A TBM cutterhead rotating speed adaptive control method

Abstract

The application relates to the technical field of full-face tunnel boring machines, in particular to a TBM cutterhead rotating speed self-adaptive control method. The method adjusts the rotating speed of the cutterhead according to real-time parameters by establishing a relevant database and a random forest algorithm optimization model, realizes self-adaptive control of the rotating speed of the cutterhead, and can greatly improve the efficiency of advancing construction and the operation accuracy of the TBM.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of full-face tunnel boring machine, and particularly relates to a TBM cutterhead rotating speed adaptive control method. BACKGROUND

[0002] With the transportation infrastructure projects moving to the west, it is necessary to build deep mountain tunnel, and in this kind of environment with high excavation difficulty and complex geological conditions, full-face tunnel boring machine (TBM) is usually used as the main construction method.

[0003] In the process of TBM tunneling, its tunneling efficiency is affected by multiple factors, and the most important factors include geological parameters during tunneling and mechanical parameters of the TBM itself. The geological related parameters mainly include: strength, brittleness, wear resistance of rock, integrity of surrounding rock mass, etc. The mechanical parameters of the TBM itself include TBM cutterhead rotating speed, cutterhead torque, thrust, tunneling speed and penetration, etc. The traditional TBM cutterhead rotating speed adopts a manual constant speed method, which has certain limitations in different geological environments, has relatively low efficiency and is easy to damage the machine body. Therefore, a more advanced way is needed to control the TBM cutterhead rotating speed in actual use. SUMMARY

[0004] The purpose of the present application is to provide a TBM cutterhead rotating speed adaptive control method, which can greatly improve the efficiency of the advancing construction and the accuracy of the TBM operation by establishing a related database and a mathematical model to adjust the rotating speed of the cutterhead according to real-time parameters and realize adaptive control of the cutterhead rotating speed.

[0005] The above purpose of the present application is achieved by the following technical scheme:

[0006] A TBM cutterhead rotating speed adaptive control method, comprising the following steps:

[0007] S1: Establishing a geological database

[0008] Reading the TBM machine itself real-time geological information module through the PLC controller, obtaining the TBM tunneling geological information of the previous 1000 rings, and establishing a geological information database with ring number as the primary key;

[0009] S2: Establishing a TBM tunneling parameter database

[0010] Reading the TBM tunneling parameters of the previous 1000 rings through the PLC, and establishing a tunneling parameter database with ring number as the foreign key;

[0011] S3: Data processing

[0012] The geological information database and the tunneling parameter database are associated by taking the ring number as a primary key and a foreign key to form a new TBM tunneling geological and tunneling parameter database, and then the database is subjected to missing value processing, abnormal value processing and data screening;

[0013] S4: Establishing a random forest optimization parameter model

[0014] The random forest algorithm is used to model the parameters affecting the cutterhead speed in the above-processed data, to determine a cutterhead speed control scheme, so as to achieve the best advancing efficiency and working state, and to calculate the speed range of the cutterhead;

[0015] S5: Real-time updating of parameters

[0016] The geological parameters are updated in real time by using the PLC to read the geological conditions encountered by the TBM detected by the sensors in real time;

[0017] S6: Performing cutterhead speed adaptive control

[0018] According to the real-time updated geological parameters and the established random forest optimization parameter model, the speed of the TBM cutterhead is adjusted, and is written into the TBM through the PLC system, and at the same time, a warning value is set to ensure that the cutterhead speed is controlled within the speed range determined in step S4, so as to realize cutterhead speed adaptive control;

[0019] S7: Repeating steps S5 and S6 until the advancing distance is reached.

[0020] Preferably, the geological information in step S1 includes rock type, hardness and density, and the tunneling parameters in step S2 include cutterhead speed, cutterhead torque, advancing force, tunneling speed and penetration.

[0021] Preferably, the establishment of the random forest algorithm in step S4 is to select the data of the first 800 rings from the tunneling geological and tunneling parameter database, use the BootStrap method to randomly extract 800 ring data sets to form a training set D, and then extract features from the parameters in the tunneling geological and tunneling parameter database. If there are M parameters in the database, m features are randomly extracted, wherein A single decision tree is constructed by using the information gain strategy to split the nodes, and the formula is as follows:

[0022] g(D,A) = H(D) - H(D|A)

[0023] Wherein, A is a certain feature affecting the cutterhead speed, g(D,A) is the information gain of feature A to the training set D, H(D) is the empirical entropy of set D, and H(D|A) is the empirical conditional entropy of D under the given condition of feature A;

[0024] The above process is repeated to construct the entire random forest, and the prediction of the cutterhead rotating speed can be regarded as a classification problem, and the final formula of the random forest algorithm for the classification problem is:

[0025]

[0026] Wherein, c tj represents the prediction value of the jth leaf node of the th tree, I(x tj ) represents an indicator function, and f t (x) is an information gain function.

[0027] Therefore, the application has the beneficial technical effects:

[0028] The TBM sensor data is read based on the PLC, the geological information and tunneling parameter database is established, the cutterhead rotating speed is adjusted in real time according to the real-time geological conditions, historical parameters and random forest algorithm parameter model, the self-adaptive control of the cutterhead rotating speed is realized, the manpower is saved to a certain extent, and the safety and reliability of the tunnel excavation operation are greatly improved. BRIEF DESCRIPTION OF DRAWINGS

[0029] Figure 1 is the overall flowchart of the application. DETAILED DESCRIPTION

[0030] The application will be further described in detail below with reference to the accompanying drawings.

[0031] Reference Figure 1 A TBM cutterhead rotating speed self-adaptive control method is disclosed, which comprises the following steps: firstly, establishing a geological database, reading the real-time geological information module in the TBM upper computer through the PLC, obtaining the geological data of the first 1000 rings, including rock types, strength of surrounding rock, brittleness, wear resistance, integrity of surrounding rock mass and other geological parameters, and storing the data into a computer connected with the PLC to establish a geological information database with ring number as the primary key.

[0032] Then, the historical tunneling parameter data of the first 1000 rings stored in the TBM upper computer is read through the PLC, including cutterhead rotating speed, cutterhead torque, thrust, tunneling speed and penetration, and the data is stored into a computer connected with the PLC to establish a tunneling parameter database. Then, the geological database and the tunneling parameter database are associated with ring number as the primary key and the foreign key to form a new TBM tunneling geological and tunneling parameter database. The data in the tunneling geological and tunneling parameter database is processed for missing value, abnormal value, data conversion and data screening.

[0033] The random forest optimization parameter model is established. The data of the first 800 rings are selected from the tunneling geological and tunneling parameter database, and the BootStrap method is used to randomly extract 800 ring data sets to form a training set D. Then, the parameters in the tunneling geological and tunneling parameter database are extracted. If there are M parameters in the database, m features are randomly extracted, wherein A single decision tree is constructed by using the information gain strategy to split the node, and the formula is as follows:

[0034] g(D,A)=H(D)-H(D|A)

[0035] Wherein, A is a feature that affects the speed of the cutter head, g(D,A) is the information gain of feature A to the training set D, H(D) is the empirical entropy of set D, and H(D|A) is the empirical conditional entropy of D under the condition of feature A. The greater the information gain is, the greater the "purity improvement" obtained by using feature A to divide is. Pruning is performed by setting an information gain threshold, so that the decision tree correctly divides the training data.

[0036] The above process is repeated to construct the entire random forest. Here, the prediction of the speed of the cutter head can be regarded as a classification problem. For the classification problem, the final formula of the random forest algorithm is:

[0037]

[0038] Wherein, c tj represents the prediction value of the jth leaf node of the th tree, I(x∈R tj ) represents an indicator function, f t (x) is the information gain function, and the final prediction result is the average of all decision tree prediction results.

[0039] The random forest algorithm is used to model the parameters affecting the speed of the cutter head in the above processed data, to determine the cutter head speed control scheme, to achieve the best propulsion efficiency and working state, and to calculate the speed range of the cutter head.

[0040] During the tunneling process of the TBM, the influence of the change of the geological condition on the control parameters of the TBM is very important, therefore, the geological condition encountered by the TBM can be read by the PLC in real time through sensors and other devices, and the geological parameters are updated in real time. That is, the geological condition parameters detected by the sensors and other devices of the TBM are read by the PLC, including the TBM cutterhead speed, cutterhead torque, thrust, tunneling speed and penetration, etc., and are written into the geological information database in real time. Then the geological and tunneling parameter database is updated in real time, the trained random forest model is used to predict the cutterhead speed to be selected by the current TBM in real time. The accurately predicted and real-time changing cutterhead speed is written into the TBM system through the PLC system, and the cutterhead speed of the TBM is controlled in real time and effectively. At the same time, the warning value is set, and once the abnormality or the potential danger to the TBM tunneling occurs, the PLC system and the TBM protection system will immediately respond, and the cutterhead speed is adjusted in time and effectively or the cutterhead is directly stopped. The parameter information is read continuously until the tunneling process is reached.

[0041] Obviously, the above embodiments are only examples for clearly illustrating the present application, and are not intended to limit the embodiments of the present application. Based on the above description, other different forms of changes or variations can also be made by those of ordinary skill in the art. All the embodiments do not need to be exhausted here, and the obvious changes or variations derived from the spirit of the present application are still within the protection scope of the present application.

Claims

1. A TBM cutterhead speed adaptive control method, characterized in that: Comprise the following steps: S1: Establish a geological database Read the TBM machine itself real-time geological information module through the PLC controller, obtain the TBM tunneling geological information of the previous 1000 rings, and establish a geological information database with ring number as the primary key; S2: Establish a TBM tunneling parameter database Read the TBM tunneling parameters of the previous 1000 rings through the PLC, and establish a tunneling parameter database with ring number as the foreign key; S3: Data processing Associate the geological information database and the tunneling parameter database with ring number as the primary key and the foreign key to form a new TBM tunneling geological and tunneling parameter database, and then perform missing value processing, outlier processing and data filtering on the database; S4: Establish a random forest optimization parameter model Use the random forest algorithm to model the parameters affecting the cutterhead speed in the above processed data, determine the cutterhead speed control scheme to achieve the best propulsion efficiency and working state, and calculate the cutterhead speed range; Wherein, the establishment of random forest algorithm is from the tunneling geological and tunneling parameter database to select the data of the first 800 rings, using BootStrap method, that is, randomly drawing 800 ring data to form a training set D, then extracting features from the parameters in the tunneling geological and tunneling parameter database, if there are M parameters in the database, randomly extract m features, wherein A single decision tree is constructed by using information gain strategy for node splitting, and the formula is as follows: g(D,A)=H(D)-H(D|A) Where A is a certain feature affecting the cutterhead speed, g(D,A) is the information gain of feature A on the training set D, H(D) is the empirical entropy of set D, and H(D|A) is the empirical conditional entropy of D under the given condition of feature A; Repeat the above process to build the entire random forest, where the prediction of the cutterhead speed can be regarded as a classification problem. The final formula of the random forest algorithm for classification problems is: where c tj represents the prediction value of the jth leaf node of the tth tree, I(x∈R tj ) represents an indicator function, and f t (x) is an information gain function. S5: Real-time update parameters Use the PLC to read the geological conditions encountered by the TBM detected by the sensor in real time to update the geological parameters in real time; S6: Perform cutterhead speed adaptive control Adjust the cutterhead speed of the TBM according to the real-time updated geological parameters and the established random forest optimization parameter model, and write the cutterhead speed into the TBM through the PLC system. At the same time, set an early warning value to ensure that the cutterhead speed is controlled within the speed range determined in step S4, realizing cutterhead speed adaptive control; S7: Repeat steps S5 and S6 until the tunneling distance is reached.

2. The TBM cutterhead speed adaptive control method of claim 1, wherein: The geological information in step S1 includes rock type, hardness and density, and the tunneling parameters in step S2 include cutterhead speed, cutterhead torque, propulsion force, tunneling speed and penetration.

Images