequence of the future time zone is predicted based on the soil characteristic distribution sequence.

[0018] Furthermore, step A100 of the present invention further includes:

[0019] A110: The soil characteristic data in the tunnel are collected and recorded at fixed points by using an acoustic wave survey instrument, the distribution of P soil characteristics at P time points in the historical time zone is obtained, and a soil characteristic distribution sequence is generated, where the soil characteristics include soil type, soil hardness, soil structure and moisture content, and P is an integer greater than 10.

[0020] Specifically, during the tunneling construction process of the shield machine, changes in soil quality have a direct impact on tunneling efficiency and accuracy. In particular, when encountering uneven or complex soil layers (such as soft soil, clay, sand or rock layers), the shield machine may encounter greater propulsion resistance or increased tool wear. Therefore, it is necessary to grasp the characteristics of the underground soil in real time in order to adjust the tunneling parameters and optimize the operation strategy in time; acoustic wave survey technology can efficiently obtain relevant data on underground soil by emitting sound waves and analyzing their propagation characteristics in the soil layer. It has the advantages of being non-invasive, efficient and accurate, and has gradually become an important monitoring method in tunneling construction.

[0021] During the tunneling process of the shield machine, the soil in the tunnel is monitored regularly (according to the predetermined monitoring nodes, such as collecting data every 3 minutes) using an acoustic survey instrument, and soil characteristic data are collected through the differences in the propagation speed and reflection of sound waves in different soil layers, wherein the soil characteristic data include soil type (such as soft soil, sand, clay, rock layer, etc.), soil hardness, soil structure (such as particle size, particle arrangement, etc.) and moisture content; a historical time zone is set, that is, the soil in the past period of time is regularly collected during the tunneling process, and the historical time zone includes P time points, which can be set according to the actual operation scenario. For example, if data is collected every 3 minutes within the last hour, P is 20; P soil characteristic distributions at P time points in the historical time zone are obtained (including soil characteristic information at multiple locations in the tunneling area); the P soil characteristic distributions are further arranged in the order of collection time to generate a soil characteristic distribution sequence.

[0022] A120: Use a generative adversarial network to construct a soil characteristic change prediction model, and input the soil characteristic distribution sequence into the soil characteristic change prediction model, and output a predicted soil characteristic distribution sequence for a future time zone, wherein the predicted soil characteristic distribution sequence includes P predicted soil characteristic distributions at P time points in the future time zone.

[0023] Further, step A120 of the present invention further includes:

[0024] A121: Based on the historical excavation operation log of the shield machine, with P time points as screening constraints, collect a set of sample soil characteristic distribution sequence sets, and collect the soil characteristic distribution of each sample soil characteristic distribution sequence at the next P time points to obtain a set of sample predicted soil characteristic distribution sequence sets; A122: Using the sample soil characteristic distribution sequence set and the sample predicted soil characteristic distribution sequence set, iteratively train the generator and discriminator of the generative adversarial network until both the generation loss function and the discriminant loss function converge, and obtain the soil characteristic change prediction model.

[0025] Specifically, query the historical tunneling operation log of the shield machine, including the soil characteristic data collected in real time during each shield machine tunneling operation. These data usually include parameters such as soil type, hardness, structure, moisture content, etc., and are recorded in the time sequence of tunneling. Further, with P time points as constraints, the historical tunneling operation log is screened, that is, data points that meet the constraints are selected from the historical log. These time points are the soil characteristic data collected by the shield machine every P time intervals during the tunneling process, ensuring that the selected data is representative and complete, and obtaining a sample soil characteristic distribution sequence set. Then, the soil characteristic distribution of each sample soil characteristic distribution sequence at the next P time points is collected and set as a sample predicted soil characteristic distribution sequence, where the sample predicted soil characteristic distribution sequence corresponds to the sample soil characteristic distribution sequence one by one, and a sample predicted soil characteristic distribution sequence set is obtained.

[0026] Generative adversarial network is a model composed of generator and discriminator. In this scheme, the change of soil characteristics is predicted by using generative adversarial network. Generative adversarial network includes generator and discriminator. The goal of generator is to generate a predicted soil characteristic distribution sequence that conforms to the law of future soil change from the input noise or historical soil characteristic data. Generator tries to create soil characteristic prediction data as close to the real one as possible. The goal of discriminator is to distinguish the prediction data generated by generator from the actual sample prediction data. The discriminator will output a probability value, indicating whether the input data comes from the real sample prediction data set or the forged data from the generator. Then, the sample soil characteristic distribution sequence set and the sample predicted soil characteristic distribution sequence set are used as sample training data to iteratively train the generator and the discriminator, wherein the generator and the discriminator play a game during the training process, the generator deceives the discriminator by generating more and more realistic predicted data, and the discriminator continuously learns how to distinguish between real data and generated data. Through this game, the generator and the discriminator will continuously optimize their own parameters. During the training process, the generator and the discriminator have their own loss functions. The generation loss function measures the difference between the predicted soil characteristic sequence generated by the generator and the real soil characteristic distribution sequence. The goal is to minimize the generation loss so that the generated soil prediction sequence is as close to the real data as possible. The discriminant loss function measures the ability of the discriminator to distinguish between generated data and real data. The goal is to minimize the discriminant loss and improve the accuracy of the discriminator's judgment.

[0027] During the training process, the generator and the discriminator are updated alternately. The generator continuously improves the soil characteristic prediction sequence it generates, while the discriminator continuously learns how to identify false prediction data. Through multiple iterations, the soil prediction sequence generated by the generator becomes closer and closer to the real data, and the discriminator's discrimination ability becomes stronger and stronger. Repeat the iterative training until both the generation loss function and the discriminant loss function reach a certain convergence standard, that is, the generator can generate sufficiently real soil characteristic prediction data, and the discriminator cannot effectively distinguish between the generated data and the real data. At this time, the trained soil characteristic change prediction model is obtained.

[0028] Then the soil characteristic distribution sequence is input into the soil characteristic change prediction model for prediction, and a predicted soil characteristic distribution sequence for the future time zone is output, wherein the predicted soil characteristic distribution sequence includes P predicted soil characteristic distributions at P time points in the future time zone.

[0029] A200: The tool status information of several cutting tools in the cutter head of the shield machine is collected through fixed-point monitoring by sensors, the tool wear distribution sequence of the historical time zone is obtained, and the predicted tool wear distribution sequence of the future time zone is obtained based on the tool wear distribution sequence prediction.

[0030] Further, step A200 of the present invention further includes:

[0031] A210: Collect tool status information of several cutting tools in the cutter head of the shield machine through fixed-point monitoring by sensors, wherein the tool status information includes tool temperature, pressure data and vibration characteristics; A220: Evaluate tool wear status based on the tool temperature, pressure data and vibration characteristics, obtain tool wear distribution at P time points in the historical time zone, and generate a tool wear distribution sequence; A230: Use a generative adversarial network to predict and obtain a predicted tool wear distribution sequence for the future time zone based on the tool wear distribution sequence, wherein the predicted tool wear distribution sequence includes P predicted tool wear distributions at P time points in the future time zone.

[0032] Specifically, inside the shield machine cutterhead, sensors are used to monitor several cutting tools at fixed points and collect cutting tool status information in real time. These sensors include temperature sensors, pressure sensors, and vibration sensors, which are used to monitor the working temperature of the cutting tool (reflecting the heat generated by the cutting tool when working. Too high a temperature may mean that the cutting tool is overloaded or severely worn), the pressure on the cutting tool (monitoring the pressure generated when the cutting tool contacts the soil layer. Changes in pressure can reveal whether the cutting tool is in normal working condition), and the vibration characteristics of the cutting tool (by measuring the vibration of the cutting tool, it can be determined whether the cutting tool is damaged or abnormally worn).

[0033] Next, the tool wear state is evaluated based on the tool temperature, pressure data and vibration characteristics. For example, the temperature of the tool is usually directly related to the cutting force and the degree of tool wear. Too high a temperature may indicate that the tool surface has been worn or overloaded. By analyzing the pressure that the tool is subjected to during operation, it is evaluated whether the tool is in a normal load range. Excessive pressure may mean that the tool is in an overloaded working state, which in turn leads to increased wear. By analyzing the vibration characteristics, it is detected whether the tool is abnormally worn. Abnormal vibration usually means damage to the tool or uneven wear. The temperature, pressure and vibration data are fused, and all aspects of tool wear are comprehensively evaluated through multidimensional data analysis technology. The fused data can provide a more comprehensive basis for wear judgment. For example, a machine learning model (such as a regression analysis model) is constructed using historical data and known wear cases, and the tool wear state is predicted through multidimensional data to quantify the tool wear coefficient. The tool wear distribution at P time points in the historical time zone (including the wear coefficient of tools at several different positions in the tool disc) is obtained, and the tool wear distribution at P time points is arranged in chronological order to obtain a tool wear distribution sequence.

[0034] Further, based on the historical excavation operation log of the shield machine, we collected sample tool wear distribution sequence sets and sample predicted tool wear distribution sequence sets; and used the sample tool wear distribution sequence sets and sample predicted tool wear distribution sequence sets to iteratively train the generator and discriminator of the generative adversarial network. The task of the generator is to generate the predicted tool wear distribution sequence of the future time zone based on the input historical tool wear distribution sequence, that is, to predict the trend and change of future tool wear by learning the pattern of historical data; the task of the discriminator is to distinguish the difference between the predicted data generated by the generator and the real future tool wear distribution data. The discriminator optimizes the parameters of the model based on the comparison between the real data and the generated data; through iterative training, the generator and the discriminator compete with each other and continuously optimize. The generator learns how to generate more and more realistic predicted tool wear data, while the discriminator continuously improves its ability to distinguish generated data from real data. After multiple iterations, the model gradually converges, allowing the generator to accurately predict future tool wear conditions. During the training process, the generator generates a realistic predicted tool wear distribution as much as possible by minimizing the generation loss function, while the discriminator improves its ability to distinguish true and false data by minimizing the discriminant loss function. Finally, a trained tool wear change prediction model is obtained.

[0035] Then, the tool wear change prediction model is used to predict the tool wear distribution sequence to obtain the predicted tool wear distribution sequence in the future time zone, wherein the predicted tool wear distribution sequence includes P predicted tool wear distributions at P time points in the future time zone.

[0036] By constructing soil characteristic change prediction models and tool wear change prediction models based on generative adversarial networks, the intelligence of soil characteristic change prediction and tool wear change prediction can be improved, thereby effectively improving the accuracy and efficiency of the prediction, achieving more accurate soil change prediction and tool wear prediction, and providing reliable data support for the subsequent optimization analysis of the tunneling control parameters of the shield machine.

[0037] A300: Perform time domain alignment and cutting area mapping alignment on the predicted soil characteristic distribution sequence and the predicted tool wear distribution sequence, and integrate them to obtain a predicted operation status data sequence, wherein the cutting area mapping alignment is performed according to the interval between the tool and the center of the cutter disc.

[0038] Specifically, the predicted soil characteristic distribution sequence and the predicted tool wear distribution sequence are then aligned in the time domain and aligned in the cutting area mapping. The purpose of the time domain alignment is to align the predicted soil characteristic distribution sequence and the tool wear distribution sequence in time, to ensure that the soil characteristics and tool wear data are matched at the same time point, which is convenient for subsequent data integration and analysis; the cutting area mapping alignment refers to the alignment based on the distance between the tool and the center of the cutter disc. During the tunneling process of the shield machine, the positions of the tools are not evenly distributed. They are usually arranged at different radii (i.e., the distance between the tool and the center of the cutter disc), so their working areas and influence ranges are also different. The cutting area mapping alignment aligns the tool wear distribution with the specific position of the tool so as to accurately evaluate the impact of each tool on the tunneling process. The predicted soil characteristic distribution and the predicted tool wear distribution at the same predicted time point after alignment are further integrated as a set of predicted operation status data to obtain a predicted operation status data sequence.

[0039] A400: Based on the predicted operation status data sequence, with the minimum error of the overall excavation operation of the cutter head as the expected goal, the excavation control parameters of the shield machine are iteratively optimized and analyzed until the predetermined optimization scale is met, and the optimal excavation parameter sequence is output to control the shield machine to execute tunnel construction in the future time zone.

[0040] Further, step A400 of the present invention further includes:

[0041] A410: Select the first predicted operation status data from the predicted operation status data sequence, the first predicted operation status data including the first predicted soil characteristic distribution and the first predicted tool wear distribution at the first time point; A420: Perform soil distribution uniformity analysis based on the first predicted soil characteristic distribution, evaluate the overall tool wear state based on the first predicted tool wear distribution, and obtain the first soil distribution uniformity coefficient and the first tool wear coefficient; A430: Perform operation accuracy requirement analysis based on the first soil distribution uniformity coefficient and the first tool wear coefficient, and set the first optimization scale according to the first operation accuracy, wherein the optimization scale is the number of iterative optimization times, which is negatively correlated with the soil distribution uniformity coefficient and positively correlated with the tool wear coefficient.

[0042] Specifically, first, the first predicted operation status data is selected from the predicted operation status data sequence, and the first predicted operation status data includes the first predicted soil characteristic distribution and the first predicted tool wear distribution at the first time point (the earliest time point of P time points in the future time zone). Then, the soil distribution uniformity analysis is performed based on the first predicted soil characteristic distribution. The soil distribution uniformity analysis is to evaluate whether the soil layer encountered by the shield machine is uniform at the first time point, and whether there is a large soil change area (such as the alternation of hard soil and soft soil, etc.). The uniformity of soil distribution is evaluated by calculating the change range and change rate between different soil layers. If the soil changes dramatically, the uniformity is poor, otherwise the uniformity is good. For example, a standard is set, based on parameters such as the frequency of soil changes and the proportion of soil types, the soil distribution uniformity at that time point is calculated to obtain the first soil distribution uniformity coefficient. The more uniform the soil distribution, the larger the soil distribution uniformity coefficient, which indicates that the soil layer faced by the shield machine during the excavation process has almost no major changes, or the changes are relatively gentle. Such soil conditions usually mean that the cutter will not wear excessively, the excavation process is easier to control, the equipment workload is smaller, and the operating efficiency is higher.

[0043] According to the first predicted tool wear distribution, the overall tool wear state is evaluated, such as calculating the wear degree of each tool, and by combining these data, the wear degree of the entire tool disc is obtained. The average wear degree of all tools and the uniformity of the distribution of severely worn tools are calculated, and the overall wear degree is calculated according to the wear degree of each tool (such as percentage) and the uniformity of wear distribution to obtain a comprehensive wear coefficient. If the wear distribution is uniform and the wear is small, the wear coefficient is low; if the tool wear is severe and uneven, the wear coefficient is high; the first tool wear coefficient is obtained.

[0044] Further based on the first soil distribution uniformity coefficient and the first tool wear coefficient, an operation accuracy requirement analysis is performed. The goal of the operation accuracy requirement analysis is to determine the accuracy requirements of the shield machine in future operations based on these two parameters. For example, when the soil distribution uniformity coefficient is large, the soil changes relatively slowly and the operation accuracy requirements are low. When the tool wear coefficient is large, the tool wear is more serious and unevenly distributed, and the operation accuracy requirements are high, so as to obtain the first operation accuracy. Then, the first optimization scale is set according to the first operating accuracy. The optimization scale indicates the number of times the shield machine performs optimization and adjustment during the operation process, including the optimization frequency of tunneling parameters (such as thrust, cutter head speed, tool adjustment, etc.). The size of the optimization scale directly affects the operating accuracy and tunneling efficiency. The optimization scale is negatively correlated with the soil distribution uniformity coefficient: when the soil distribution is relatively uniform, the operating accuracy requirement is low, and the shield machine has less demand for optimization and adjustment of tunneling parameters, so the optimization scale is small. On the contrary, when the soil changes drastically and the soil is uneven, the operating accuracy requirement is increased, and the optimization scale needs to be increased to cope with the unevenness of the soil. The optimization scale is positively correlated with the tool wear coefficient: when the tool wear is more serious and unevenly distributed, the working state of the shield machine is affected, and the tool needs more adjustment and optimization to extend its service life and maintain operating accuracy. Therefore, the optimization scale will increase with the increase of the wear coefficient to ensure operating accuracy through more optimization and adjustment.

[0045] By adjusting the optimization frequency according to the uniformity of the soil and the severity of tool wear, the adaptability of the optimization frequency to the actual operation can be improved, thereby avoiding unnecessary over-optimization or under-optimization and ensuring the flexibility and real-time nature of the optimization strategy; by reducing unnecessary optimization frequency, the consumption of computing resources and optimization time can be significantly reduced, thereby ensuring the accuracy and quality of the optimization of excavation control parameters.

[0046] A440: Based on the first predicted soil characteristic distribution and the first predicted tool wear distribution, with the minimization of the overall excavation operation error of the cutterhead as the expected goal, the excavation control parameters of the shield machine are iteratively optimized and analyzed until the first optimization scale is met, and the first optimal excavation parameters are output and added to the optimal excavation parameter sequence.

[0047] Further, step A440 of the present invention further includes:

[0048] A441: Build a tunneling operation simulation space for the shield machine, and adjust the parameters of the tunneling operation simulation space according to the first predicted soil characteristic distribution and the first predicted tool wear distribution to obtain a first tunneling operation simulation space; A442: Obtain the tunneling control parameter threshold of the shield machine, and randomly generate a number of initial tunneling control parameters, wherein the tunneling control parameters include propulsion force, propulsion speed, cutter head rotation speed and cutter head torque; A443: In the first tunneling operation simulation space, simulate the tunneling operation according to the several initial tunneling control parameters, and output several cutter head overall tunneling operation errors, wherein the cutter head overall tunneling operation error is the sum of the tunneling depth dimension deviations of the several tool coverage areas; A444: Mark the several initial tunneling control parameters with operation fitness according to the several cutter head overall tunneling operation errors to obtain several operation fitnesses.

[0049] Specifically, first, a tunneling operation simulation space for the shield machine is built. The tunneling operation simulation space is a virtual environment that can be simulated and modeled through digital twins to simulate the tunneling process of the shield machine in actual construction. The purpose of the simulation space is to provide an efficient and accurate virtual test platform for the shield machine to verify and optimize the operation parameters before actual construction; then, according to the first predicted soil characteristic distribution and the first predicted tool wear distribution, the parameters of the tunneling operation simulation space are adjusted, that is, in the tunneling simulation space, the soil characteristics (such as soil type, hardness, moisture content, etc.) predicted at the first time point are mapped to the virtual tunneling space to simulate the specific changes in the soil layer; the tool wear prediction data (such as tool wear degree, wear distribution position) at the first time point is added to the simulation space to simulate the working state and wear of the tool under different soil conditions; through the input of these data, the soil characteristics and tool wear distribution in the simulation space are adjusted to obtain a first tunneling operation simulation space that is more in line with actual working conditions, which can accurately reflect the working environment and tool status of the shield machine in the future time zone.

[0050] Next, the tunneling control parameter threshold of the shield machine is obtained. The tunneling control parameter threshold refers to the maximum or minimum operating parameter limit that the shield machine can withstand under different soil conditions. It can be obtained through historical data analysis and equipment technical specifications. The tunneling control parameters include thrust, thrust speed, cutter head speed and cutter head torque. Further, a number of initial tunneling control parameters are randomly generated within the tunneling control parameter threshold.

[0051] Then, in the first tunneling operation simulation space, several initial tunneling control parameters that have been obtained are input, and a simulated tunneling operation is performed. The purpose of the simulated tunneling operation is to simulate the performance of the shield machine under different operating conditions through a virtual environment, and then analyze the operation error and performance. The overall tunneling operation error of the cutterhead refers to the error caused by factors such as soil changes, tool wear, and improper parameter settings during the simulated tunneling process. These errors will affect the tunneling accuracy of the shield machine, especially in terms of tunneling depth; each tool is responsible for cutting a certain layer of soil in the working area of ​​the cutterhead, and its accuracy is affected by many factors, such as tool wear, soil type, etc. Etc., each tool's working area will have a certain depth deviation, which can be calculated through simulation results; then, for each tool's coverage area, by comparing the difference between the actual depth of simulated excavation and the ideal depth (or design trajectory), the excavation depth deviation of each tool is calculated. The source of the deviation may be factors such as uneven soil quality, uneven tool wear, and too high or too low cutter head speed; then the excavation depth deviations of all tools are summarized to obtain the total cutter head overall excavation operation error. This error value is one of the key indicators in the excavation operation process, which is used to evaluate the operating accuracy of the shield machine and output a number of cutter head overall excavation operation errors.

[0052] Furthermore, the several initial excavation control parameters are marked with operational adaptability according to the several overall excavation operation errors of the several cutter heads, that is, the adaptability of each set of initial excavation control parameters is evaluated according to the excavation operation error of the shield machine. When the overall excavation operation error of the cutter head is small, it means that the operation accuracy is high, the excavation control parameters are well matched with the operating conditions such as soil quality and tool status, and the operational adaptability is high; conversely, when the overall excavation operation error of the cutter head is large, it means that there are large errors in the excavation process, which may cause the tunnel excavation size to not meet the design requirements, and the operational adaptability is low; and several operational adaptabilities are obtained.

[0053] A445: Based on the several operation fitnesses and the several initial excavation control parameters, perform iterative optimization analysis of the excavation control parameters until the first optimization scale is met, and output the first optimal excavation parameters.

[0054] Further, step A445 of the present invention also includes:

[0055] A4451: Set the initial excavation control parameters as the initial solution, arrange several initial solutions from large to small according to the operation fitness, and generate an initial solution sequence; A4452: Set the first K solutions of the initial solution sequence as optimal solutions, and the last M solutions as inferior solutions, and based on the K optimal solutions, randomly divide and cluster the M inferior solutions to obtain K solution sets, where K and M are both positive integers, and M is an integer multiple of K; A4453: In the K solution sets, take the optimal solution as the adjustment direction, adjust the inferior solutions in the solution set once according to the predetermined parameter adjustment range, and obtain K updated solution sets, Among them, if the adjusted inferior solution does not meet the threshold of the tunneling control parameter, an initial solution is randomly generated to replace the inferior solution; A4454: identify the K updated solution sets, if in the same updated solution set, the fitness of the inferior solution is greater than the fitness of the superior solution, the superior solution is replaced by the inferior solution; A4455: perform iterative optimization until the first optimization scale is reached, output K current solution sets, and sum the fitness of the K current solution sets respectively, set the current solution set with the largest sum of fitness as the current optimal solution set, and output the superior solution of the current optimal solution set as the first optimal tunneling parameter.

[0056] Specifically, first, the initial excavation control parameters are set as the initial solution. Based on the several operational fitnesses, several initial solutions are arranged from large to small according to the operational fitnesses. The operational fitnesses reflect the adaptability of the control parameter combination in the actual excavation operation. The higher the fitness of the solution, the better the operation performance under this condition. An initial solution sequence is generated, in which the solutions with higher operational fitness are arranged in the front and the solutions with lower fitness are arranged in the back. This sequence is used for the subsequent optimization process. Next, the first K solutions (solutions with higher operational fitnesses) in the initial solution sequence are set as optimal solutions, and the last M solutions (solutions with lower fitnesses) are set as inferior solutions. Here, K and M are positive integers and satisfy that M is an integer multiple of K, that is, the number of inferior solutions is an integer multiple of the number of optimal solutions; then, based on the K optimal solutions, the M inferior solutions are randomly divided and clustered. The purpose of this process is to generate a set of inferior solutions for each optimal solution. These inferior solutions will be adjusted according to the optimal solution in the subsequent optimization process. After division and clustering, K solution sets will be obtained. Each solution set contains one optimal solution and several inferior solutions. These solution sets will serve as the basis for subsequent optimization.

[0057] Then, in the K solution sets, the inferior solution in each solution set is adjusted with the optimal solution as the adjustment direction. The adjustment method can be to optimize the inferior solution through a certain predetermined parameter adjustment range (including the thrust force adjustment step, the thrust speed adjustment step, the cutter head speed adjustment step and the cutter head torque adjustment step); after the adjustment, it is necessary to check whether each inferior solution meets the tunneling control parameter threshold. If the adjusted inferior solution does not meet the threshold limit, an initial solution is randomly generated to replace the inferior solution to ensure that each solution is within the valid parameter range; through the adjustment, a new K updated solution set is generated. Each updated solution set contains the optimal solution and the adjusted inferior solution. Then the K updated solution sets are identified. If the fitness of the inferior solution is greater than the fitness of the optimal solution in the same solution set, the superior solution is replaced by the inferior solution, that is, if the fitness of the inferior solution in the same solution set is higher than that of the optimal solution, it means that the current inferior solution performs better under this condition and the original optimal solution should be replaced. This helps to jump out of the local optimal solution and find more suitable control parameters during the optimization process.

[0058] Multiple iterative optimizations are performed, and each iteration will perform further optimization based on the updated solution set. The optimization process will continue until the set first optimization scale (such as 500 times) is reached, that is, the predetermined number of optimizations is reached, and K current solution sets are output; then the fitness of the K current solution sets is summed up respectively, and the larger the sum of the fitness, the better the overall performance of the solution set under the current conditions. The current solution set with the largest sum of fitness is set as the current optimal solution set, which represents the current best combination of control parameters; further, the optimal solution is extracted from the current optimal solution set and set as the first optimal excavation parameter.

[0059] Further, step A400 of the present invention further includes:

[0060] A450: In the process of controlling the shield machine to perform tunnel construction in the future time zone according to the optimal tunneling parameter sequence, real-time soil characteristic distribution and real-time tool wear distribution are collected, and prediction error analysis is performed based on the predicted soil characteristic distribution and predicted tool wear distribution at the same time point to determine the soil characteristic prediction error and tool wear prediction error; A460: An error compensation database is constructed using sample data, and tunneling parameter mapping compensation coefficients are obtained based on the matching of the soil characteristic prediction error and the tool wear prediction error; A470: Several tunneling parameter mapping compensation coefficients of several time nodes before the current time node are obtained, and the several tunneling parameter mapping compensation coefficients are comprehensively calculated according to predetermined gradient weights to obtain parameter compensation coefficients, wherein the weights are negatively correlated with the time intervals; A480: The optimal tunneling parameters at the current time node are optimized and adjusted using the parameter compensation coefficients to obtain the adjusted optimal tunneling parameters to control the shield machine to perform tunnel construction at the current time node.

[0061] Specifically, in the process of controlling the shield machine to execute tunnel construction in the future time zone according to the optimal tunneling parameter sequence, sensors and monitoring equipment are used to collect and obtain real-time soil characteristic distribution and real-time tool wear distribution; then, the real-time soil characteristic distribution is compared with the predicted soil characteristic distribution, and the soil characteristic prediction error is calculated. If the difference between the real-time data and the predicted data is large, it means that the soil prediction model fails to accurately reflect the actual situation, and the prediction algorithm or tunneling strategy may need to be adjusted; the real-time tool wear distribution is compared with the predicted tool wear distribution, and the tool wear prediction error is calculated. If there is a large deviation between the actual wear and the predicted value, it means that the speed of tool wear in the actual tunneling process is different from the prediction, and the tool management strategy may need to be optimized or the tunneling parameters may need to be adjusted.

[0062] Then, an error compensation database is constructed, that is, sample data is collected, and the historical soil characteristics prediction error and tool wear prediction error are combined with the corresponding tunneling control parameters to form an error compensation database. The database records the tunneling parameter adjustments required by the shield machine under different soil characteristics and tool wear conditions. The error can be mapped to the corresponding adjustment parameters through machine learning models or regression analysis to provide an accurate compensation model. Further, according to the real-time collected soil characteristics prediction error and tool wear prediction error, the corresponding tunneling parameter adjustment compensation coefficient is matched from the error compensation database. This process is based on the actual differences in soil and tool wear, and the most similar error conditions and their corresponding tunneling parameter adjustment strategies in the historical data are found. The mapping compensation coefficient is to associate the soil characteristics prediction error and tool wear prediction error with the tunneling control parameters to obtain a coefficient that can adjust the tunneling parameters. For example, if the soil characteristics prediction error is large or the tool wear prediction error is large, the system will adjust the tunneling control parameters accordingly according to the compensation coefficient in the historical data. In this way, the impact caused by the error can be reduced in actual operations to ensure the accuracy and efficiency of tunneling.

[0063] Then, at the current time node, by looking back at historical data, the excavation parameter mapping compensation coefficients of the previous several time nodes (for example, the previous 5 time nodes) are obtained. These coefficients are obtained by analyzing the historical soil characteristics prediction error and tool wear prediction error. They reflect the adjustment requirements of the excavation parameters at different time points. The excavation parameter mapping compensation coefficient of each time node is calculated based on the soil characteristics and tool wear error at that time, indicating how the excavation control parameters should be adjusted at this time node. Further obtain the predetermined gradual weight, the gradual weight is negatively correlated with the time interval, which means that the historical compensation coefficient closer to the current time node (that is, the smaller the time interval) will be given a higher weight, and the compensation coefficient of the farther time node will be given a lower weight according to the time interval; and the several excavation parameter mapping compensation coefficients are weighted according to the predetermined gradual weight to obtain a comprehensive parameter compensation coefficient, that is, the parameter compensation coefficient.

[0064] Finally, the parameter compensation coefficient is used to optimize and adjust the optimal excavation parameters at the current time node, that is, according to the parameter compensation coefficient, the control parameters such as propulsion force, propulsion speed, and cutter head speed are fine-tuned. For example, if the soil quality changes greatly or the tool wear is serious, the compensation coefficient may prompt to increase the propulsion force or adjust the cutter head speed to maintain the excavation accuracy, and obtain the adjusted optimal excavation parameters. These parameters can better adapt to the current soil quality and tool wear conditions, thereby improving the operation accuracy and efficiency; finally, according to the adjustment of the optimal excavation parameters, the shield machine is controlled to perform tunnel construction at the current time node. By obtaining the excavation parameter mapping compensation coefficients of the previous several time nodes and performing weighted calculations according to the predetermined gradient weights, a comprehensive parameter compensation coefficient is obtained, which is used to optimize and adjust the current optimal excavation parameters to ensure that the shield machine can adapt to the changes in soil quality and tool wear in the tunnel construction at the current time node, and improve the operation accuracy and efficiency.

[0065] The shield machine excavation error compensation control method provided by the embodiment of the present invention has at least the following technical effects:

[0066] 1. By introducing a dynamic adjustment mechanism based on real-time monitoring and prediction, the tunneling control parameters can be quickly and accurately optimized according to future soil changes and tool wear conditions, and the automation and intelligence of the shield machine tunneling control can be improved, thereby effectively reducing the tunneling operation error, significantly improving the accuracy and operating efficiency of the shield machine during tunneling, and realizing efficient and accurate tunnel construction.

[0067] 2. By constructing a soil characteristic change prediction model and a tool wear change prediction model based on a generative adversarial network, the intelligence of the soil characteristic change prediction and the tool wear change prediction can be improved, thereby effectively improving the accuracy and efficiency of the prediction, achieving more accurate soil change prediction and tool wear prediction, and providing reliable data support for the subsequent optimization analysis of the tunneling control parameters of the shield machine.

[0068] 3. By adjusting the optimization frequency according to the uniformity of the soil and the severity of tool wear, the adaptability of the optimization frequency to the actual operating state can be improved, thereby avoiding unnecessary over-optimization or under-optimization, and ensuring the flexibility and real-time nature of the optimization strategy; by reducing unnecessary optimization frequency, the consumption of computing resources and optimization time can be significantly reduced, thereby ensuring the accuracy and quality of the optimization of tunneling control parameters.

[0069] 4. By obtaining the excavation parameter mapping compensation coefficients of the previous several time nodes and performing weighted calculation according to the predetermined gradient weights, a comprehensive parameter compensation coefficient is obtained. This coefficient is used to optimize and adjust the current optimal excavation parameters to ensure that the shield machine can adapt to changes in soil quality and tool wear during tunnel construction at the current time node, thereby improving operation accuracy and efficiency.

[0070] Embodiment 2, as Figure 2 As shown, based on the same inventive concept as the shield machine excavation error compensation control method provided in the first embodiment, the embodiment of the present invention further provides a shield machine excavation error compensation control system, comprising:

[0071] The predicted soil characteristic distribution sequence acquisition unit 01 is used to collect and record the soil characteristic data in the tunnel through the sonic survey instrument at a fixed point during the shield machine excavation construction process, obtain the soil characteristic distribution sequence of the historical time zone, and predict the predicted soil characteristic distribution sequence of the future time zone based on the soil characteristic distribution sequence; the predicted tool wear distribution sequence acquisition unit 02 is used to collect the tool status information of several tools in the shield machine cutter head through the sensor fixed point monitoring, obtain the tool wear distribution sequence of the historical time zone, and predict the predicted tool wear distribution sequence of the future time zone based on the tool wear distribution sequence; predict the operation The state data sequence obtaining unit 03 is used to perform time domain alignment and cutting area mapping alignment on the predicted soil characteristic distribution sequence and the predicted tool wear distribution sequence, and integrate them to obtain a predicted operation state data sequence, wherein the cutting area mapping alignment is performed according to the interval between the tool and the center of the cutter disc; the tunneling control parameter iterative optimization unit 04 is used to iteratively optimize and analyze the tunneling control parameters of the shield machine based on the predicted operation state data sequence, with the minimum error of the overall tunneling operation of the cutter disc as the expected goal, until the predetermined optimization scale is met, and the optimal tunneling parameter sequence is output to control the shield machine to execute tunnel construction in the future time zone.

[0072] Furthermore, the shield machine excavation error compensation control system is also used for: in the process of controlling the shield machine to perform tunnel construction in the future time zone according to the optimal excavation parameter sequence, collecting and acquiring real-time soil characteristic distribution and real-time tool wear distribution, and performing prediction error analysis based on the predicted soil characteristic distribution and predicted tool wear distribution at the same time point to determine the soil characteristic prediction error and the tool wear prediction error; constructing an error compensation database using sample data, and obtaining excavation parameter mapping compensation coefficients based on the matching of the soil characteristic prediction error and the tool wear prediction error; obtaining several excavation parameter mapping compensation coefficients of several time nodes before the current time node, and comprehensively calculating the several excavation parameter mapping compensation coefficients according to predetermined gradient weights to obtain parameter compensation coefficients, wherein the weights are negatively correlated with the time intervals; and optimizing and adjusting the optimal excavation parameters at the current time node using the parameter compensation coefficients to obtain the adjusted optimal excavation parameters to control the shield machine to perform tunnel construction at the current time node.

[0073] Furthermore, the shield machine excavation error compensation control system is also used to: collect and record soil characteristic data in the tunnel at a fixed point through an acoustic survey instrument, obtain P soil characteristic distributions at P time points in a historical time zone, and generate a soil characteristic distribution sequence, wherein the soil characteristics include soil type, soil hardness, soil structure and moisture content, and P is an integer greater than 10; use a generative adversarial network to construct a soil characteristic change prediction model, and input the soil characteristic distribution sequence into the soil characteristic change prediction model, and output a predicted soil characteristic distribution sequence for a future time zone, wherein the predicted soil characteristic distribution sequence includes P predicted soil characteristic distributions at P time points in the future time zone.

[0074] Furthermore, the shield machine excavation error compensation control system is also used to: based on the historical excavation operation log of the shield machine, with P time points as screening constraints, collect a set of sample soil characteristic distribution sequences, and collect the soil characteristic distribution of each sample soil characteristic distribution sequence at P time points in the future to obtain a set of sample predicted soil characteristic distribution sequences; use the sample soil characteristic distribution sequence set and the sample predicted soil characteristic distribution sequence set to iteratively train the generator and discriminator of the generative adversarial network until both the generation loss function and the discriminant loss function converge, thereby obtaining the soil characteristic change prediction model.

[0075] Furthermore, the shield machine excavation error compensation control system is also used to: collect tool status information of several tools in the shield machine cutter head through fixed-point monitoring by sensors, wherein the tool status information includes tool temperature, pressure data and vibration characteristics; evaluate tool wear status according to the tool temperature, pressure data and vibration characteristics, obtain tool wear distribution at P time points in the historical time zone, and generate a tool wear distribution sequence; use a generative adversarial network to predict and obtain a predicted tool wear distribution sequence for the future time zone according to the tool wear distribution sequence, wherein the predicted tool wear distribution sequence includes P predicted tool wear distributions at P time points in the future time zone.

[0076] Furthermore, the shield machine excavation error compensation control system is also used to: select first predicted operating state data from the predicted operating state data sequence, the first predicted operating state data including a first predicted soil characteristic distribution and a first predicted tool wear distribution at a first time point; perform soil distribution uniformity analysis based on the first predicted soil characteristic distribution, perform overall tool wear state evaluation based on the first predicted tool wear distribution, and obtain a first soil distribution uniformity coefficient and a first tool wear coefficient; perform operating accuracy requirement analysis based on the first soil distribution uniformity coefficient and the first tool wear coefficient, and set a first optimization scale based on the first operating accuracy, wherein the optimization scale is the number of iterative optimization times, is negatively correlated with the soil distribution uniformity coefficient, and is positively correlated with the tool wear coefficient; based on the first predicted soil characteristic distribution and the first predicted tool wear distribution, with the overall excavation operation error of the cutter disc minimized as the expected goal, perform iterative optimization analysis on the excavation control parameters of the shield machine until the first optimization scale is met, and output the first optimal excavation parameters to add to the optimal excavation parameter sequence.

[0077] Furthermore, the tunneling error compensation control system of a shield machine is also used to: build a tunneling operation simulation space of the shield machine, and adjust the parameters of the tunneling operation simulation space according to the first predicted soil characteristic distribution and the first predicted tool wear distribution to obtain a first tunneling operation simulation space; obtain the tunneling control parameter threshold of the shield machine, and randomly generate a number of initial tunneling control parameters, wherein the tunneling control parameters include propulsion force, propulsion speed, cutter head speed and cutter head torque; in the first tunneling operation simulation space, simulate the tunneling operation according to the several initial tunneling control parameters, and output a number of cutter head overall tunneling operation errors, wherein the cutter head overall tunneling operation error is the sum of the tunneling depth dimension deviations of the several tool coverage areas; mark the several initial tunneling control parameters with operation fitness according to the several cutter head overall tunneling operation errors, and obtain a number of operation fitnesses; based on the several operation fitnesses and several initial tunneling control parameters, perform iterative optimization analysis of the tunneling control parameters until the first optimization scale is met, and output the first optimal tunneling parameters.

[0078] Furthermore, the shield machine excavation error compensation control system is also used to: set the initial excavation control parameter as the initial solution, arrange a number of initial solutions from large to small according to the operation fitness, and generate an initial solution sequence; set the first K solutions of the initial solution sequence as optimal solutions, and the last M solutions as inferior solutions, and based on the K optimal solutions, randomly divide and cluster the M inferior solutions to obtain K solution sets, wherein K and M are both positive integers, and M is an integer multiple of K; within the K solution sets, take the optimal solution as the adjustment direction, and adjust the inferior solutions in the solution set once according to the predetermined parameter adjustment range , obtaining K updated solution sets, wherein if the adjusted inferior solution does not meet the tunneling control parameter threshold, randomly generating an initial solution to replace the inferior solution; identifying the K updated solution sets, if within the same updated solution set, the fitness of the inferior solution is greater than the fitness of the superior solution, replacing the superior solution with the inferior solution; performing iterative optimization until the first optimization scale is reached, outputting K current solution sets, and summing up the fitness of the K current solution sets respectively, setting the current solution set with the largest sum of fitness as the current optimal solution set, and outputting the superior solution of the current optimal solution set as the first optimal tunneling parameter.

[0079] Although preferred embodiments of the present invention have been described, additional changes and modifications may occur to these embodiments once those skilled in the art understand the basic inventive concepts.

[0080] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention belong to the scope of the present invention and its equivalent technologies, the present invention is also intended to include these changes and variations.

Claims

1. A shield machine excavation error compensation control method, characterized in that the method include: During the tunneling construction process of the shield machine, the soil characteristic data in the tunnel is collected and recorded at fixed points by an acoustic survey instrument to obtain the soil characteristic distribution sequence of the historical time zone, and the predicted soil characteristic distribution sequence of the future time zone is predicted based on the soil characteristic distribution sequence; The tool status information of several tools in the shield machine cutter head is collected by fixed-point monitoring with sensors, the tool wear distribution sequence of the historical time zone is obtained, and the predicted tool wear distribution sequence of the future time zone is obtained based on the tool wear distribution sequence prediction; Performing time domain alignment and cutting area mapping alignment on the predicted soil characteristic distribution sequence and the predicted tool wear distribution sequence, and integrating them to obtain a predicted operation status data sequence, wherein the cutting area mapping alignment is performed according to the interval between the tool and the center of the cutter disc; Based on the predicted operation status data sequence, with the minimum error of the overall excavation operation of the cutterhead as the expected goal, the excavation control parameters of the shield machine are iteratively optimized and analyzed until the predetermined optimization scale is met, the optimal excavation parameter sequence is output, and the shield machine is controlled to execute tunnel construction in the future time zone; The controlling of the shield machine to perform tunnel construction in the future time zone includes: In the process of controlling the shield machine to execute tunnel construction in the future time zone according to the optimal tunneling parameter sequence, real-time soil characteristic distribution and real-time tool wear distribution are collected and obtained, and prediction error analysis is performed based on the predicted soil characteristic distribution and predicted tool wear distribution at the same time point to determine the soil characteristic prediction error and tool wear prediction error; Using the sample data to construct an error compensation database, and obtaining a tunneling parameter mapping compensation coefficient based on the matching of the soil characteristic prediction error and the tool wear prediction error; Acquire several tunneling parameter mapping compensation coefficients of several time nodes before the current time node, and perform comprehensive calculation on the several tunneling parameter mapping compensation coefficients according to a predetermined gradual weight to obtain a parameter compensation coefficient, wherein the weight is negatively correlated with the time interval; The parameter compensation coefficient is used to optimize and adjust the optimal excavation parameters at the current time node, and the optimal excavation parameters are adjusted to control the shield machine to perform tunnel construction at the current time node.

2. A shield machine excavation error compensation control method according to claim 1, characterized in that: The soil characteristic data in the tunnel are collected and recorded at fixed points by an acoustic survey instrument to obtain a soil characteristic distribution sequence of a historical time zone, and a predicted soil characteristic distribution sequence of a future time zone is predicted based on the soil characteristic distribution sequence, including: The soil characteristic data in the tunnel are collected and recorded at fixed points by using an acoustic survey instrument, and P soil characteristic distributions at P time points in the historical time zone are obtained to generate a soil characteristic distribution sequence, wherein the soil characteristics include soil type, soil hardness, soil structure and water content, and P is an integer greater than 10; A soil characteristic change prediction model is constructed using a generative adversarial network, and the soil characteristic distribution sequence is input into the soil characteristic change prediction model to output a predicted soil characteristic distribution sequence for a future time zone, wherein the predicted soil characteristic distribution sequence includes P predicted soil characteristic distributions at P time points in the future time zone.

3. A shield machine excavation error compensation control method according to claim 2, characterized in that: The soil characteristics change prediction model is constructed using the generative adversarial network, including: According to the historical excavation operation log of the shield machine, with P time points as the screening constraints, a sample soil characteristic distribution sequence set is collected, and the soil characteristic distribution of each sample soil characteristic distribution sequence at the next P time points is collected to obtain a sample predicted soil characteristic distribution sequence set; The generator and discriminator of the generative adversarial network are iteratively trained using the sample soil characteristic distribution sequence set and the sample predicted soil characteristic distribution sequence set until both the generation loss function and the discriminant loss function converge, thereby obtaining the soil characteristic change prediction model.

4. A shield machine excavation error compensation control method according to claim 3, characterized in that: The tool status information of several tools in the shield machine cutter head is collected by fixed-point monitoring with sensors, the tool wear distribution sequence of the historical time zone is obtained, and the predicted tool wear distribution sequence of the future time zone is obtained according to the tool wear distribution sequence prediction, including: The tool status information of several tools in the shield machine cutter head is collected through fixed-point monitoring by sensors, wherein the tool status information includes tool temperature, pressure data and vibration characteristics; Evaluate the tool wear state according to the tool temperature, pressure data and vibration characteristics, obtain the tool wear distribution at P time points in the historical time zone, and generate a tool wear distribution sequence; A generative adversarial network is used to predict the tool wear distribution sequence for a future time zone, wherein the predicted tool wear distribution sequence includes P predicted tool wear distributions at P time points in the future time zone.

5. A shield machine excavation error compensation control method according to claim 3, characterized in that: Based on the predicted operation status data sequence, with the minimum error of the overall excavation operation of the cutterhead as the expected goal, the excavation control parameters of the shield machine are iteratively optimized and analyzed until the predetermined optimization scale is met, and the optimal excavation parameter sequence is output, including: Selecting first predicted operation status data from the predicted operation status data sequence, wherein the first predicted operation status data includes a first predicted soil characteristic distribution and a first predicted tool wear distribution at a first time point; Perform soil distribution uniformity analysis based on the first predicted soil characteristic distribution, evaluate the overall tool wear state based on the first predicted tool wear distribution, and obtain a first soil distribution uniformity coefficient and a first tool wear coefficient; Based on the first soil distribution uniformity coefficient and the first tool wear coefficient, an operation accuracy requirement analysis is performed, and a first optimization scale is set according to the first operation accuracy, wherein the optimization scale is the number of iterative optimizations, is negatively correlated with the soil distribution uniformity coefficient, and is positively correlated with the tool wear coefficient; Based on the first predicted soil characteristic distribution and the first predicted tool wear distribution, with the minimization of the overall excavation operation error of the cutter head as the expected goal, the excavation control parameters of the shield machine are iteratively optimized and analyzed until the first optimization scale is met, and the first optimal excavation parameters are output and added to the optimal excavation parameter sequence.

6. A shield machine excavation error compensation control method according to claim 5, characterized in that: Based on the first predicted soil characteristic distribution and the first predicted tool wear distribution, with the minimum error of the overall excavation operation of the cutter head as the expected goal, the excavation control parameters of the shield machine are iteratively optimized and analyzed until the first optimization scale is met, and the first optimal excavation parameters are output, including: Building a tunneling operation simulation space for a shield machine, and adjusting parameters of the tunneling operation simulation space according to the first predicted soil characteristic distribution and the first predicted tool wear distribution to obtain a first tunneling operation simulation space; Obtaining the tunneling control parameter threshold of the shield machine and randomly generating a number of initial tunneling control parameters, wherein the tunneling control parameters include propulsion force, propulsion speed, cutter head speed and cutter head torque; In the first tunneling operation simulation space, a tunneling operation is simulated according to the plurality of initial tunneling control parameters, and a plurality of cutter head overall tunneling operation errors are output, wherein the cutter head overall tunneling operation error is the sum of the tunneling depth dimension deviations of the plurality of cutter cover areas; Marking the several initial excavation control parameters with operation fitness according to the several cutterhead overall excavation operation errors to obtain several operation fitnesses; Based on the several operation fitnesses and the several initial excavation control parameters, an iterative optimization analysis of the excavation control parameters is performed until the first optimization scale is met, and the first optimal excavation parameters are output.

7. A shield machine excavation error compensation control method according to claim 6, characterized in that: Based on the plurality of operation fitnesses and the plurality of initial excavation control parameters, performing iterative optimization analysis of the excavation control parameters until the first optimization scale is met, and outputting the first optimal excavation parameters, including: The initial excavation control parameters are set as the initial solution, and several initial solutions are arranged from large to small according to the operation fitness to generate an initial solution sequence; The first K solutions of the initial solution sequence are set as optimal solutions, and the last M solutions are set as inferior solutions, and based on the K optimal solutions, the M inferior solutions are randomly divided and clustered to obtain K solution sets, where K and M are both positive integers, and M is an integer multiple of K; In the K solution sets, taking the best solution as the adjustment direction, adjusting the inferior solution in the solution set once according to the predetermined parameter adjustment range to obtain K updated solution sets, wherein if the adjusted inferior solution does not meet the tunneling control parameter threshold, randomly generating an initial solution to replace the inferior solution; Identify the K update solution sets, and if the fitness of the inferior solution is greater than the fitness of the superior solution in the same update solution set, replace the superior solution with the inferior solution; Perform iterative optimization until the first optimization scale is reached, output K current solution sets, and sum the fitness of the K current solution sets respectively, set the current solution set with the largest sum of fitness as the current optimal solution set, and output the optimal solution of the current optimal solution set as the first optimal excavation parameter.

8. A shield machine excavation error compensation control system, characterized in that: The steps for implementing a shield machine excavation error compensation control method as described in any one of claims 1 to 7 include: The predicted soil characteristic distribution sequence acquisition unit is used to collect and record the soil characteristic data in the tunnel at fixed points through an acoustic survey instrument during the shield machine excavation construction process, obtain the soil characteristic distribution sequence of the historical time zone, and predict and obtain the predicted soil characteristic distribution sequence of the future time zone based on the soil characteristic distribution sequence; A predicted tool wear distribution sequence acquisition unit is used to collect tool status information of a plurality of tools in the shield machine cutter head through fixed-point monitoring by sensors, acquire tool wear distribution sequences in historical time zones, and predict and acquire predicted tool wear distribution sequences in future time zones based on the tool wear distribution sequences; A predicted operation status data sequence obtaining unit is used to perform time domain alignment and cutting area mapping alignment on the predicted soil characteristic distribution sequence and the predicted tool wear distribution sequence, and integrate them to obtain a predicted operation status data sequence, wherein the cutting area mapping alignment is performed according to the interval between the tool and the center of the cutter disc; The tunneling control parameter iterative optimization unit is used to iteratively optimize and analyze the tunneling control parameters of the shield machine based on the predicted operation state data sequence and with the minimum error of the overall tunneling operation of the cutterhead as the expected goal, until the predetermined optimization scale is met, output the optimal tunneling parameter sequence, and control the shield machine to execute tunnel construction in the future time zone; Wherein, the tunneling control parameter iterative optimization unit is also used for: In the process of controlling the shield machine to execute tunnel construction in the future time zone according to the optimal tunneling parameter sequence, real-time soil characteristic distribution and real-time tool wear distribution are collected and obtained, and prediction error analysis is performed based on the predicted soil characteristic distribution and predicted tool wear distribution at the same time point to determine the soil characteristic prediction error and tool wear prediction error; Using the sample data to construct an error compensation database, and obtaining a tunneling parameter mapping compensation coefficient based on the matching of the soil characteristic prediction error and the tool wear prediction error; Acquire several tunneling parameter mapping compensation coefficients of several time nodes before the current time node, and perform comprehensive calculation on the several tunneling parameter mapping compensation coefficients according to a predetermined gradual weight to obtain a parameter compensation coefficient, wherein the weight is negatively correlated with the time interval; The parameter compensation coefficient is used to optimize and adjust the optimal excavation parameters at the current time node, and the optimal excavation parameters are adjusted to control the shield machine to perform tunnel construction at the current time node.

Images