WO2019063024A1 - Smart decision making method and system for boring control parameters of hard rock tbm - Google Patents

Smart decision making method and system for boring control parameters of hard rock tbm Download PDF

Info

Publication number
WO2019063024A1
WO2019063024A1 PCT/CN2018/112521 CN2018112521W WO2019063024A1 WO 2019063024 A1 WO2019063024 A1 WO 2019063024A1 CN 2018112521 W CN2018112521 W CN 2018112521W WO 2019063024 A1 WO2019063024 A1 WO 2019063024A1
Authority
WO
WIPO (PCT)
Prior art keywords
parameters
rock
parameter
model
excavation
Prior art date
Application number
PCT/CN2018/112521
Other languages
French (fr)
Chinese (zh)
Inventor
李建斌
荆留杰
李鹏宇
徐受天
杨晨
焦妮
Original Assignee
中铁工程装备集团有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中铁工程装备集团有限公司 filed Critical 中铁工程装备集团有限公司
Priority to DE112018004244.8T priority Critical patent/DE112018004244T5/en
Publication of WO2019063024A1 publication Critical patent/WO2019063024A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
    • G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
    • G05B13/0265—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric the criterion being a learning criterion
    • E—FIXED CONSTRUCTIONS
    • E21—EARTH OR ROCK DRILLING; MINING
    • E21D—SHAFTS; TUNNELS; GALLERIES; LARGE UNDERGROUND CHAMBERS
    • E21D9/00—Tunnels or galleries, with or without linings; Methods or apparatus for making thereof; Layout of tunnels or galleries
    • E21D9/003—Arrangement of measuring or indicating devices for use during driving of tunnels, e.g. for guiding machines

Definitions

  • the invention relates to the technical field of tunnelling equipment construction intelligent control, in particular to an intelligent decision method and system for hard rock TBM tunneling control parameters.
  • Hard rock tunnel boring machine (hereinafter referred to as hard rock TBM) is a large-scale high-tech construction equipment specially used for excavation of rock tunnels and underground passages.
  • the driver in the main control room evaluates the state parameters of the surrounding rock by slowly testing the progress process, and then repeatedly adjusts the excavation parameters until the excavation parameters remain stable.
  • this kind of operation mode will lead to a large amount of construction time consumption due to the continuous slow test and repeated adjustment of the excavation parameters.
  • the surrounding rock changes drastically the parameters of the current surrounding rock cannot be effectively and effectively detected in real time, resulting in excavation.
  • the parameters can not adapt to the current tunneling environment, causing abnormal wear of the tool and the rock breaking performance of the cutter head system.
  • the present invention proposes an intelligent decision-making method and system for control parameters of hard rock TBM tunneling, which uses data mining and machine learning techniques to predict the driving parameters, and can adapt to different circumferences through self-learning and self-updating.
  • an intelligent decision method for hard rock TBM tunneling control parameters the steps are as follows:
  • Step 1 Establish an engineering database including a surrounding rock state database and a tunneling parameter database
  • Step 2 According to the surrounding rock state parameters, obtain a continuous surrounding rock grade W; use the three-layer neural network, support vector machine and least squares regression method to carry out data mining and machine for surrounding rock state parameters and tunneling control parameters in the engineering database. Learning, obtaining mathematical model of rock machine and intelligent control decision model through mathematical calculation;
  • Step 3 The rock-machine mutual-feeding model predicts the surrounding rock state parameters of the current excavation environment according to the real-time acquisition parameters of the TBM host computer.
  • the intelligent control decision model pre-predicts the optimal tunneling control parameters according to the obtained surrounding rock state parameters. Estimate, display the cumulative average estimated parameters and current estimated parameters on the TBM;
  • Step 4 Automatically or manually adjust the current tunneling parameters of the TBM by using the cumulative average estimated parameter and the current estimated parameter;
  • Step 5 Transfer the TBM real-time excavation parameters and other TBM engineering databases to the engineering database, and proceed to step 2 to self-learn and self-update the rock machine mutual feedback model and the intelligent control decision model.
  • the data contained in the surrounding rock state database includes: single-knife thrust Ft, single-pole torque Tn, penetration P, cutter speed n, and propulsion speed V;
  • the data included in the excavation parameter database includes: rock uniaxial saturation resistance The compressive strength Rc, the joint number Jv of the unit rock volume, and the surrounding rock grade W.
  • the method for obtaining a continuous surrounding rock grade W according to the surrounding rock state parameter is:
  • Step 2 According to the engineering rock mass grading standard, the uniaxial saturated compressive strength Rc and the fitting value Kv' are used to obtain the value of the basic quality index BQ of the rock mass;
  • the method for obtaining the rock machine mutual feed model is:
  • Step 2 Establish a three-layer neural network, take the ascending segment excavation parameter matrix M1 as input, the surrounding rock state parameter matrix N as the output, and supervise the learning of the initial neural network; select 70% of the data in the excavation parameter database for training. 30% of the data is tested, and a mature neural network Net1 is obtained, and the prediction result of surrounding rock state parameters is obtained Ynet1;
  • Step 3 Using the support vector machine to take the ascending segment excavation parameter matrix M1 as input, the surrounding rock state parameter matrix N as the output, and perform data regression; select 70% of the data in the excavation parameter database for training, and 30% of the data to be tested. , get a mature regression learning machine svm1, obtain the prediction result Ysvm1 of surrounding rock state parameters;
  • Step 4 Using the least squares regression method to take the ascending segment heading parameter matrix M1 as input, and the surrounding rock state parameter matrix N as the output to obtain the mathematical model as follows:
  • Step 5 Mathematical average of the prediction results of the surrounding rock state parameters Ynet1, Ysvm1 and Yreg1 obtained in steps 2-4: The obtained rock machine mutual feed model Y1.
  • the method for obtaining the intelligent control decision model is:
  • the average value of the cyclic excavation parameters constitutes the steady-state segmentation parameter matrix
  • Step 2 Establish a three-layer neural network with the surrounding rock state parameter matrix N as the input, the steady-state segmentation parameter matrix M2 as the output, and supervise the learning of the initial neural network; select 70% of the data in the excavation parameter database. Training, 30% of the data is tested, and a mature neural network Net2 is obtained, and the prediction result of surrounding rock state parameters is obtained Ynet2;
  • Step 3 Using the support vector machine to take the surrounding rock state parameter matrix N as input, the steady-state segment excavation parameter matrix M2 as the output, and perform data regression; select 70% of the data in the excavation parameter database for training, and 30% of the data is performed. Test, get a mature regression learning machine svm2, obtain the surrounding rock state parameter prediction result Ysvm2;
  • Step 4 Using the minimum regression method, the surrounding rock state parameter matrix N is taken as the input, and the steady state segment mining parameter matrix M2 is taken as the output to obtain the mathematical model as follows:
  • Step 5 Perform the mathematical average of the predicted results of the steady state segment excavation state parameters Ynet2, Ysvm2 and Yreg2 obtained in steps 2-4.
  • the current estimated parameter is: an average value of the surrounding rock parameter Nk and the steady-state heading parameter Mk predicted by the current k-th group driving parameter; the cumulative average estimated parameter is: the surrounding rock according to the rock-machine mutual feeding model Y1
  • the state parameters are estimated, and the surrounding rock parameters Ns predicted from the k-2th to the k-1 sections of the excavation parameters are obtained; if the average deviation of the current estimated parameters and the cumulative average estimated parameters is less than 10%, the current surrounding rock is stable. Segment, the current tunneling parameter effect is stable; if the average value of the estimated parameter and the cumulative average estimated parameter deviation is greater than 90%, the current tunneling parameters are unstable and need to be adjusted.
  • the manual adjustment in the fourth step is that the TBM main driver adjusts the current excavation parameters such as the cutter speed n and the propulsion speed V in real time according to the difference between the cumulative average estimated parameter and the current estimated parameter, and controls other excavation parameters within a stable range.
  • the automatic adjustment in the fourth step is: according to the difference between the pre-cumulative average estimated parameter and the current estimated parameter, the current tunneling parameters such as the cutter speed n and the advance speed are adjusted in real time by the PLC controller. V, control other excavation parameters within a stable range, and keep the TBM safe and efficient.
  • the artificial or upper computer can self-learn and self-renew the rock machine mutual feedback model and the intelligent control decision model in the TBM during the downtime or grouting time to obtain a new rock machine.
  • the mutual-feeding model and the intelligent control decision-making model; the current rock-machine mutual-feeding model and the intelligent control decision-making model predominate, and the prediction results of the new rock-machine mutual-feeding model are better than the current rock-machine mutual-feeding model or intelligent control decision-making.
  • the new rock machine mutual feedback model replaces the current rock machine mutual feedback model.
  • the new intelligent control decision model replaces the current intelligent control decision. model.
  • An intelligent decision system for hard rock TBM tunneling control parameters comprising:
  • the engineering database unit regularly acquires surrounding rock state data and tunneling parameter data from the TBM host computer;
  • the rock machine mutual feed model unit estimates the surrounding rock state parameters of the current excavation environment by using the excavation parameters of the engineering database unit;
  • the intelligent control decision model unit estimates the optimal tunneling control parameters according to the obtained surrounding rock state parameters
  • the model parameter outputs a real-time display module, communicates with the host computer through the I/O interface, and displays the output parameters on the main driver operation interface;
  • the automatic/manual excavation parameter control module is set on the upper computer operation interface, and the main driver or the system controls the TBM excavation parameters through the PLC controller according to the estimated excavation parameters;
  • the model self-learning self-updating unit is set on the upper computer operation interface, and the model is updated by the main driver to set the model update time or the appropriate time is selected by the background.
  • the appropriate time includes but is not limited to the downtime and the grouting time.
  • the parameters of the rock machine mutual feed model unit and the intelligent control decision model unit can be read and written. After the model self-learning self-updating unit is started, the parameters of the rock machine mutual feed model unit and the intelligent control decision model can be written. In, other processes can only be read and cannot be written.
  • the data obtained by the engineering database unit from the TBM host computer includes the TBM real-time heading parameters and other TBM engineering databases.
  • the other TBM engineering databases include, but are not limited to, the same structure type or a similar TBM engineering database.
  • the rock machine mutual feed model unit predicts the current surrounding rock state parameters according to the equipment operating parameters, and perceives the TBM tunneling environment in real time;
  • the intelligent control decision model unit predicts the optimal control tunneling parameters according to the current tunneling environment;
  • the output unit displays the surrounding rock state parameters and the optimal control excavation parameters in real time on the main driver operation interface;
  • the automatic/manual excavation parameter control module is used to select the excavation parameter control mode to adjust the current excavation parameters according to the recommended optimal control excavation parameters.
  • it has a self-learning self-refreshing unit to adapt to the use of different surrounding rocks, different diameters, different performance TBMs and different stages of the same TBM life cycle.
  • FIG. 1 is a flow chart of an intelligent decision method for TBM tunneling control parameters according to the present invention.
  • FIG. 2 is a schematic structural view of an intelligent decision making system for TBM tunneling control parameters according to the present invention.
  • FIG. 3 is a flow chart of obtaining a rock machine mutual feed model according to the present invention.
  • FIG. 4 is a flow chart of obtaining an intelligent control decision model according to the present invention.
  • Figure 5 is a flow chart showing the calculation and output of the model parameters of the present invention.
  • an intelligent decision method for control parameters of hard rock TBM tunneling is as follows:
  • Step 1 Establish an engineering database containing the surrounding rock state database and the excavation parameter database.
  • the engineering database includes the surrounding rock state parameters and the tunneling control parameter library for efficient excavation as a sample for subsequent use.
  • the data contained in the surrounding rock state database includes at least: single-knife thrust Ft, single-pole torque Tn, penetration P, cutter speed n, and propulsion speed V.
  • the data contained in the excavation parameter database includes at least: rock uniaxial saturated compressive strength Rc, joint number Jv of unit rock mass, and rock mass basic quality grade W.
  • the surrounding rock of the area where the surrounding rock state parameter library is located is taken and sampled, and the surrounding rock joint diagram is drawn. According to the retrieved core, an indoor experiment is performed to obtain its uniaxial saturated compressive strength Rc. According to the surrounding rock joint diagram near the core pile, the Jv value of the joint volume of the rock mass is calculated according to the engineering rock mass classification standard.
  • Step 2 According to the surrounding rock state parameters, obtain a continuous surrounding rock grade W; use the three-layer neural network, support vector machine and least squares regression method to carry out data mining and machine for surrounding rock state parameters and tunneling control parameters in the engineering database. Learning, mathematical model to obtain the rock machine mutual feedback model and intelligent control decision model.
  • the rock machine mutual feedback model is obtained based on the surrounding rock state parameter library and the ascending section tunneling control parameter library information.
  • the intelligent control decision model is obtained based on the surrounding rock state parameter library and the stability segment tunneling control parameter library information.
  • the division of the ascending and stable sections of the tunneling control parameter is to take the 10th data point of the tunneling cycle as the segmentation critical point, and the segmentation threshold includes but is not limited to the 10th data point of the tunneling cycle, or may be phase Proportional points in other locations.
  • Mathematical calculations include, but are not limited to, mathematical averages, geometric averages, or other weighted average forms.
  • the rock-machine mutual feedback model is used to predict the surrounding rock state parameters of the current excavation environment, and the optimal control parameters are predicted by the rock-machine intelligent control decision-making model.
  • the present invention improves the process by fitting the values using the values at the nodes.
  • the method for obtaining a continuous surrounding rock grade W based on the surrounding rock state parameters is:
  • the mean square error MSE 0.016 of the fitted value Kv' and the true value Kv, the relative error is less than 6%, and the prediction accuracy is greater than 94%, which is in line with engineering needs.
  • Step 2 According to the engineering rock mass grading standard, the uniaxial saturated compressive strength Rc and the fitted value Kv' are used to obtain the value of the basic quality index BQ of the rock mass.
  • Step 3 In the process of transforming the BQ value of the basic quality index of the rock mass into the surrounding rock grade W, the classification standard of the engineering rock mass is only linearly divided into five grades, which is not conducive to achieving fine control. Therefore, the present invention improves the process by fitting the values using the values at the nodes.
  • the acquisition method of the rock machine mutual feed model is:
  • Step 2 Establish a three-layer neural network, take the ascending segment excavation parameter matrix M1 as input, the surrounding rock state parameter matrix N as the output, and supervise the learning of the initial neural network; select 70% of the data in the excavation parameter database for training. 30% of the data was tested to obtain a mature neural network Net1, and the prediction result Ynet1 of the surrounding rock state parameters was obtained.
  • the three-layer neural network has 20, 10, and 10 nodes per layer. After testing, the prediction accuracy of neural network Net1 for rock uniaxial saturated compressive strength Rc, unit rock mass joint number Jv, rock mass basic quality grade W is 92.7%, 85.4%, 95.9%, respectively, to meet engineering needs.
  • Step 3 Using the support vector machine to take the ascending segment excavation parameter matrix M1 as input, the surrounding rock state parameter matrix N as the output, and perform data regression; select 70% of the data in the excavation parameter database for training, and 30% of the data to be tested. , get a mature regression learning machine svm1, get the prediction result Ysvm1 of the surrounding rock state parameters.
  • the support vector machine SVM is used for data regression. After testing, the prediction accuracy of the svm1 of the regression learning machine on the rock uniaxial saturated compressive strength Rc, the joint number JV of the unit rock volume and the basic quality grade W of the rock mass are 93.2%, 84.2%, 94.7%, respectively, to meet the engineering needs. .
  • Step 4 Using the least squares regression method to take the ascending segment excavation parameter matrix M1 as input, and the surrounding rock state parameter matrix N as the output to obtain the mathematical model as follows:
  • the neural network or support vector machine SVM is used to obtain the implicit function model, and the physical interpretation is not strong. Therefore, the present invention returns the data according to the least squares regression method.
  • the relative accuracy of the above three models established by the above three models for rock uniaxial saturated compressive compressive strength Rc, unit rock mass joint number JV, and rock mass basic quality grade W is 92.5%, respectively. 82.4%, 90.3%, to meet engineering needs.
  • Step 5 Mathematical average of the prediction results of the surrounding rock state parameters Ynet1, Ysvm1 and Yreg1 obtained in steps 2-4: The obtained rock machine mutual feed model Y1.
  • Mathematical averaging is performed on the prediction results of the three models of neural network model, support vector machine model and least squares regression model, which avoids the calculation error caused by single mathematical method in data prediction, and rationally exerts the three models.
  • the advantage of improving the accuracy of the forecast is performed on the prediction results of the three models of neural network model, support vector machine model and least squares regression model, which avoids the calculation error caused by single mathematical method in data prediction, and rationally exerts the three models.
  • the method for obtaining the intelligent control decision model is:
  • the average value of the cyclic excavation parameters constitutes the steady-state segmentation parameter matrix
  • Step 2 Establish a three-layer neural network with the surrounding rock state parameter matrix N as input, the steady-state segmentation parameter matrix M2 as the output, and supervise learning of the initial neural network; select 70% of the data in the excavation parameter database. Training, 30% of the data was tested, and a mature neural network Net2 was obtained, and the prediction result Ym2 of the surrounding rock state parameters was obtained.
  • the number of nodes in each layer of the three-layer neural network is 20, 10, and 10, respectively.
  • Tested, neural network Net2 pair The forecast accuracy rates are 93.4%, 82.4%, 91.9%, 95.4%, and 88.9%, respectively, to meet engineering needs.
  • Step 3 Using the support vector machine to take the surrounding rock state parameter matrix N as input, the steady state segment mining parameter matrix M2 as the output, and perform data regression; select 70% of the data in the tunneling parameter database for training, and 30% of the data is performed. Test, get a mature regression learning machine svm2, get the prediction result Ysvm2 of surrounding rock state parameters.
  • Step 4 Using the method of minimum regression, the surrounding rock state parameter matrix N is taken as the input, and the steady-state segmenting parameter matrix M2 is taken as the output to obtain the mathematical model as follows:
  • the steady state segment excavation state parameter prediction result Yreg2 is obtained.
  • the neural network or support vector machine SVM is used to obtain the implicit function model.
  • the physical interpretation is not meaningful. Therefore, the data is trained and predicted according to the least squares regression method.
  • the relative accuracy of the above five mathematical models for the prediction results of the test data were 91.7%, 83.6%, 90.3%, 92.8%, and 87.6%, respectively, to meet the engineering needs.
  • Step 5 Perform the mathematical average of the predicted results of the steady state segment excavation state parameters Ynet2, Ysvm2 and Yreg2 obtained in steps 2-4.
  • the comprehensive models obtained by mathematical averages are obtained based on the three models to obtain the prediction results Ynet2, Ysvm2 and Yreg2.
  • the mathematical average of the prediction results of the three models avoids the calculation error caused by the single mathematical method in the data prediction, and rationally exerts the advantages of the three models to improve the accuracy of the prediction.
  • Step 3 The rock-machine mutual-feeding model predicts the surrounding rock state parameters of the current excavation environment according to the real-time acquisition parameters of the TBM host computer.
  • the intelligent control decision model pre-predicts the optimal tunneling control parameters according to the obtained surrounding rock state parameters. Estimate, the cumulative average estimated parameters and current estimated parameters are displayed on the TBM.
  • the tunneling parameters are saved once per second and uploaded to the host computer.
  • the excavation parameters are read once every 10 seconds from the host computer and saved as m1, m2, ..., mk, respectively, assuming that the currently obtained tunneling parameter is the kth segment.
  • the interval time can be adjusted. Under normal law, when the state of the surrounding rock changes drastically, the interval time is shortened, and when the surrounding rock state is stable, the interval time can be appropriately extended.
  • the surrounding rock state parameters are estimated, and the surrounding rock state parameters are obtained.
  • the steady-state excavation parameters are estimated according to the tunneling parameter intelligent decision model Y2.
  • the current estimated parameters are: the average value of the surrounding rock parameters Nk and the steady-state driving parameters Mk predicted by the current k-th group driving parameters.
  • the cumulative average estimation parameters are: According to the rock-machine mutual feedback model Y1, the surrounding rock state parameters are estimated, and the surrounding rock parameters Ns predicted from the k-2th to the k-1 sections of the tunneling parameters are obtained.
  • the current estimated parameters are displayed in the second column of the operation interface, and the cumulative average estimated parameters are displayed in the first column of the operation interface.
  • the current surrounding rock is in the stable section, and the current tunneling parameter effect is stable; if the average value of the estimated parameter and the cumulative average estimated parameter deviation is greater than 90%, the current tunneling
  • the parameters are unstable and need to be adjusted.
  • Step 4 Automatically or manually adjust the current tunneling parameters of the TBM using the cumulative average estimated parameters and the current estimated parameters.
  • the excavation parameter control method selects the manual mode, that is, the manual adjustment is that the TBM main driver adjusts the current excavation parameters such as the cutter rotation speed n and the propulsion speed V in real time according to the difference between the cumulative average estimation parameter and the current estimation parameter, and controls other excavation parameters. Keep the TBM safe and efficient in the stable range.
  • the excavation parameter control method selects the automatic mode, that is, the automatic adjustment is: according to the difference value between the pre-accumulated average estimation parameter and the current estimation parameter, the current excavation parameters such as the cutter head rotation speed n and the propulsion speed V are controlled in real time by the PLC controller, and the control is performed. Other excavation parameters are within a stable range to keep the TBM safe and efficient.
  • Step 5 Transfer the TBM real-time excavation parameters and other TBM project databases to the engineering database, and proceed to step 2 for self-learning and self-updating of the rock machine mutual feedback model and the intelligent control decision model.
  • the model self-learning self-updating algorithm is used to update the rock machine mutual feedback model and the intelligent control decision model in real time.
  • the model self-learning self-updating algorithm is derived from the continuous enrichment of engineering data during the excavation process.
  • the engineering database can directly read the excavation parameters from the local host computer or from other TBM engineering databases.
  • Other TBMs include but are not limited to the same structure type or TBM similar to geological excavation.
  • the new engineering database is used to update the rock machine mutual feedback model and the intelligent control decision model in real time to match the use of different surrounding rocks, different diameters, different performance TBMs and different stages of the same TBM life cycle.
  • the appropriate time includes but is not limited to downtime, grouting time, and so on.
  • the artificial or upper computer performs self-learning and self-updating of the rock machine mutual feedback model and the intelligent control decision model in the TBM during downtime or grouting time.
  • a new rock machine mutual feedback model and an intelligent control decision model are obtained.
  • the current rock-machine mutual feedback model and intelligent control decision-making model dominate.
  • the new rock-machine mutual feedback model replaces the current rock-machine mutual feedback model.
  • the new intelligent control decision-making model replaces the current intelligent control decision-making model.
  • an intelligent decision system for hard rock TBM tunneling control parameters includes:
  • the engineering database unit regularly acquires surrounding rock state data and tunneling parameter data from the TBM host computer.
  • the data obtained by the engineering database unit from the TBM host computer includes the TBM real-time heading parameters and other TBM engineering databases.
  • the other TBM engineering databases include but are not limited to the same structure type or the engineering database of the similar TBM-like TBM.
  • the rock machine mutual feedback model unit uses the excavation parameters of the engineering database unit to estimate the surrounding rock state parameters of the current excavation environment.
  • the intelligent control decision model unit estimates the optimal tunneling control parameters according to the obtained surrounding rock state parameters.
  • the parameters of the rock machine mutual feed model unit and the intelligent control decision model unit can be read and written. After the model self-learning self-updating unit is started, the parameters of the rock machine mutual feed model unit and the intelligent control decision model can be written. Other processes can only be read and cannot be written.
  • the model parameter outputs a real-time display module, communicates with the host computer through the I/O interface, and displays the output parameters on the main driver operation interface.
  • the output parameters include cumulative average estimation parameters and current estimation parameters to guide the tunneling of the TBM.
  • the automatic/manual excavation parameter control module is set on the upper computer operation interface, and the main driver or system controls the TBM excavation parameters to adjust according to the estimated excavation parameters.
  • the model self-learning self-updating unit is set on the upper computer operation interface, and the model is updated by the main driver to set the model update time or the appropriate time is selected by the background.
  • the appropriate time includes but is not limited to the downtime and the grouting time.
  • the engineering database unit of the invention can continuously enrich the large database of engineering parameters, and uses the data mining and machine learning technology to create a rock machine mutual feedback model and an intelligent control decision model by using neural network, support vector machine and least squares regression method.
  • the prediction of the excavation parameters; the parameter prediction result of the model is displayed on the upper computer operation interface through the model parameter output real-time display unit module, and the excavation parameter control mode can be selected by the automatic/manual excavation parameter control module;
  • the self-learning self-updating unit of the model can be According to the continuous updating of the large database of engineering parameters, the rock-machine mutual feedback model and the intelligent control decision-making model are self-learning and self-renewing, which are used to adapt to different surrounding rock, different diameters, different performance TBMs and different stages of the same life cycle of the same TBM.

Landscapes

  • Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Mining & Mineral Resources (AREA)
  • Health & Medical Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Geochemistry & Mineralogy (AREA)
  • Geology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Environmental & Geological Engineering (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Life Sciences & Earth Sciences (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Earth Drilling (AREA)
  • Feedback Control In General (AREA)

Abstract

A smart decision making method and system for boring control parameters of a hard rock TBM. The method comprises the following components: establishing an engineering database; establishing a rock-machine mutual feedback model and a smart control decision making model according to the engineering database; model parameter output display and update; an automatic/manual boring parameter control method; a model self-learning and self-updating algorithm. The system comprises the following components: an engineering database unit; a rock-machine mutual feedback model unit; a smart control decision making model unit; a real-time model parameter output display module; an automatic/manual boring parameter control module; a model self-learning and self-learning updating unit. The rock-machine mutual feedback model predicts, according to the apparatus operation parameters, the current wall rock status parameters, and senses the TBM boring environment in real time; the smart control decision making model predicts, according to the current boring environment, the optimal boring control parameters, and timely adjusts the current boring parameters by controlling the boring parameters, so as to adapt to the current boring environment and ensure the boring process safe, effective, and stable.

Description

硬岩TBM掘进控制参数智能决策方法及系统Intelligent decision method and system for hard rock TBM tunneling control parameters 技术领域Technical field
本发明涉及隧道掘进装备施工智能控制的技术领域，尤其涉及一种硬岩TBM掘进控制参数智能决策方法及系统。The invention relates to the technical field of tunnelling equipment construction intelligent control, in particular to an intelligent decision method and system for hard rock TBM tunneling control parameters.
背景技术Background technique
硬岩隧道掘进机(以下简称硬岩TBM)是一种专门应用于开挖岩石隧道与地下通道工程的大型高科技施工装备。传统的施工方式中，主控室司机通过缓慢试掘进过程评估围岩状态参数，再通过反复调整掘进参数直到掘进参数保持稳定。这样的操作方式一方面会由于不断的缓慢试掘进和反复调整掘进参数导致大量施工时间的消耗，另一方面当围岩剧烈变化时，无法实时有效的对当前围岩的参数进行感知，导致掘进参数无法适应当前掘进环境，造成刀具的非正常磨损和刀盘系统破岩性能的下降，严重情况下会导致关键部件的破坏和停机，影响TBM的使用寿命。因此，找到一套既能实时感知当前围岩状态参数，又能快速决策寻找适合当前掘进环境的最优掘进参数的智能决策方法十分重要。Hard rock tunnel boring machine (hereinafter referred to as hard rock TBM) is a large-scale high-tech construction equipment specially used for excavation of rock tunnels and underground passages. In the traditional construction mode, the driver in the main control room evaluates the state parameters of the surrounding rock by slowly testing the progress process, and then repeatedly adjusts the excavation parameters until the excavation parameters remain stable. On the one hand, this kind of operation mode will lead to a large amount of construction time consumption due to the continuous slow test and repeated adjustment of the excavation parameters. On the other hand, when the surrounding rock changes drastically, the parameters of the current surrounding rock cannot be effectively and effectively detected in real time, resulting in excavation. The parameters can not adapt to the current tunneling environment, causing abnormal wear of the tool and the rock breaking performance of the cutter head system. In severe cases, the damage and shutdown of key components will be caused, which will affect the service life of the TBM. Therefore, it is very important to find an intelligent decision-making method that can sense the current surrounding rock state parameters in real time and quickly determine the optimal driving parameters suitable for the current tunneling environment.
发明内容Summary of the invention
针对现有的操作方式中仅能凭借主司机的人为经验去猜测围岩状态参数，依靠不断尝试不同掘进参数所导致的耗时耗力、TBM刀具非正常磨损、刀盘系统破岩性能下降、关键部件破坏和停机等技术问题，本发明提出一种硬岩TBM掘进控制参数智能决策方法及系统，利用数据挖掘和机器学习技术对掘进参数进行预测，通过自学习和自更新可以适应于不同围岩、不同直径、不同性能TBM和同一TBM全生命周期不同阶段的使用。In view of the existing operation mode, only the main driver's human experience can be used to guess the surrounding rock state parameters, and the time-consuming and labor-consuming force caused by constantly trying different tunneling parameters, the abnormal wear of the TBM cutter, and the rock-breaking performance of the cutterhead system are reduced. Technical problems such as destruction and shutdown of key components, the present invention proposes an intelligent decision-making method and system for control parameters of hard rock TBM tunneling, which uses data mining and machine learning techniques to predict the driving parameters, and can adapt to different circumferences through self-learning and self-updating. The use of rock, different diameters, different performance TBMs and different stages of the same life cycle of the same TBM.
为了达到上述目的，本发明的技术方案是这样实现的：一种硬岩TBM掘进控制参数智能决策方法，其步骤如下：In order to achieve the above object, the technical solution of the present invention is implemented as follows: an intelligent decision method for hard rock TBM tunneling control parameters, the steps are as follows:
步骤一：建立包含围岩状态数据库和掘进参数数据库的工程数据库；Step 1: Establish an engineering database including a surrounding rock state database and a tunneling parameter database;
步骤二：根据围岩状态参数得到连续的围岩等级W；利用三层神经网络、支持向量机和最小二乘回归的方法对工程数据库中的围岩状态参数和掘进控制参数进行数据挖掘和机器学习，通过数学计算获取岩机互馈模型和智能控制决策模型；Step 2: According to the surrounding rock state parameters, obtain a continuous surrounding rock grade W; use the three-layer neural network, support vector machine and least squares regression method to carry out data mining and machine for surrounding rock state parameters and tunneling control parameters in the engineering database. Learning, obtaining mathematical model of rock machine and intelligent control decision model through mathematical calculation;
步骤三：岩机互馈模型根据TBM上位机在线实时获取的掘进参数对当前掘进环境的围岩状态参数进行预估，智能控制决策模型根据获得的围岩状态参数对最佳掘进控制参数进行预估，将累积平均预估参数和当前预估参数显示在TBM上；Step 3: The rock-machine mutual-feeding model predicts the surrounding rock state parameters of the current excavation environment according to the real-time acquisition parameters of the TBM host computer. The intelligent control decision model pre-predicts the optimal tunneling control parameters according to the obtained surrounding rock state parameters. Estimate, display the cumulative average estimated parameters and current estimated parameters on the TBM;
步骤四：利用累积平均预估参数和当前预估参数对TBM的当前掘进参数进行自动或手 动调整；Step 4: Automatically or manually adjust the current tunneling parameters of the TBM by using the cumulative average estimated parameter and the current estimated parameter;
步骤五：将TBM实时掘进参数和其他TBM工程数据库传送至工程数据库，进入步骤二对岩机互馈模型和智能控制决策模型进行自学习自更新。Step 5: Transfer the TBM real-time excavation parameters and other TBM engineering databases to the engineering database, and proceed to step 2 to self-learn and self-update the rock machine mutual feedback model and the intelligent control decision model.
所述围岩状态数据库中包含的数据有：单刀推力Ft、单刀扭矩Tn、贯入度P、刀盘转速n、推进速度V；所述掘进参数数据库中包含的数据有：岩石单轴饱和抗压强度Rc、单位岩体体积的节理数Jv、围岩等级W。The data contained in the surrounding rock state database includes: single-knife thrust Ft, single-pole torque Tn, penetration P, cutter speed n, and propulsion speed V; the data included in the excavation parameter database includes: rock uniaxial saturation resistance The compressive strength Rc, the joint number Jv of the unit rock volume, and the surrounding rock grade W.
所述根据围岩状态参数得到连续的围岩等级W的方法是：The method for obtaining a continuous surrounding rock grade W according to the surrounding rock state parameter is:
步骤一：将单位岩体体积的节理数Jv转化为完整程度指标Kv的节点处进行插值，通过拟合得到拟合值Kv'＝e (-0.05Jv-0.11)； Step 1: Interpolate the node number Jv of the unit rock volume into the completeness index Kv, and obtain the fitting value Kv'=e (-0.05Jv-0.11) by fitting;
步骤二：根据工程岩体分级标准，利用单轴饱和抗压强度Rc和拟合值Kv’求取岩体基本质量指标BQ的值；Step 2: According to the engineering rock mass grading standard, the uniaxial saturated compressive strength Rc and the fitting value Kv' are used to obtain the value of the basic quality index BQ of the rock mass;
步骤三：将岩体基本质量指标BQ转化为围岩等级W的节点处进行插值，通过拟合得到围岩等级：W＝7-0.01*BQ，得到连续的围岩等级。Step 3: The basic quality index BQ of the rock mass is transformed into the node of the surrounding rock grade W, and the surrounding rock grade is obtained by fitting: W=7-0.01*BQ, and the continuous surrounding rock grade is obtained.
所述岩机互馈模型的获取方法为：The method for obtaining the rock machine mutual feed model is:
步骤一：利用工程数据库中的围岩状态数据库求取围岩状态参数矩阵N＝[Rc，Jv，W]；提取掘进参数数据库中对掘进循环有用的前10％掘进参数数据组成上升段掘进参数矩阵M1＝[Ft，Tn，P，n，V]；Step 1: Use the surrounding rock state database in the engineering database to obtain the surrounding rock state parameter matrix N=[Rc, Jv, W]; extract the top 10% of the excavation parameter data useful for the excavation cycle in the excavation parameter database to form the ascending section excavation parameters Matrix M1=[Ft,Tn,P,n,V];
步骤二：建立一个三层神经网络，将上升段掘进参数矩阵M1作为输入，围岩状态参数矩阵N作为输出，对初始神经网络进行有监督的学习；选取掘进参数数据库中70％的数据进行训练，30％的数据进行测试，得到一个成熟的神经网络Net1，获得围岩状态参数预测结果Ynet1；Step 2: Establish a three-layer neural network, take the ascending segment excavation parameter matrix M1 as input, the surrounding rock state parameter matrix N as the output, and supervise the learning of the initial neural network; select 70% of the data in the excavation parameter database for training. 30% of the data is tested, and a mature neural network Net1 is obtained, and the prediction result of surrounding rock state parameters is obtained Ynet1;
步骤三：利用支持向量机将上升段掘进参数矩阵M1作为输入，围岩状态参数矩阵N作为输出，对其进行数据回归；选取掘进参数数据库中70％的数据进行训练，30％的数据进行测试，得到一个成熟的回归学习机svm1，获得围岩状态参数预测结果Ysvm1；Step 3: Using the support vector machine to take the ascending segment excavation parameter matrix M1 as input, the surrounding rock state parameter matrix N as the output, and perform data regression; select 70% of the data in the excavation parameter database for training, and 30% of the data to be tested. , get a mature regression learning machine svm1, obtain the prediction result Ysvm1 of surrounding rock state parameters;
步骤四：利用最小二乘回归的方法将上升段掘进参数矩阵M1作为输入，围岩状态参数矩阵N作为输出得到数学模型如下：Step 4: Using the least squares regression method to take the ascending segment heading parameter matrix M1 as input, and the surrounding rock state parameter matrix N as the output to obtain the mathematical model as follows:
Figure PCTCN2018112521-appb-000001
Figure PCTCN2018112521-appb-000001
Figure PCTCN2018112521-appb-000002
Figure PCTCN2018112521-appb-000002
Figure PCTCN2018112521-appb-000003
获得围岩状态参数预测结果Yreg1；
Figure PCTCN2018112521-appb-000003
Obtaining the prediction result of the surrounding rock state parameter Yreg1;
步骤五：将步骤二-四得到的围岩状态参数预测结果Ynet1、Ysvm1和Yreg1进行数学平均：
Figure PCTCN2018112521-appb-000004
获得的岩机互馈模型Y1。 Step 5: Mathematical average of the prediction results of the surrounding rock state parameters Ynet1, Ysvm1 and Yreg1 obtained in steps 2-4:
Figure PCTCN2018112521-appb-000004
The obtained rock machine mutual feed model Y1.
所述智能控制决策模型的获取方法为：The method for obtaining the intelligent control decision model is:
步骤一：利用工程数据库中的围岩状态数据库求取围岩状态参数矩阵N＝[Rc，Jv，W]；提取掘进参数数据库中对掘进循环有用的后90％掘进参数数据，求取该掘进循环掘进参数的平均值组成稳态段掘进参数矩阵
Figure PCTCN2018112521-appb-000005
Step 1: Use the surrounding rock state database in the engineering database to obtain the surrounding rock state parameter matrix N=[Rc, Jv, W]; extract the 90% of the excavation parameter data useful for the tunneling cycle in the excavation parameter database, and obtain the excavation data. The average value of the cyclic excavation parameters constitutes the steady-state segmentation parameter matrix
Figure PCTCN2018112521-appb-000005
步骤二：建立一个三层神经网络，将围岩状态参数矩阵N作为输入，稳态段掘进参数矩阵M2作为输出，对初始神经网络进行有监督的学习；选取掘进参数数据库中70％的数据进行训练，30％的数据进行测试，得到一个成熟的神经网络Net2，获得围岩状态参数预测结果Ynet2；Step 2: Establish a three-layer neural network with the surrounding rock state parameter matrix N as the input, the steady-state segmentation parameter matrix M2 as the output, and supervise the learning of the initial neural network; select 70% of the data in the excavation parameter database. Training, 30% of the data is tested, and a mature neural network Net2 is obtained, and the prediction result of surrounding rock state parameters is obtained Ynet2;
步骤三：利用支持向量机将围岩状态参数矩阵N作为输入，稳态段掘进参数矩阵M2作为输出，对其进行数据回归；选取掘进参数数据库中70％的数据进行训练，30％的数据进行测试，得到一个成熟的回归学习机svm2，获得围岩状态参数预测结果Ysvm2；Step 3: Using the support vector machine to take the surrounding rock state parameter matrix N as input, the steady-state segment excavation parameter matrix M2 as the output, and perform data regression; select 70% of the data in the excavation parameter database for training, and 30% of the data is performed. Test, get a mature regression learning machine svm2, obtain the surrounding rock state parameter prediction result Ysvm2;
步骤四：利用最小回归的方法将将围岩状态参数矩阵N作为输入，稳态段掘进参数矩阵M2作为输出得到数学模型如下：Step 4: Using the minimum regression method, the surrounding rock state parameter matrix N is taken as the input, and the steady state segment mining parameter matrix M2 is taken as the output to obtain the mathematical model as follows:
Ft＝464.853-9.012*Jv-80.4*W+0.118*RcFt=464.853-9.012*Jv-80.4*W+0.118*Rc
Tn＝61.865-7.012*Jv-7.4*W+0.064*RcTn=61.865-7.012*Jv-7.4*W+0.064*Rc
P＝0.1105+0.12*Jv+2.466*W+0.058*RcP=0.1105+0.12*Jv+2.466*W+0.058*Rc
n＝11.743+0.16*Jv-1.132*W-0.044*Rcn=11.743+0.16*Jv-1.132*W-0.044*Rc
V＝p*n；V=p*n;
获得稳态段掘进状态参数预测结果Yreg2；Obtaining the steady state segment excavation state parameter prediction result Yreg2;
步骤五：将步骤二-四得到的稳态段掘进状态参数的预测结果Ynet2、Ysvm2和Yreg2,进行数学平均：
Figure PCTCN2018112521-appb-000006
获得的智能控制决策模型Y2。 Step 5: Perform the mathematical average of the predicted results of the steady state segment excavation state parameters Ynet2, Ysvm2 and Yreg2 obtained in steps 2-4.
Figure PCTCN2018112521-appb-000006
The obtained intelligent control decision model Y2.
所述当前预估参数为：当前第k组掘进参数预测得到的围岩参数Nk和稳态掘进参数Mk 的平均值；所述累积平均预估参数为：根据岩机互馈模型Y1对围岩状态参数进行预估，得到掘进参数第k-2段到k-1段预测得到的围岩参数Ns；若当前预估参数和累积平均预估参数偏差平均值小于10％，当前围岩处于稳定段，当前掘进参数效果稳定；如果预估参数和累积平均预估参数偏差平均值大于90％，当前掘进参数不稳定，需要调整。The current estimated parameter is: an average value of the surrounding rock parameter Nk and the steady-state heading parameter Mk predicted by the current k-th group driving parameter; the cumulative average estimated parameter is: the surrounding rock according to the rock-machine mutual feeding model Y1 The state parameters are estimated, and the surrounding rock parameters Ns predicted from the k-2th to the k-1 sections of the excavation parameters are obtained; if the average deviation of the current estimated parameters and the cumulative average estimated parameters is less than 10%, the current surrounding rock is stable. Segment, the current tunneling parameter effect is stable; if the average value of the estimated parameter and the cumulative average estimated parameter deviation is greater than 90%, the current tunneling parameters are unstable and need to be adjusted.
所述步骤四中手动调整是TBM主司机根据累积平均预估参数和当前预估参数的差值大小，实时调整当前掘进参数如刀盘转速n和推进速度V，控制其他掘进参数在稳定范围内，保持TBM安全高效掘进；所述步骤四中自动调整是：根据预累积平均预估参数和当前预估参数的差值大小，通过PLC控制器实时调整当前掘进参数如刀盘转速n和推进速度V，控制其他掘进参数在稳定范围内，保持TBM安全高效掘进。The manual adjustment in the fourth step is that the TBM main driver adjusts the current excavation parameters such as the cutter speed n and the propulsion speed V in real time according to the difference between the cumulative average estimated parameter and the current estimated parameter, and controls other excavation parameters within a stable range. Keep the TBM safe and efficient; the automatic adjustment in the fourth step is: according to the difference between the pre-cumulative average estimated parameter and the current estimated parameter, the current tunneling parameters such as the cutter speed n and the advance speed are adjusted in real time by the PLC controller. V, control other excavation parameters within a stable range, and keep the TBM safe and efficient.
当工程数据库中的数据增幅大于30％时，由人工或上位机在停机时间或注浆时间对TBM中岩机互馈模型和智能控制决策模型进行自学习和自更新，从而获得新的岩机互馈模型和智能控制决策模型；更新后当前的岩机互馈模型和智能控制决策模型占主导地位，新的岩机互馈模型的预测结果优于当前的岩机互馈模型或智能控制决策模型时，新的岩机互馈模型替换当前的岩机互馈模型，新的智能控制决策模型的预测结果优于当前的智能控制决策模型时，新的智能控制决策模型替换当前的智能控制决策模型。When the data in the engineering database increases by more than 30%, the artificial or upper computer can self-learn and self-renew the rock machine mutual feedback model and the intelligent control decision model in the TBM during the downtime or grouting time to obtain a new rock machine. The mutual-feeding model and the intelligent control decision-making model; the current rock-machine mutual-feeding model and the intelligent control decision-making model predominate, and the prediction results of the new rock-machine mutual-feeding model are better than the current rock-machine mutual-feeding model or intelligent control decision-making. In the model, the new rock machine mutual feedback model replaces the current rock machine mutual feedback model. When the prediction result of the new intelligent control decision model is better than the current intelligent control decision model, the new intelligent control decision model replaces the current intelligent control decision. model.
一种硬岩TBM掘进控制参数智能决策系统，包括：An intelligent decision system for hard rock TBM tunneling control parameters, comprising:
工程数据库单元，定期从TBM上位机获取围岩状态数据和掘进参数数据；The engineering database unit regularly acquires surrounding rock state data and tunneling parameter data from the TBM host computer;
岩机互馈模型单元，利用工程数据库单元的掘进参数对当前掘进环境的围岩状态参数进行预估；The rock machine mutual feed model unit estimates the surrounding rock state parameters of the current excavation environment by using the excavation parameters of the engineering database unit;
智能控制决策模型单元，根据获得的围岩状态参数对最佳掘进控制参数进行预估；The intelligent control decision model unit estimates the optimal tunneling control parameters according to the obtained surrounding rock state parameters;
模型参数输出实时显示模块，通过I/O接口与上位机进行通讯，将输出参数在主司机操作界面显示；The model parameter outputs a real-time display module, communicates with the host computer through the I/O interface, and displays the output parameters on the main driver operation interface;
自动/手动掘进参数控制模块，设置在上位机操作界面上，由主司机或系统依据预估的掘进参数，通过PLC控制器控制TBM掘进参数进行调整；The automatic/manual excavation parameter control module is set on the upper computer operation interface, and the main driver or the system controls the TBM excavation parameters through the PLC controller according to the estimated excavation parameters;
模型自学习自更新单元，设置在上位机操作界面上，由主司机设置模型更新时间或者由后台选择恰当时间对模型进行自动更新，恰当时间包含但不限于停机时间和注浆时间。The model self-learning self-updating unit is set on the upper computer operation interface, and the model is updated by the main driver to set the model update time or the appropriate time is selected by the background. The appropriate time includes but is not limited to the downtime and the grouting time.
所述岩机互馈模型单元和智能控制决策模型单元的参数可以被读取和写入，当模型自学习自更新单元启动后，岩机互馈模型单元和智能控制决策模型的参数可以被写入，其他过程只能被读取，不能写入。The parameters of the rock machine mutual feed model unit and the intelligent control decision model unit can be read and written. After the model self-learning self-updating unit is started, the parameters of the rock machine mutual feed model unit and the intelligent control decision model can be written. In, other processes can only be read and cannot be written.
所述工程数据库单元从TBM上位机获取的数据包括TBM实时掘进参数和其他TBM的工程数据库，其他TBM的工程数据库包含但不限于同一结构类型或类似地质掘进的TBM的 工程数据库。The data obtained by the engineering database unit from the TBM host computer includes the TBM real-time heading parameters and other TBM engineering databases. The other TBM engineering databases include, but are not limited to, the same structure type or a similar TBM engineering database.
本发明的有益效果：岩机互馈模型单元根据设备运行参数预测当前围岩状态参数，实时感知TBM掘进环境；智能控制决策模型单元根据当前掘进环境，预测出最优控制掘进参数；通过模型参数输出单元将围岩状态参数和最优控制掘进参数实时显示在主司机操作界面上；自动/手动掘进参数控制模块用于选择掘进参数控制方式从而根据推荐的最优控制掘进参数及时调整当前掘进参数以适应当前掘进环境，保持掘进的安全高效稳定；同时具备自学习自更新单元，以适应不同围岩、不同直径、不同性能TBM和同一TBM全生命周期不同阶段的使用。The beneficial effects of the invention: the rock machine mutual feed model unit predicts the current surrounding rock state parameters according to the equipment operating parameters, and perceives the TBM tunneling environment in real time; the intelligent control decision model unit predicts the optimal control tunneling parameters according to the current tunneling environment; The output unit displays the surrounding rock state parameters and the optimal control excavation parameters in real time on the main driver operation interface; the automatic/manual excavation parameter control module is used to select the excavation parameter control mode to adjust the current excavation parameters according to the recommended optimal control excavation parameters. In order to adapt to the current excavation environment, to maintain safe, efficient and stable excavation; at the same time, it has a self-learning self-refreshing unit to adapt to the use of different surrounding rocks, different diameters, different performance TBMs and different stages of the same TBM life cycle.
附图说明DRAWINGS
为了更清楚地说明本发明实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments or the description of the prior art will be briefly described below. Obviously, the drawings in the following description are only It is a certain embodiment of the present invention, and other drawings can be obtained from those skilled in the art without any creative work.
图1为本发明TBM掘进控制参数智能决策方法的流程图。1 is a flow chart of an intelligent decision method for TBM tunneling control parameters according to the present invention.
图2为本发明TBM掘进控制参数智能决策系统的结构示意图。2 is a schematic structural view of an intelligent decision making system for TBM tunneling control parameters according to the present invention.
图3为本发明岩机互馈模型的求取流程图。FIG. 3 is a flow chart of obtaining a rock machine mutual feed model according to the present invention.
图4为本发明智能控制决策模型的求取流程图。FIG. 4 is a flow chart of obtaining an intelligent control decision model according to the present invention.
图5为本发明模型参数的计算和输出流程图。Figure 5 is a flow chart showing the calculation and output of the model parameters of the present invention.
图6为本发明模型自学习自更新流程图。6 is a flow chart of self-learning self-updating of the model of the present invention
具体实施方式Detailed ways
下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有付出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。The technical solutions in the embodiments of the present invention are clearly and completely described in the following with reference to the accompanying drawings in the embodiments of the present invention. It is obvious that the described embodiments are only a part of the embodiments of the present invention, but not all embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments of the present invention without departing from the inventive scope are the scope of the present invention.
如图1所示，一种硬岩TBM掘进控制参数智能决策方法，其步骤如下：As shown in Figure 1, an intelligent decision method for control parameters of hard rock TBM tunneling is as follows:
步骤一：建立包含围岩状态数据库和掘进参数数据库的工程数据库。Step 1: Establish an engineering database containing the surrounding rock state database and the excavation parameter database.
工程数据库包括高效掘进时的围岩状态参数和掘进控制参数库，其作为样本以供后续使用。围岩状态数据库中包含的数据至少包括有：单刀推力Ft、单刀扭矩Tn、贯入度P、刀盘转速n、推进速度V。掘进参数数据库中包含的数据至少包括有：岩石单轴饱和抗压强度Rc、单位岩体体积的节理数Jv、岩体基本质量等级W。The engineering database includes the surrounding rock state parameters and the tunneling control parameter library for efficient excavation as a sample for subsequent use. The data contained in the surrounding rock state database includes at least: single-knife thrust Ft, single-pole torque Tn, penetration P, cutter speed n, and propulsion speed V. The data contained in the excavation parameter database includes at least: rock uniaxial saturated compressive strength Rc, joint number Jv of unit rock mass, and rock mass basic quality grade W.
对围岩状态参数库所在区域的围岩进行取芯留样，并绘制围岩节理图。根据取回的岩芯，做 室内实验求取其单轴饱和抗压强度Rc。根据取芯桩号附近的围岩节理图，依据工程岩体分级标准统计单位岩体体积的节理数Jv值。The surrounding rock of the area where the surrounding rock state parameter library is located is taken and sampled, and the surrounding rock joint diagram is drawn. According to the retrieved core, an indoor experiment is performed to obtain its uniaxial saturated compressive strength Rc. According to the surrounding rock joint diagram near the core pile, the Jv value of the joint volume of the rock mass is calculated according to the engineering rock mass classification standard.
步骤二：根据围岩状态参数得到连续的围岩等级W；利用三层神经网络、支持向量机和最小二乘回归的方法对工程数据库中的围岩状态参数和掘进控制参数进行数据挖掘和机器学习，通过数学计算获取岩机互馈模型和智能控制决策模型。Step 2: According to the surrounding rock state parameters, obtain a continuous surrounding rock grade W; use the three-layer neural network, support vector machine and least squares regression method to carry out data mining and machine for surrounding rock state parameters and tunneling control parameters in the engineering database. Learning, mathematical model to obtain the rock machine mutual feedback model and intelligent control decision model.
岩机互馈模型是根据围岩状态参数库和上升段掘进控制参数库信息进行求取的。智能控制决策模型是根据围岩状态参数库和稳定段掘进控制参数库信息进行求取的。掘进控制参数的上升段和稳定段的划分是取该掘进循环的第10％个数据点作为分割临界点，分割临界点包含但不限于该掘进循环的第10％个数据点，也可以是相临比例的其他位置点。数学计算的方式包含但不限于数学平均，几何平均或其他加权平均形式。利用岩机互馈模型对当前掘进环境的围岩状态参数进行预测，利用岩机智能控制决策模型对最佳掘进参数进行预测。The rock machine mutual feedback model is obtained based on the surrounding rock state parameter library and the ascending section tunneling control parameter library information. The intelligent control decision model is obtained based on the surrounding rock state parameter library and the stability segment tunneling control parameter library information. The division of the ascending and stable sections of the tunneling control parameter is to take the 10th data point of the tunneling cycle as the segmentation critical point, and the segmentation threshold includes but is not limited to the 10th data point of the tunneling cycle, or may be phase Proportional points in other locations. Mathematical calculations include, but are not limited to, mathematical averages, geometric averages, or other weighted average forms. The rock-machine mutual feedback model is used to predict the surrounding rock state parameters of the current excavation environment, and the optimal control parameters are predicted by the rock-machine intelligent control decision-making model.
从单位岩体体积的节理数Jv值转化为完整程度指标Kv值的过程中，由于工程岩体分级标准给出的转化方式求得的是离散的Kv值，不利于实现精细化控制。因此，本发明对该过程进行改进，利用节点处的值进行拟合插值。根据围岩状态参数得到连续的围岩等级W的方法是：In the process of transforming the joint number Jv value of the unit rock mass into the Kv value of the completeness index, the discrete Kv value is obtained due to the transformation mode given by the engineering rock mass grading standard, which is not conducive to achieving fine control. Therefore, the present invention improves the process by fitting the values using the values at the nodes. The method for obtaining a continuous surrounding rock grade W based on the surrounding rock state parameters is:
步骤1：将单位岩体体积的节理数Jv转化为完整程度指标Kv的节点处进行插值，通过拟合得到拟合值Kv'＝e (-0.05Jv-0.11)。 Step 1: Interpolate the joint number Jv of the unit rock mass into the completeness index Kv, and obtain the fitted value Kv'=e (-0.05Jv-0.11) by fitting.
经计算，拟合值Kv’与真实值Kv的均方误差MSE＝0.016，相对误差小于6％,预测精度大于94％，符合工程需要。After calculation, the mean square error MSE=0.016 of the fitted value Kv' and the true value Kv, the relative error is less than 6%, and the prediction accuracy is greater than 94%, which is in line with engineering needs.
步骤2：根据工程岩体分级标准，利用单轴饱和抗压强度Rc和拟合值Kv’求取岩体基本质量指标BQ的值。Step 2: According to the engineering rock mass grading standard, the uniaxial saturated compressive strength Rc and the fitted value Kv' are used to obtain the value of the basic quality index BQ of the rock mass.
步骤3：从岩体基本质量指标BQ值转化为围岩等级W的过程中，由于工程岩体分级标准只是线性的划分为5个等级，不利于实现精细化控制。因此，本发明对该过程进行改进，利用节点处的值进行拟合插值。将岩体基本质量指标BQ转化为围岩等级W的节点处进行插值，通过拟合得到围岩等级：W＝7-0.01*BQ，得到围岩等级W一个连续的输出。Step 3: In the process of transforming the BQ value of the basic quality index of the rock mass into the surrounding rock grade W, the classification standard of the engineering rock mass is only linearly divided into five grades, which is not conducive to achieving fine control. Therefore, the present invention improves the process by fitting the values using the values at the nodes. The basic quality index BQ of the rock mass is transformed into the node of the surrounding rock grade W, and the surrounding rock grade is obtained by fitting: W=7-0.01*BQ, and a continuous output of the surrounding rock grade W is obtained.
如图3所示，岩机互馈模型的获取方法为：As shown in Figure 3, the acquisition method of the rock machine mutual feed model is:
步骤1：利用工程数据库中的围岩状态数据库求取围岩状态参数矩阵N＝[Rc，Jv，W]；提取掘进参数数据库中对掘进循环有用的前10％掘进参数数据组成上升段掘进参数矩阵M1＝[Ft，Tn，P，n，V]；Step 1: Use the surrounding rock state database in the engineering database to obtain the surrounding rock state parameter matrix N=[Rc, Jv, W]; extract the top 10% of the excavation parameter data useful for the tunneling cycle in the excavation parameter database to form the ascending section excavation parameters Matrix M1=[Ft,Tn,P,n,V];
步骤2：建立一个三层神经网络，将上升段掘进参数矩阵M1作为输入，围岩状态参数矩阵N作为输出，对初始神经网络进行有监督的学习；选取掘进参数数据库中70％的数据进行 训练，30％的数据进行测试，得到一个成熟的神经网络Net1，获得围岩状态参数预测结果Ynet1。Step 2: Establish a three-layer neural network, take the ascending segment excavation parameter matrix M1 as input, the surrounding rock state parameter matrix N as the output, and supervise the learning of the initial neural network; select 70% of the data in the excavation parameter database for training. 30% of the data was tested to obtain a mature neural network Net1, and the prediction result Ynet1 of the surrounding rock state parameters was obtained.
三层神经网络，每层节点数分别为20，10，10。经测试，神经网络Net1对岩石单轴饱和抗压强度Rc、单位岩体体积的节理数Jv、岩体基本质量等级W的预测准确率分别为92.7％、85.4％、95.9％，满足工程需要。The three-layer neural network has 20, 10, and 10 nodes per layer. After testing, the prediction accuracy of neural network Net1 for rock uniaxial saturated compressive strength Rc, unit rock mass joint number Jv, rock mass basic quality grade W is 92.7%, 85.4%, 95.9%, respectively, to meet engineering needs.
步骤3：利用支持向量机将上升段掘进参数矩阵M1作为输入，围岩状态参数矩阵N作为输出，对其进行数据回归；选取掘进参数数据库中70％的数据进行训练，30％的数据进行测试，得到一个成熟的回归学习机svm1，获得围岩状态参数预测结果Ysvm1。Step 3: Using the support vector machine to take the ascending segment excavation parameter matrix M1 as input, the surrounding rock state parameter matrix N as the output, and perform data regression; select 70% of the data in the excavation parameter database for training, and 30% of the data to be tested. , get a mature regression learning machine svm1, get the prediction result Ysvm1 of the surrounding rock state parameters.
为提高预测准确度，同时使用支持向量机SVM进行数据回归。经测试，回归学习机svm1对岩石单轴饱和抗压强度Rc、单位岩体体积的节理数JV、岩体基本质量等级W的预测准确率分别为93.2％，84.2％，94.7％，满足工程需要。In order to improve the prediction accuracy, the support vector machine SVM is used for data regression. After testing, the prediction accuracy of the svm1 of the regression learning machine on the rock uniaxial saturated compressive strength Rc, the joint number JV of the unit rock volume and the basic quality grade W of the rock mass are 93.2%, 84.2%, 94.7%, respectively, to meet the engineering needs. .
步骤4：利用最小二乘回归的方法将上升段掘进参数矩阵M1作为输入，围岩状态参数矩阵N作为输出得到数学模型如下：Step 4: Using the least squares regression method to take the ascending segment excavation parameter matrix M1 as input, and the surrounding rock state parameter matrix N as the output to obtain the mathematical model as follows:
Figure PCTCN2018112521-appb-000007
Figure PCTCN2018112521-appb-000007
Figure PCTCN2018112521-appb-000008
Figure PCTCN2018112521-appb-000008
Figure PCTCN2018112521-appb-000009
获得围岩状态参数预测结果Yreg1。
Figure PCTCN2018112521-appb-000009
The prediction result YGE1 of the surrounding rock state parameter is obtained.
使用神经网络或者支持向量机SVM求取的都是隐函数模型，物理解释意义不强，因此，本发明根据最小二乘回归的方法对数据进行回归。经测试，最小回归的方法建立的上述三个模型对岩石单轴饱和抗压强度Rc、单位岩体体积的节理数JV、岩体基本质量等级W的预测结果的相对准确率分别为92.5％，82.4％，90.3％，满足工程需要。The neural network or support vector machine SVM is used to obtain the implicit function model, and the physical interpretation is not strong. Therefore, the present invention returns the data according to the least squares regression method. After testing, the relative accuracy of the above three models established by the above three models for rock uniaxial saturated compressive compressive strength Rc, unit rock mass joint number JV, and rock mass basic quality grade W is 92.5%, respectively. 82.4%, 90.3%, to meet engineering needs.
步骤5：将步骤二-四得到的围岩状态参数预测结果Ynet1、Ysvm1和Yreg1进行数学平均：
Figure PCTCN2018112521-appb-000010
获得的岩机互馈模型Y1。 Step 5: Mathematical average of the prediction results of the surrounding rock state parameters Ynet1, Ysvm1 and Yreg1 obtained in steps 2-4:
Figure PCTCN2018112521-appb-000010
The obtained rock machine mutual feed model Y1.
将神经网络模型、支持向量机模型和最小二乘回归模型三大模型的预测结果进行数学平均，既避免了单一的数学方法在进行数据预测时造成的计算误差，同时合理的发挥三大模型各自的优势，提高预测的准确度。Mathematical averaging is performed on the prediction results of the three models of neural network model, support vector machine model and least squares regression model, which avoids the calculation error caused by single mathematical method in data prediction, and rationally exerts the three models. The advantage of improving the accuracy of the forecast.
如图4所示，所述智能控制决策模型的获取方法为：As shown in FIG. 4, the method for obtaining the intelligent control decision model is:
步骤1：利用工程数据库中的围岩状态数据库求取围岩状态参数矩阵N＝[Rc，Jv，W]；提取掘进参数数据库中对掘进循环有用的后90％掘进参数数据，求取该掘进循环掘进参数的平均值组成稳态段掘进参数矩阵
Figure PCTCN2018112521-appb-000011
Step 1: Use the surrounding rock state database in the engineering database to obtain the surrounding rock state parameter matrix N=[Rc, Jv, W]; extract the 90% of the excavation parameter data useful for the tunneling cycle in the excavation parameter database, and obtain the excavation data. The average value of the cyclic excavation parameters constitutes the steady-state segmentation parameter matrix
Figure PCTCN2018112521-appb-000011
步骤2：建立一个三层神经网络，将围岩状态参数矩阵N作为输入，稳态段掘进参数矩阵M2作为输出，对初始神经网络进行有监督的学习；选取掘进参数数据库中70％的数据进行训练，30％的数据进行测试，得到一个成熟的神经网络Net2，获得围岩状态参数预测结果Ynet2。Step 2: Establish a three-layer neural network with the surrounding rock state parameter matrix N as input, the steady-state segmentation parameter matrix M2 as the output, and supervise learning of the initial neural network; select 70% of the data in the excavation parameter database. Training, 30% of the data was tested, and a mature neural network Net2 was obtained, and the prediction result Ym2 of the surrounding rock state parameters was obtained.
三层神经网络的每层节点数分别为20，10，10。经测试，神经网络Net2对
Figure PCTCN2018112521-appb-000012
Figure PCTCN2018112521-appb-000013
的预测准确率分别为93.4％，82.4％，91.9％，95.4％，88.9％，满足工程需要。 The number of nodes in each layer of the three-layer neural network is 20, 10, and 10, respectively. Tested, neural network Net2 pair
Figure PCTCN2018112521-appb-000012
Figure PCTCN2018112521-appb-000013
The forecast accuracy rates are 93.4%, 82.4%, 91.9%, 95.4%, and 88.9%, respectively, to meet engineering needs.
步骤3：利用支持向量机将围岩状态参数矩阵N作为输入，稳态段掘进参数矩阵M2作为输出，对其进行数据回归；选取掘进参数数据库中70％的数据进行训练，30％的数据进行测试，得到一个成熟的回归学习机svm2，获得围岩状态参数预测结果Ysvm2。Step 3: Using the support vector machine to take the surrounding rock state parameter matrix N as input, the steady state segment mining parameter matrix M2 as the output, and perform data regression; select 70% of the data in the tunneling parameter database for training, and 30% of the data is performed. Test, get a mature regression learning machine svm2, get the prediction result Ysvm2 of surrounding rock state parameters.
为提高预测准确度，使用支持向量机SVM进行数据回归。经测试，回归学习机svm2对
Figure PCTCN2018112521-appb-000014
Figure PCTCN2018112521-appb-000015
的预测准确率分别得到92.6％，84.7％，90.3％，94.8％，90.2％，满足工程需要。 To improve prediction accuracy, support vector machine SVM is used for data regression. After testing, the regression learning machine svm2 pair
Figure PCTCN2018112521-appb-000014
Figure PCTCN2018112521-appb-000015
The prediction accuracy rate was 92.6%, 84.7%, 90.3%, 94.8%, and 90.2%, respectively, to meet the engineering needs.
步骤4：利用最小回归的方法将将围岩状态参数矩阵N作为输入，稳态段掘进参数矩阵M2作为输出得到数学模型如下：Step 4: Using the method of minimum regression, the surrounding rock state parameter matrix N is taken as the input, and the steady-state segmenting parameter matrix M2 is taken as the output to obtain the mathematical model as follows:
Ft＝464.853-9.012*Jv-80.4*W+0.118*RcFt=464.853-9.012*Jv-80.4*W+0.118*Rc
Tn＝61.865-7.012*Jv-7.4*W+0.064*RcTn=61.865-7.012*Jv-7.4*W+0.064*Rc
P＝0.1105+0.12*Jv+2.466*W+0.058*RcP=0.1105+0.12*Jv+2.466*W+0.058*Rc
n＝11.743+0.16*Jv-1.132*W-0.044*Rcn=11.743+0.16*Jv-1.132*W-0.044*Rc
V＝p*n；V=p*n;
获得稳态段掘进状态参数预测结果Yreg2。The steady state segment excavation state parameter prediction result Yreg2 is obtained.
使用神经网络或者支持向量机SVM求取的都是隐函数模型，物理解释意义不强，因此根据最小二乘回归的方法对数据进行训练和预测。上述五个数学模型对测试数据预测结果的相对准确率分别为91.7％，83.6％，90.3％，92.8％，87.6％，满足工程需要。The neural network or support vector machine SVM is used to obtain the implicit function model. The physical interpretation is not meaningful. Therefore, the data is trained and predicted according to the least squares regression method. The relative accuracy of the above five mathematical models for the prediction results of the test data were 91.7%, 83.6%, 90.3%, 92.8%, and 87.6%, respectively, to meet the engineering needs.
步骤5：将步骤二-四得到的稳态段掘进状态参数的预测结果Ynet2、Ysvm2和Yreg2,进 行数学平均：
Figure PCTCN2018112521-appb-000016
获得的智能控制决策模型Y2。 Step 5: Perform the mathematical average of the predicted results of the steady state segment excavation state parameters Ynet2, Ysvm2 and Yreg2 obtained in steps 2-4.
Figure PCTCN2018112521-appb-000016
The obtained intelligent control decision model Y2.
在利用围岩参数对稳态段掘进状态参数进行预测时，同时根据三大模型求取预测结果Ynet2,Ysvm2,Yreg2进行数学平均得到的综合模型。将三大模型的预测结果进行数学平均，既避免了单一的数学方法在进行数据预测时造成的计算误差，同时合理的发挥三大模型各自的优势，提高预测的准确度。When using the surrounding rock parameters to predict the state parameters of the steady-state section, the comprehensive models obtained by mathematical averages are obtained based on the three models to obtain the prediction results Ynet2, Ysvm2 and Yreg2. The mathematical average of the prediction results of the three models avoids the calculation error caused by the single mathematical method in the data prediction, and rationally exerts the advantages of the three models to improve the accuracy of the prediction.
步骤三：岩机互馈模型根据TBM上位机在线实时获取的掘进参数对当前掘进环境的围岩状态参数进行预估，智能控制决策模型根据获得的围岩状态参数对最佳掘进控制参数进行预估，将累积平均预估参数和当前预估参数显示在TBM上。Step 3: The rock-machine mutual-feeding model predicts the surrounding rock state parameters of the current excavation environment according to the real-time acquisition parameters of the TBM host computer. The intelligent control decision model pre-predicts the optimal tunneling control parameters according to the obtained surrounding rock state parameters. Estimate, the cumulative average estimated parameters and current estimated parameters are displayed on the TBM.
如图5所示，当掘进机TBM进入一个新的掘进循环时，掘进参数按每秒保存一次上传至上位机。从上位机每隔10秒读取一次掘进参数，分别保存为m1，m2，…，mk，假设当前得到的掘进参数为第k段。间隔时间可调节，一般规律下，围岩状态变化剧烈时，缩短间隔时间，围岩状态较稳定时，可适当延长间隔时间。As shown in Figure 5, when the roadheader TBM enters a new tunneling cycle, the tunneling parameters are saved once per second and uploaded to the host computer. The excavation parameters are read once every 10 seconds from the host computer and saved as m1, m2, ..., mk, respectively, assuming that the currently obtained tunneling parameter is the kth segment. The interval time can be adjusted. Under normal law, when the state of the surrounding rock changes drastically, the interval time is shortened, and when the surrounding rock state is stable, the interval time can be appropriately extended.
按照岩机互馈模型Y1对围岩状态参数进行预估，得到围岩状态参数。按照掘进参数智能决策模型Y2对稳态掘进参数进行预估。当前预估参数为：当前第k组掘进参数预测得到的围岩参数Nk和稳态掘进参数Mk的平均值。累积平均预估参数为：根据岩机互馈模型Y1对围岩状态参数进行预估，得到掘进参数第k-2段到k-1段预测得到的围岩参数Ns。当前预估参数显示在操作界面第二列，累积平均预估参数显示在操作界面第一列。According to the rock-machine mutual feedback model Y1, the surrounding rock state parameters are estimated, and the surrounding rock state parameters are obtained. The steady-state excavation parameters are estimated according to the tunneling parameter intelligent decision model Y2. The current estimated parameters are: the average value of the surrounding rock parameters Nk and the steady-state driving parameters Mk predicted by the current k-th group driving parameters. The cumulative average estimation parameters are: According to the rock-machine mutual feedback model Y1, the surrounding rock state parameters are estimated, and the surrounding rock parameters Ns predicted from the k-2th to the k-1 sections of the tunneling parameters are obtained. The current estimated parameters are displayed in the second column of the operation interface, and the cumulative average estimated parameters are displayed in the first column of the operation interface.
若当前预估参数和累积平均预估参数偏差平均值小于10％，当前围岩处于稳定段，当前掘进参数效果稳定；如果预估参数和累积平均预估参数偏差平均值大于90％，当前掘进参数不稳定，需要调整。If the average deviation of the current estimated parameter and the cumulative average estimated parameter is less than 10%, the current surrounding rock is in the stable section, and the current tunneling parameter effect is stable; if the average value of the estimated parameter and the cumulative average estimated parameter deviation is greater than 90%, the current tunneling The parameters are unstable and need to be adjusted.
步骤四：利用累积平均预估参数和当前预估参数对TBM的当前掘进参数进行自动或手动调整。Step 4: Automatically or manually adjust the current tunneling parameters of the TBM using the cumulative average estimated parameters and the current estimated parameters.
掘进参数控制方法选择手动模式，即手动调整是TBM主司机根据累积平均预估参数和当前预估参数的差值大小，实时调整当前掘进参数如刀盘转速n和推进速度V，控制其他掘进参数在稳定范围内，保持TBM安全高效掘进。掘进参数控制方法选择自动模式，即自动调整是：根据预累积平均预估参数和当前预估参数的差值大小，通过PLC控制器实时调整当前掘进参数如刀盘转速n和推进速度V，控制其他掘进参数在稳定范围内，保持TBM安全高效掘进。The excavation parameter control method selects the manual mode, that is, the manual adjustment is that the TBM main driver adjusts the current excavation parameters such as the cutter rotation speed n and the propulsion speed V in real time according to the difference between the cumulative average estimation parameter and the current estimation parameter, and controls other excavation parameters. Keep the TBM safe and efficient in the stable range. The excavation parameter control method selects the automatic mode, that is, the automatic adjustment is: according to the difference value between the pre-accumulated average estimation parameter and the current estimation parameter, the current excavation parameters such as the cutter head rotation speed n and the propulsion speed V are controlled in real time by the PLC controller, and the control is performed. Other excavation parameters are within a stable range to keep the TBM safe and efficient.
步骤五：将TBM实时掘进参数和其他TBM工程数据库传送至工程数据库，进入步骤二 对岩机互馈模型和智能控制决策模型进行自学习自更新。Step 5: Transfer the TBM real-time excavation parameters and other TBM project databases to the engineering database, and proceed to step 2 for self-learning and self-updating of the rock machine mutual feedback model and the intelligent control decision model.
利用模型自学习自更新算法对岩机互馈模型和智能控制决策模型进行实时快速更新。模型自学习自更新算法源于掘进过程中工程数据的不断丰富，工程数据库可直接从本机上位机读取掘进参数，或从其他TBM的工程数据库调用，其他TBM包含但不限于同一结构类型或类似地质掘进的TBM。新的工程数据库用于对岩机互馈模型和智能控制决策模型进行实时快速更新，以匹配不同围岩、不同直径、不同性能TBM和同一TBM全生命周期不同阶段的使用。当工程数据样本库不断丰富后，可人工选择何时更新模型或由后台程序选择恰当时间对模型进行自动更新，恰当时间包含但不限于停机时间、注浆时间等。The model self-learning self-updating algorithm is used to update the rock machine mutual feedback model and the intelligent control decision model in real time. The model self-learning self-updating algorithm is derived from the continuous enrichment of engineering data during the excavation process. The engineering database can directly read the excavation parameters from the local host computer or from other TBM engineering databases. Other TBMs include but are not limited to the same structure type or TBM similar to geological excavation. The new engineering database is used to update the rock machine mutual feedback model and the intelligent control decision model in real time to match the use of different surrounding rocks, different diameters, different performance TBMs and different stages of the same TBM life cycle. When the engineering data sample library is continuously enriched, you can manually select when to update the model or automatically update the model by the background program at the appropriate time. The appropriate time includes but is not limited to downtime, grouting time, and so on.
如图6所示，当工程数据库中的数据增幅大于30％时，由人工或上位机在停机时间或注浆时间对TBM中岩机互馈模型和智能控制决策模型进行自学习和自更新，从而获得新的岩机互馈模型和智能控制决策模型。更新后当前的岩机互馈模型和智能控制决策模型占主导地位，新的岩机互馈模型的预测结果优于当前的岩机互馈模型或智能控制决策模型时，新的岩机互馈模型替换当前的岩机互馈模型，新的智能控制决策模型的预测结果优于当前的智能控制决策模型时，新的智能控制决策模型替换当前的智能控制决策模型。As shown in Fig. 6, when the data in the engineering database increases by more than 30%, the artificial or upper computer performs self-learning and self-updating of the rock machine mutual feedback model and the intelligent control decision model in the TBM during downtime or grouting time. Thereby, a new rock machine mutual feedback model and an intelligent control decision model are obtained. After the update, the current rock-machine mutual feedback model and intelligent control decision-making model dominate. When the prediction result of the new rock-machine mutual feedback model is better than the current rock-machine mutual-feeding model or intelligent control decision-making model, the new rock-machine mutual feedback The model replaces the current rock-machine mutual feedback model. When the prediction result of the new intelligent control decision-making model is better than the current intelligent control decision-making model, the new intelligent control decision-making model replaces the current intelligent control decision-making model.
如图2所示，一种硬岩TBM掘进控制参数智能决策系统，包括：As shown in Figure 2, an intelligent decision system for hard rock TBM tunneling control parameters includes:
工程数据库单元，定期从TBM上位机获取围岩状态数据和掘进参数数据。工程数据库单元从TBM上位机获取的数据包括TBM实时掘进参数和其他TBM的工程数据库，其他TBM的工程数据库包含但不限于同一结构类型或类似地质掘进的TBM的工程数据库。The engineering database unit regularly acquires surrounding rock state data and tunneling parameter data from the TBM host computer. The data obtained by the engineering database unit from the TBM host computer includes the TBM real-time heading parameters and other TBM engineering databases. The other TBM engineering databases include but are not limited to the same structure type or the engineering database of the similar TBM-like TBM.
岩机互馈模型单元，利用工程数据库单元的掘进参数对当前掘进环境的围岩状态参数进行预估；智能控制决策模型单元，根据获得的围岩状态参数对最佳掘进控制参数进行预估。The rock machine mutual feedback model unit uses the excavation parameters of the engineering database unit to estimate the surrounding rock state parameters of the current excavation environment. The intelligent control decision model unit estimates the optimal tunneling control parameters according to the obtained surrounding rock state parameters.
岩机互馈模型单元和智能控制决策模型单元的参数可以被读取和写入，当模型自学习自更新单元启动后，岩机互馈模型单元和智能控制决策模型的参数可以被写入，其他过程只能被读取，不能写入。The parameters of the rock machine mutual feed model unit and the intelligent control decision model unit can be read and written. After the model self-learning self-updating unit is started, the parameters of the rock machine mutual feed model unit and the intelligent control decision model can be written. Other processes can only be read and cannot be written.
模型参数输出实时显示模块，通过I/O接口与上位机进行通讯，将输出参数在主司机操作界面显示。输出参数包括累积平均预估参数和当前预估参数，用于指导TBM的掘进。The model parameter outputs a real-time display module, communicates with the host computer through the I/O interface, and displays the output parameters on the main driver operation interface. The output parameters include cumulative average estimation parameters and current estimation parameters to guide the tunneling of the TBM.
自动/手动掘进参数控制模块，设置在上位机操作界面上，由主司机或系统依据预估的掘进参数，通过PLC控制器控制TBM掘进参数进行调整。The automatic/manual excavation parameter control module is set on the upper computer operation interface, and the main driver or system controls the TBM excavation parameters to adjust according to the estimated excavation parameters.
模型自学习自更新单元，设置在上位机操作界面上，由主司机设置模型更新时间或者由后台选择恰当时间对模型进行自动更新，恰当时间包含但不限于停机时间和注浆时间。The model self-learning self-updating unit is set on the upper computer operation interface, and the model is updated by the main driver to set the model update time or the appropriate time is selected by the background. The appropriate time includes but is not limited to the downtime and the grouting time.
本发明的工程数据库单元可不断丰富工程参数大数据库，利用数据挖掘和机器学习技术，使用神经网络、支持向量机、最小二乘回归的方法创建了岩机互馈模型和智能控制决策模型， 进行掘进参数的预估；模型的参数预测结果通过模型参数输出实时显示单元模块显示在上位机操作界面上，可通过自动/手动掘进参数控制模块选择掘进参数控制方式；模型的自学习自更新单元可根据工程参数大数据库的不断更新而对岩机互馈模型和智能控制决策模型进行自学习自更新，用于适应不同围岩、不同直径、不同性能TBM和同一TBM全生命周期不同阶段的使用。The engineering database unit of the invention can continuously enrich the large database of engineering parameters, and uses the data mining and machine learning technology to create a rock machine mutual feedback model and an intelligent control decision model by using neural network, support vector machine and least squares regression method. The prediction of the excavation parameters; the parameter prediction result of the model is displayed on the upper computer operation interface through the model parameter output real-time display unit module, and the excavation parameter control mode can be selected by the automatic/manual excavation parameter control module; the self-learning self-updating unit of the model can be According to the continuous updating of the large database of engineering parameters, the rock-machine mutual feedback model and the intelligent control decision-making model are self-learning and self-renewing, which are used to adapt to different surrounding rock, different diameters, different performance TBMs and different stages of the same life cycle of the same TBM.
以上对本发明一种TBM掘进控制参数智能决策方法及系统进行了详细介绍，但以上开发步骤的的说明只是用于帮助理解本发明的方法及其核心思想，不应理解为对本发明的限制。本技术领域的技术人员在本发明揭露的技术范围内，可轻易想到的变化或替换，都应涵盖在本发明的保护范围之内。The above is a detailed description of the method and system for intelligently determining the TBM tunneling control parameters of the present invention. However, the description of the above development steps is only for helping to understand the method and the core idea of the present invention, and should not be construed as limiting the present invention. Those skilled in the art will be able to devise variations or alternatives within the scope of the present invention within the scope of the present invention.

Claims (11)

  1. 一种硬岩TBM掘进控制参数智能决策方法，其特征在于，其步骤如下：An intelligent decision making method for hard rock TBM tunneling control parameters, characterized in that the steps are as follows:
    步骤一：建立包含围岩状态数据库和掘进参数数据库的工程数据库；Step 1: Establish an engineering database including a surrounding rock state database and a tunneling parameter database;
    步骤二：根据围岩状态参数得到连续的围岩等级W；利用三层神经网络、支持向量机和最小二乘回归的方法对工程数据库中的围岩状态参数和掘进控制参数进行数据挖掘和机器学习，通过数学计算获取岩机互馈模型和智能控制决策模型；Step 2: According to the surrounding rock state parameters, obtain a continuous surrounding rock grade W; use the three-layer neural network, support vector machine and least squares regression method to carry out data mining and machine for surrounding rock state parameters and tunneling control parameters in the engineering database. Learning, obtaining mathematical model of rock machine and intelligent control decision model through mathematical calculation;
    步骤三：岩机互馈模型根据TBM上位机在线实时获取的掘进参数对当前掘进环境的围岩状态参数进行预估，智能控制决策模型根据获得的围岩状态参数对最佳掘进控制参数进行预估，将累积平均预估参数和当前预估参数显示在TBM上；Step 3: The rock-machine mutual-feeding model predicts the surrounding rock state parameters of the current excavation environment according to the real-time acquisition parameters of the TBM host computer. The intelligent control decision model pre-predicts the optimal tunneling control parameters according to the obtained surrounding rock state parameters. Estimate, display the cumulative average estimated parameters and current estimated parameters on the TBM;
    步骤四：利用累积平均预估参数和当前预估参数对TBM的当前掘进参数进行自动或手动调整；Step 4: Automatically or manually adjust the current tunneling parameters of the TBM by using the cumulative average estimated parameter and the current estimated parameter;
    步骤五：将TBM实时掘进参数和其他TBM工程数据库传送至工程数据库，进入步骤二对岩机互馈模型和智能控制决策模型进行自学习自更新。Step 5: Transfer the TBM real-time excavation parameters and other TBM engineering databases to the engineering database, and proceed to step 2 to self-learn and self-update the rock machine mutual feedback model and the intelligent control decision model.
  2. 根据权利要求1所述的硬岩TBM掘进控制参数智能决策方法，其特征在于，所述围岩状态数据库中包含的数据有：单刀推力Ft、单刀扭矩Tn、贯入度P、刀盘转速n、推进速度V；所述掘进参数数据库中包含的数据有：岩石单轴饱和抗压强度Rc、单位岩体体积的节理数Jv、围岩等级W。The intelligent rock TBM tunneling control parameter intelligent decision making method according to claim 1, wherein the data included in the surrounding rock state database comprises: a single-knife thrust Ft, a single-pole torque Tn, a penetration degree P, and a cutter disk rotation speed n. The propulsion speed V; the data included in the excavation parameter database includes: rock uniaxial saturated compressive strength Rc, joint number Jv of unit rock mass, and surrounding rock grade W.
  3. 根据权利要求1或2所述的硬岩TBM掘进控制参数智能决策方法，其特征在于，所述根据围岩状态参数得到连续的围岩等级W的方法是：The intelligent rock TBM tunneling control parameter intelligent decision making method according to claim 1 or 2, wherein the method for obtaining a continuous surrounding rock grade W according to the surrounding rock state parameter is:
    步骤一：将单位岩体体积的节理数Jv转化为完整程度指标Kv的节点处进行插值，通过拟合得到拟合值Kv'＝e (-0.05Jv-0.11)； Step 1: Interpolate the node number Jv of the unit rock volume into the completeness index Kv, and obtain the fitting value Kv'=e (-0.05Jv-0.11) by fitting;
    步骤二：根据工程岩体分级标准，利用单轴饱和抗压强度Rc和拟合值Kv’求取岩体基本质量指标BQ的值；Step 2: According to the engineering rock mass grading standard, the uniaxial saturated compressive strength Rc and the fitting value Kv' are used to obtain the value of the basic quality index BQ of the rock mass;
    步骤三：将岩体基本质量指标BQ转化为围岩等级W的节点处进行插值，通过拟合得到围岩等级：W＝7-0.01*BQ，得到连续的围岩等级。Step 3: The basic quality index BQ of the rock mass is transformed into the node of the surrounding rock grade W, and the surrounding rock grade is obtained by fitting: W=7-0.01*BQ, and the continuous surrounding rock grade is obtained.
  4. 根据权利要求3所述的硬岩TBM掘进控制参数智能决策方法，其特征在于，所述岩机互馈模型的获取方法为：The intelligent rock TBM tunneling control parameter intelligent decision making method according to claim 3, wherein the rock machine mutual feedback model is obtained by:
    步骤一：利用工程数据库中的围岩状态数据库求取围岩状态参数矩阵N＝[Rc，Jv，W]；提取掘进参数数据库中对掘进循环有用的前10％掘进参数数据组成上升段掘进参数矩阵M1＝[Ft，Tn，P，n，V]；Step 1: Use the surrounding rock state database in the engineering database to obtain the surrounding rock state parameter matrix N=[Rc, Jv, W]; extract the top 10% of the excavation parameter data useful for the excavation cycle in the excavation parameter database to form the ascending section excavation parameters Matrix M1=[Ft,Tn,P,n,V];
    步骤二：建立一个三层神经网络，将上升段掘进参数矩阵M1作为输入，围岩状态参数矩 阵N作为输出，对初始神经网络进行有监督的学习；选取掘进参数数据库中70％的数据进行训练，30％的数据进行测试，得到一个成熟的神经网络Net1，获得围岩状态参数预测结果Ynet1；Step 2: Establish a three-layer neural network, take the ascending segment excavation parameter matrix M1 as input, the surrounding rock state parameter matrix N as the output, and supervise the learning of the initial neural network; select 70% of the data in the excavation parameter database for training. 30% of the data is tested, and a mature neural network Net1 is obtained, and the prediction result of surrounding rock state parameters is obtained Ynet1;
    步骤三：利用支持向量机将上升段掘进参数矩阵M1作为输入，围岩状态参数矩阵N作为输出，对其进行数据回归；选取掘进参数数据库中70％的数据进行训练，30％的数据进行测试，得到一个成熟的回归学习机svm1，获得围岩状态参数预测结果Ysvm1；Step 3: Using the support vector machine to take the ascending segment excavation parameter matrix M1 as input, the surrounding rock state parameter matrix N as the output, and perform data regression; select 70% of the data in the excavation parameter database for training, and 30% of the data to be tested. , get a mature regression learning machine svm1, obtain the prediction result Ysvm1 of surrounding rock state parameters;
    步骤四：利用最小二乘回归的方法将上升段掘进参数矩阵M1作为输入，围岩状态参数矩阵N作为输出得到数学模型如下：Step 4: Using the least squares regression method to take the ascending segment heading parameter matrix M1 as input, and the surrounding rock state parameter matrix N as the output to obtain the mathematical model as follows:
    Figure PCTCN2018112521-appb-100001
    Figure PCTCN2018112521-appb-100001
    Figure PCTCN2018112521-appb-100002
    Figure PCTCN2018112521-appb-100002
    Figure PCTCN2018112521-appb-100003
    Figure PCTCN2018112521-appb-100003
    获得围岩状态参数预测结果Yreg1；Obtaining the prediction result of the surrounding rock state parameter Yreg1;
    步骤五：将步骤二-四得到的围岩状态参数预测结果Ynet1、Ysvm1和Yreg1进行数学平均：
    Figure PCTCN2018112521-appb-100004
    获得的岩机互馈模型Y1。     Step 5: Mathematical average of the prediction results of the surrounding rock state parameters Ynet1, Ysvm1 and Yreg1 obtained in steps 2-4:
    Figure PCTCN2018112521-appb-100004
    The obtained rock machine mutual feed model Y1.
  5. 根据权利要求3所述的硬岩TBM掘进控制参数智能决策方法，其特征在于，所述智能控制决策模型的获取方法为：The method for intelligently determining a control parameter of a hard rock TBM tunneling control according to claim 3, wherein the method for obtaining the intelligent control decision model is:
    步骤一：利用工程数据库中的围岩状态数据库求取围岩状态参数矩阵N＝[Rc，Jv，W]；提取掘进参数数据库中对掘进循环有用的后90％掘进参数数据，求取该掘进循环掘进参数的平均值组成稳态段掘进参数矩阵
    Figure PCTCN2018112521-appb-100005
    Step 1: Use the surrounding rock state database in the engineering database to obtain the surrounding rock state parameter matrix N=[Rc, Jv, W]; extract the 90% of the excavation parameter data useful for the tunneling cycle in the excavation parameter database, and obtain the excavation data. The average value of the cyclic excavation parameters constitutes the steady-state segmentation parameter matrix
    Figure PCTCN2018112521-appb-100005
    步骤二：建立一个三层神经网络，将围岩状态参数矩阵N作为输入，稳态段掘进参数矩阵M2作为输出，对初始神经网络进行有监督的学习；选取掘进参数数据库中70％的数据进行训练，30％的数据进行测试，得到一个成熟的神经网络Net2，获得围岩状态参数预测结果Ynet2；Step 2: Establish a three-layer neural network with the surrounding rock state parameter matrix N as the input, the steady-state segmentation parameter matrix M2 as the output, and supervise the learning of the initial neural network; select 70% of the data in the excavation parameter database. Training, 30% of the data is tested, and a mature neural network Net2 is obtained, and the prediction result of surrounding rock state parameters is obtained Ynet2;
    步骤三：利用支持向量机将围岩状态参数矩阵N作为输入，稳态段掘进参数矩阵M2作为输出，对其进行数据回归；选取掘进参数数据库中70％的数据进行训练，30％的数据进行测试，得到一个成熟的回归学习机svm2，获得围岩状态参数预测结果Ysvm2；Step 3: Using the support vector machine to take the surrounding rock state parameter matrix N as input, the steady-state segment excavation parameter matrix M2 as the output, and perform data regression; select 70% of the data in the excavation parameter database for training, and 30% of the data is performed. Test, get a mature regression learning machine svm2, obtain the surrounding rock state parameter prediction result Ysvm2;
    步骤四：利用最小回归的方法将将围岩状态参数矩阵N作为输入，稳态段掘进参数矩阵M2作为输出得到数学模型如下：Step 4: Using the minimum regression method, the surrounding rock state parameter matrix N is taken as the input, and the steady state segment mining parameter matrix M2 is taken as the output to obtain the mathematical model as follows:
    Ft＝464.853-9.012*Jv-80.4*W+0.118*RcFt=464.853-9.012*Jv-80.4*W+0.118*Rc
    Tn＝61.865-7.012*Jv-7.4*W+0.064*RcTn=61.865-7.012*Jv-7.4*W+0.064*Rc
    P＝0.1105+0.12*Jv+2.466*W+0.058*RcP=0.1105+0.12*Jv+2.466*W+0.058*Rc
    n＝11.743+0.16*Jv-1.132*W-0.044*Rcn=11.743+0.16*Jv-1.132*W-0.044*Rc
    V＝p*n；V=p*n;
    获得稳态段掘进状态参数预测结果Yreg2；Obtaining the steady state segment excavation state parameter prediction result Yreg2;
    步骤五：将步骤二-四得到的稳态段掘进状态参数的预测结果Ynet2、Ysvm2和Yreg2，进行数学平均：
    Figure PCTCN2018112521-appb-100006
    获得的智能控制决策模型Y2。     Step 5: Perform the mathematical average of the predicted results of the steady state segment excavation state parameters Ynet2, Ysvm2 and Yreg2 obtained in steps 2-4.
    Figure PCTCN2018112521-appb-100006
    The obtained intelligent control decision model Y2.
  6. 根据权利要求1所述的硬岩TBM掘进控制参数智能决策方法，其特征在于，所述当前预估参数为：当前第k组掘进参数预测得到的围岩参数Nk和稳态掘进参数Mk的平均值；所述累积平均预估参数为：根据岩机互馈模型Y1对围岩状态参数进行预估，得到掘进参数第k-2段到k-1段预测得到的围岩参数Ns；若当前预估参数和累积平均预估参数偏差平均值小于10％，当前围岩处于稳定段，当前掘进参数效果稳定；如果预估参数和累积平均预估参数偏差平均值大于90％，当前掘进参数不稳定，需要调整。The intelligent rock TBM tunneling control parameter intelligent decision making method according to claim 1, wherein the current estimated parameter is: an average of the surrounding rock parameter Nk and the steady state driving parameter Mk predicted by the current k-th group driving parameter prediction. The cumulative average estimation parameter is: estimating the surrounding rock state parameter according to the rock machine mutual feeding model Y1, and obtaining the surrounding rock parameter Ns predicted by the k-2th to the k-1 segments of the driving parameter; The average deviation of the estimated parameters and the cumulative average estimated parameter is less than 10%. The current surrounding rock is in the stable section, and the current tunneling parameters are stable. If the average of the estimated parameters and the cumulative average estimated parameter deviation is greater than 90%, the current tunneling parameters are not Stable and need adjustment.
  7. 根据权利要求1或6所述的硬岩TBM掘进控制参数智能决策方法，其特征在于，所述步骤四中手动调整是TBM主司机根据累积平均预估参数和当前预估参数的差值大小，实时调整当前掘进参数如刀盘转速n和推进速度V，控制其他掘进参数在稳定范围内，保持TBM安全高效掘进；所述步骤四中自动调整是：根据预累积平均预估参数和当前预估参数的差值大小，通过PLC控制器实时调整当前掘进参数如刀盘转速n和推进速度V，控制其他掘进参数在稳定范围内，保持TBM安全高效掘进。The intelligent rock TBM tunneling control parameter intelligent decision making method according to claim 1 or 6, wherein the manual adjustment in the step 4 is a difference between the cumulative average estimated parameter and the current estimated parameter of the TBM master driver. Real-time adjustment of current excavation parameters such as cutter speed n and propulsion speed V, control other excavation parameters within a stable range, and maintain safe and efficient tunneling of TBM; automatic adjustment in step 4 is: based on pre-cumulative average estimation parameters and current estimation The difference value of the parameters is adjusted in real time by the PLC controller such as the cutter head speed n and the propulsion speed V, and the other tunneling parameters are controlled within a stable range to keep the TBM safe and efficient.
  8. 根据权利要求1所述的硬岩TBM掘进控制参数智能决策方法，其特征在于，当工程数据库中的数据增幅大于30％时，由人工或上位机在停机时间或注浆时间对TBM中岩机互馈模型和智能控制决策模型进行自学习和自更新，从而获得新的岩机互馈模型和智能控制决策模型；更新后当前的岩机互馈模型和智能控制决策模型占主导地位，新的岩机互馈模型的预测结果优于当前的岩机互馈模型或智能控制决策模型时，新的岩机互馈模型替换当前的岩机互馈模型，新的智能控制决策模型的预测结果优于当前的智能控制决策模型时，新的智能 控制决策模型替换当前的智能控制决策模型。The intelligent rock TBM tunneling control parameter intelligent decision making method according to claim 1, wherein when the data in the engineering database increases by more than 30%, the rock machine in the TBM is stopped by the manual or the upper computer during the downtime or the grouting time. The self-learning model and the intelligent control decision model are self-learning and self-updating, so as to obtain new rock-machine mutual feedback model and intelligent control decision-making model. After the update, the current rock-machine mutual-feeding model and intelligent control decision-making model dominate, new When the prediction result of the rock-machine mutual feedback model is better than the current rock-machine mutual-feeding model or the intelligent control decision-making model, the new rock-machine mutual-feeding model replaces the current rock-machine mutual-feeding model, and the new intelligent control decision-making model has excellent prediction results. In the current intelligent control decision model, the new intelligent control decision model replaces the current intelligent control decision model.
  9. 一种硬岩TBM掘进控制参数智能决策系统，其特征在于，包括：An intelligent decision system for hard rock TBM tunneling control parameters, characterized in that it comprises:
    工程数据库单元，定期从TBM上位机获取围岩状态数据和掘进参数数据；The engineering database unit regularly acquires surrounding rock state data and tunneling parameter data from the TBM host computer;
    岩机互馈模型单元，利用工程数据库单元的掘进参数对当前掘进环境的围岩状态参数进行预估；The rock machine mutual feed model unit estimates the surrounding rock state parameters of the current excavation environment by using the excavation parameters of the engineering database unit;
    智能控制决策模型单元，根据获得的围岩状态参数对最佳掘进控制参数进行预估；The intelligent control decision model unit estimates the optimal tunneling control parameters according to the obtained surrounding rock state parameters;
    模型参数输出实时显示模块，通过I/O接口与上位机进行通讯，将输出参数在主司机操作界面显示；The model parameter outputs a real-time display module, communicates with the host computer through the I/O interface, and displays the output parameters on the main driver operation interface;
    自动/手动掘进参数控制模块，设置在上位机操作界面上，由主司机或系统依据预估的掘进参数，通过PLC控制器控制TBM掘进参数进行调整；The automatic/manual excavation parameter control module is set on the upper computer operation interface, and the main driver or the system controls the TBM excavation parameters through the PLC controller according to the estimated excavation parameters;
    模型自学习自更新单元，设置在上位机操作界面上，由主司机设置模型更新时间或者由后台选择恰当时间对模型进行自动更新，恰当时间包含但不限于停机时间和注浆时间。The model self-learning self-updating unit is set on the upper computer operation interface, and the model is updated by the main driver to set the model update time or the appropriate time is selected by the background. The appropriate time includes but is not limited to the downtime and the grouting time.
  10. 根据权利要求9所述的硬岩TBM掘进控制参数智能决策系统，其特征在于，所述岩机互馈模型单元和智能控制决策模型单元的参数可以被读取和写入，当模型自学习自更新单元启动后，岩机互馈模型单元和智能控制决策模型的参数可以被写入，其他过程只能被读取，不能写入。The intelligent rock TBM tunneling control parameter intelligent decision system according to claim 9, wherein the parameters of the rock machine mutual feed model unit and the intelligent control decision model unit can be read and written, and the model self-learning After the update unit is started, the parameters of the rock machine mutual feedback model unit and the intelligent control decision model can be written, and other processes can only be read and cannot be written.
  11. 根据权利要求9所述的硬岩TBM掘进控制参数智能决策系统，其特征在于，所述工程数据库单元从TBM上位机获取的数据包括TBM实时掘进参数和其他TBM的工程数据库，其他TBM的工程数据库包含但不限于同一结构类型或类似地质掘进的TBM的工程数据库。The hard rock TBM tunneling control parameter intelligent decision system according to claim 9, wherein the data acquired by the engineering database unit from the TBM host computer includes TBM real-time heading parameters and other TBM engineering databases, and other TBM engineering databases. Engineering databases including, but not limited to, TBMs of the same structural type or similar geological excavation.
PCT/CN2018/112521 2017-09-30 2018-10-30 Smart decision making method and system for boring control parameters of hard rock tbm WO2019063024A1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
DE112018004244.8T DE112018004244T5 (en) 2017-09-30 2018-10-30 An intelligent decision making process and system for drilling control parameters of hard rock TBM

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710937469.1A CN107632523B (en) 2017-09-30 2017-09-30 A kind of hard rock TBM digging control parameter intelligent decision-making technique and system
CN201710937469.1 2017-09-30

Publications (1)

Publication Number Publication Date
WO2019063024A1 true WO2019063024A1 (en) 2019-04-04

Family

ID=61105259

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/112521 WO2019063024A1 (en) 2017-09-30 2018-10-30 Smart decision making method and system for boring control parameters of hard rock tbm

Country Status (3)

Country Link
CN (1) CN107632523B (en)
DE (1) DE112018004244T5 (en)
WO (1) WO2019063024A1 (en)

Cited By (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111160325A (en) * 2019-11-11 2020-05-15 北京迈格威科技有限公司 Optical fingerprint identification device, preparation method thereof and touch terminal
CN112855113A (en) * 2021-01-28 2021-05-28 北京三一智造科技有限公司 Automatic drilling method and controller of rotary drilling rig, storage medium and electronic equipment
CN113586028A (en) * 2021-07-21 2021-11-02 宁夏天地奔牛实业集团有限公司 Intelligent monitoring system of raise boring machine reaming cutter head based on digital twinning
CN113762360A (en) * 2021-08-20 2021-12-07 同济大学 Surrounding rock grade prediction method in TBM tunneling process based on SMOTE + ADACOST algorithm
CN113807004A (en) * 2021-06-30 2021-12-17 北京交通大学 Tool life prediction method, device and system based on data mining
CN113868091A (en) * 2021-10-09 2021-12-31 中国铁建重工集团股份有限公司 Data monitoring system and method for tunneling machine group
CN113958369A (en) * 2021-11-10 2022-01-21 重庆科技学院 Tunnel lining structure health monitoring method and system based on digital twinning
CN114320316A (en) * 2022-01-14 2022-04-12 重庆邮电大学 Shield tunneling machine construction early warning method and device
CN114482885A (en) * 2022-01-25 2022-05-13 西南石油大学 Pressure-controlled drilling intelligent control system
CN115268272A (en) * 2022-08-11 2022-11-01 北京交通大学 TBM control parameter decision method and device based on tunneling load prediction
CN115618222A (en) * 2022-06-21 2023-01-17 北京交通大学 Prediction method of tunneling response parameters
CN115628930A (en) * 2022-12-16 2023-01-20 太原理工大学 Method for predicting underground cutting working condition of heading machine based on RBF neural network
CN116485225A (en) * 2023-03-15 2023-07-25 西南交通大学 Automatic acquisition method and system for BQ value of surrounding rock in construction stage based on drilling parameters
CN116562156A (en) * 2023-05-15 2023-08-08 南栖仙策(南京)高新技术有限公司 Training method, device, equipment and storage medium for control decision model
CN117332992A (en) * 2023-11-24 2024-01-02 北京国联视讯信息技术股份有限公司 Collaborative manufacturing method and system for industrial Internet

Families Citing this family (22)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107632523B (en) * 2017-09-30 2019-07-23 中铁工程装备集团有限公司 A kind of hard rock TBM digging control parameter intelligent decision-making technique and system
CN108984817B (en) * 2018-05-08 2022-04-22 中铁工程装备集团有限公司 Real-time TBM cutter abrasion loss evaluation method
CN109358505B (en) * 2018-10-26 2022-03-29 中铁工程装备集团有限公司 TBM intelligent driving method and system
CN109685378B (en) * 2018-12-27 2020-04-24 中铁工程装备集团有限公司 TBM construction surrounding rock digchability grading method based on data mining
CN109376194B (en) * 2018-12-29 2021-01-26 中铁工程装备集团有限公司 Tunnel boring machine remote monitoring platform based on big data
CN109630154B (en) * 2019-01-24 2023-08-25 华能西藏雅鲁藏布江水电开发投资有限公司 Tunneling robot for tunneling and remote mobile terminal command system
CN109812257A (en) * 2019-02-21 2019-05-28 山东大学 Water jet auxiliary rock intelligence control system and method
CN109933577B (en) * 2019-03-08 2020-12-18 山东大学 Tunnel tunneling prediction method and system based on TBM rock-machine parameter dynamic interaction mechanism
CN109918822B (en) * 2019-03-15 2023-12-15 中铁高新工业股份有限公司 Shield tunneling parameter deviation calculation method and system based on random forest regression
CN110109895B (en) * 2019-03-29 2021-05-28 山东大学 Surrounding rock grading combined prediction method suitable for TBM tunneling tunnel and application
CN110020694A (en) * 2019-04-19 2019-07-16 中铁工程装备集团有限公司 A kind of TBM unfavorable geology discrimination method based on intelligent drives model
CN110110419B (en) * 2019-04-28 2022-11-18 中铁工程装备集团有限公司 TBM tunneling parameter prediction method based on multi-target learning
CN110244344A (en) * 2019-06-05 2019-09-17 中南大学 A kind of TBM method for forecasting advanced geology based on deep learning
CN111144635B (en) * 2019-12-20 2022-07-12 山东大学 TBM operation parameter decision method and system based on deep learning
CN111197500B (en) * 2020-01-06 2021-04-13 山东大学 Method and system suitable for real-time judgment of surrounding rock grade of TBM tunneling tunnel
CN110989366B (en) * 2020-03-04 2020-07-03 中国铁建重工集团股份有限公司 Control method of heading machine, heading machine and computer readable storage medium
CN111622763A (en) * 2020-04-23 2020-09-04 中国铁建重工集团股份有限公司 Direction adjusting control method and system for vertical shaft heading machine, vertical shaft heading machine and storage medium
CN111594198B (en) * 2020-04-30 2021-03-30 东北大学 Division method for tunneling cycle of open TBM construction data
CN111594201B (en) * 2020-05-28 2021-08-13 中铁工程装备集团有限公司 TBM key parameter intelligent control system and method
CN111648782A (en) * 2020-06-16 2020-09-11 中铁十四局集团隧道工程有限公司 Method for identifying stratum and adjusting tunneling parameters according to TBM (tunnel boring machine) self-vibration information
CN112183993B (en) * 2020-09-22 2022-06-17 中国科学院武汉岩土力学研究所 TBM (tunnel boring machine) tunneling control parameter intelligent prediction and optimization decision method
CN112664174B (en) * 2020-12-21 2022-05-20 中铁四局集团第五工程有限公司 Tunnel surrounding rock grade determination method and system based on multiple drill holes

Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103870677A (en) * 2014-02-07 2014-06-18 上海交通大学 Setting method for tunneling parameters of tunneling machine
CN104217124A (en) * 2014-09-15 2014-12-17 天津大学 TBM (Tunnel Boring Machine) construction surrounding rock classification method depending on engineering sample data
CN105631150A (en) * 2016-01-05 2016-06-01 石家庄铁道大学 Optimization method of shield excavation parameters under condition of compound stratum
CN105938611A (en) * 2016-07-06 2016-09-14 山东大学 Method for fast grading underground engineering surrounding rock in real time based on parameters while drilling
CN106383931A (en) * 2016-08-31 2017-02-08 广州市市政集团有限公司 Effective thrust and torque calculation method of tunnel boring machine cutterhead on the basis of CSM (Colorado School of Mines) model
JP2017066652A (en) * 2015-09-29 2017-04-06 株式会社大林組 Excavation monitoring device and excavation state determination method
CN106570275A (en) * 2016-11-07 2017-04-19 沈阳工业大学 Method based on CAI value for predicting abrasion of TBM hobbing cutter
CN106778010A (en) * 2016-12-29 2017-05-31 中铁十八局集团隧道工程有限公司 TBM cutter life Forecasting Methodologies based on data-driven support vector regression
CN107201903A (en) * 2017-04-07 2017-09-26 北京工业大学 The Intelligentized control method and system of TBM constructing tunnels
CN107248026A (en) * 2017-05-23 2017-10-13 南京工业大学 The quantitative approach of shield driving parameter is predicted using equivalent rock mass basic quality's index
CN107632523A (en) * 2017-09-30 2018-01-26 中铁工程装备集团有限公司 A kind of hard rock TBM digging control parameter intelligent decision-making techniques and system

Patent Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103870677A (en) * 2014-02-07 2014-06-18 上海交通大学 Setting method for tunneling parameters of tunneling machine
CN104217124A (en) * 2014-09-15 2014-12-17 天津大学 TBM (Tunnel Boring Machine) construction surrounding rock classification method depending on engineering sample data
JP2017066652A (en) * 2015-09-29 2017-04-06 株式会社大林組 Excavation monitoring device and excavation state determination method
CN105631150A (en) * 2016-01-05 2016-06-01 石家庄铁道大学 Optimization method of shield excavation parameters under condition of compound stratum
CN105938611A (en) * 2016-07-06 2016-09-14 山东大学 Method for fast grading underground engineering surrounding rock in real time based on parameters while drilling
CN106383931A (en) * 2016-08-31 2017-02-08 广州市市政集团有限公司 Effective thrust and torque calculation method of tunnel boring machine cutterhead on the basis of CSM (Colorado School of Mines) model
CN106570275A (en) * 2016-11-07 2017-04-19 沈阳工业大学 Method based on CAI value for predicting abrasion of TBM hobbing cutter
CN106778010A (en) * 2016-12-29 2017-05-31 中铁十八局集团隧道工程有限公司 TBM cutter life Forecasting Methodologies based on data-driven support vector regression
CN107201903A (en) * 2017-04-07 2017-09-26 北京工业大学 The Intelligentized control method and system of TBM constructing tunnels
CN107248026A (en) * 2017-05-23 2017-10-13 南京工业大学 The quantitative approach of shield driving parameter is predicted using equivalent rock mass basic quality's index
CN107632523A (en) * 2017-09-30 2018-01-26 中铁工程装备集团有限公司 A kind of hard rock TBM digging control parameter intelligent decision-making techniques and system

Cited By (24)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111160325A (en) * 2019-11-11 2020-05-15 北京迈格威科技有限公司 Optical fingerprint identification device, preparation method thereof and touch terminal
CN112855113A (en) * 2021-01-28 2021-05-28 北京三一智造科技有限公司 Automatic drilling method and controller of rotary drilling rig, storage medium and electronic equipment
CN113807004A (en) * 2021-06-30 2021-12-17 北京交通大学 Tool life prediction method, device and system based on data mining
CN113586028B (en) * 2021-07-21 2024-03-29 太原理工大学 Intelligent monitoring system of counter bore cutter head of anti-well drilling machine based on digital twin
CN113586028A (en) * 2021-07-21 2021-11-02 宁夏天地奔牛实业集团有限公司 Intelligent monitoring system of raise boring machine reaming cutter head based on digital twinning
CN113762360B (en) * 2021-08-20 2024-03-29 同济大学 Surrounding rock grade prediction method in TBM tunneling process based on SMOTE+ADACCOST algorithm
CN113762360A (en) * 2021-08-20 2021-12-07 同济大学 Surrounding rock grade prediction method in TBM tunneling process based on SMOTE + ADACOST algorithm
CN113868091A (en) * 2021-10-09 2021-12-31 中国铁建重工集团股份有限公司 Data monitoring system and method for tunneling machine group
CN113868091B (en) * 2021-10-09 2024-02-13 中国铁建重工集团股份有限公司 Data monitoring system and method for development machine group
CN113958369A (en) * 2021-11-10 2022-01-21 重庆科技学院 Tunnel lining structure health monitoring method and system based on digital twinning
CN113958369B (en) * 2021-11-10 2023-10-20 重庆科技学院 Tunnel lining structure health monitoring method and system based on digital twinning
CN114320316A (en) * 2022-01-14 2022-04-12 重庆邮电大学 Shield tunneling machine construction early warning method and device
CN114482885A (en) * 2022-01-25 2022-05-13 西南石油大学 Pressure-controlled drilling intelligent control system
CN114482885B (en) * 2022-01-25 2024-03-29 西南石油大学 Intelligent control system for pressure-controlled drilling
CN115618222B (en) * 2022-06-21 2023-05-05 北京交通大学 Tunnel tunneling response parameter prediction method
CN115618222A (en) * 2022-06-21 2023-01-17 北京交通大学 Prediction method of tunneling response parameters
CN115268272A (en) * 2022-08-11 2022-11-01 北京交通大学 TBM control parameter decision method and device based on tunneling load prediction
CN115628930A (en) * 2022-12-16 2023-01-20 太原理工大学 Method for predicting underground cutting working condition of heading machine based on RBF neural network
CN116485225B (en) * 2023-03-15 2023-11-10 西南交通大学 Automatic acquisition method and system for BQ value of surrounding rock in construction stage based on drilling parameters
CN116485225A (en) * 2023-03-15 2023-07-25 西南交通大学 Automatic acquisition method and system for BQ value of surrounding rock in construction stage based on drilling parameters
CN116562156A (en) * 2023-05-15 2023-08-08 南栖仙策(南京)高新技术有限公司 Training method, device, equipment and storage medium for control decision model
CN116562156B (en) * 2023-05-15 2024-02-06 南栖仙策(南京)高新技术有限公司 Training method, device, equipment and storage medium for control decision model
CN117332992A (en) * 2023-11-24 2024-01-02 北京国联视讯信息技术股份有限公司 Collaborative manufacturing method and system for industrial Internet
CN117332992B (en) * 2023-11-24 2024-02-09 北京国联视讯信息技术股份有限公司 Collaborative manufacturing method and system for industrial Internet

Also Published As

Publication number Publication date
CN107632523B (en) 2019-07-23
DE112018004244T5 (en) 2020-05-07
CN107632523A (en) 2018-01-26

Similar Documents

Publication Publication Date Title
WO2019063024A1 (en) Smart decision making method and system for boring control parameters of hard rock tbm
CN109685378B (en) TBM construction surrounding rock digchability grading method based on data mining
CN109630154B (en) Tunneling robot for tunneling and remote mobile terminal command system
WO2020181923A1 (en) Tunnel tunneling prediction method and system based on tbm rock-machine parameter dynamic interaction mechanism
CN103147756B (en) A kind of heading machine memory cutting control system and method thereof
CA2965289C (en) Real-time control of drilling fluid properties using predictive models
AU2017254917B2 (en) Drilling engineering analysis roadmap builder
CN105917074A (en) Managing wellbore operations using uncertainty calculations
CN108694258A (en) Drilling well underground dummy emulation method and system for arrangement and method for construction preview optimization
CN110500081A (en) A kind of automatic well drill method based on deep learning
CN113090188B (en) Intelligent drilling real-time borehole trajectory measurement method and system based on machine learning
US11321506B2 (en) Fast algorithm to simulate the response of PDC bits
CN113361043B (en) Method and system for predicting specific energy of cutter head of hard rock tunnel boring machine
CN114427430B (en) Multi-well real-time collaborative drilling parameter optimization method and system
CN106971422B (en) Method for compiling stability analysis model of arch dam abutment based on CATIA
GB2587101A (en) Adjusting well tool operation to manipulate the rate-of-penetration (ROP) of a drill bit based on multiple ROP projections
US11112514B2 (en) Systems and methods for computed resource hydrocarbon reservoir simulation and development
Liu et al. A reinforcement learning based 3d guided drilling method: Beyond ground control
CN111598366A (en) Real-time drilling auxiliary decision-making method and system
CN111852466B (en) Method for shale gas well scale production allocation and pipe network operation optimization
CN114764844B (en) Shale gas drilling optimization design method and device
US11859485B1 (en) Trip map for adjusting a tripping operation in a wellbore
Dong Development and Application of Borehole Data Processing Software for Measurement While Drilling in Coal Mine
Liu et al. Application and Progress of Artificial Intelligence in Oilfield Drilling
CN117167035A (en) Control method and device of heading machine, storage medium and program product

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18860845

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 18860845

Country of ref document: EP

Kind code of ref document: A1