CN113404502B - Device and method for wear monitoring of shield hob based on topography of ballast - Google Patents

Abstract

The invention provides a shield hob abrasion monitoring device and method based on ballast piece morphology, which comprises the following steps: step 1, acquiring a known shield engineering case in a computer and constructing a database; step 2, analyzing the mapping relation between hob structure parameters, engineering geological parameters, shield tunneling parameters, slag sheet morphology parameters and temperature parameters in a database and the abrasion loss of the hob by adopting a GRNN neural network to obtain a mapping relation analysis result; and 3, constructing a hob abrasion prediction system based on the database, the GRNN neural network and the mapping relation analysis result and setting a hob abrasion alarm threshold. According to the method, the abrasion state of the hob can be predicted in real time through actual engineering geological parameters, hob structural parameters, real-time tunneling parameters and slag parameters, an alarm function is provided when the predicted abrasion state of the hob reaches a set threshold value, and accurate prediction of hob abrasion can be improved through closed-loop feedback adaptive adjustment according to different projects.

Description

Device and method for wear monitoring of shield hob based on topography of ballast

technical field

The invention relates to the technical field of tunnel construction, in particular to a shield hob wear monitoring device and method based on the shape of a ballast piece.

Background technique

With the wide application of tunnel boring machines at home and abroad, shield construction has gradually become a main construction method in tunnel construction. For shield construction, shield hob is the main rock breaking tool. It is suitable for various rock formations, pebble gravel formations and soft and hard composite formations, and is widely used in various roadheaders.

Due to the uncertainty of the rock and the complexity of the stratum during the excavation process, the dynamics of the force change of the tool and the uncertainty of the tool wear form are caused. After a long time of rock cutting, the hob needs to be replaced due to abnormal wear or normal wear to a certain extent, which will inevitably affect the shield tunneling efficiency and construction cost.

Practice has shown that the shape and size of rock ballast and its particle size distribution are indirect feedbacks of complex surrounding rock conditions and mechanical driving performance, and the tool wear state can be predicted through ballast sheet parameters and shield operating parameters. Currently. Judging whether it is necessary to enter the warehouse to change the tool is only based on the experience of the construction personnel. The accuracy of the judgment varies from person to person, and it is likely to cause major accidents due to misjudgment. Therefore, it is necessary to provide a method for monitoring the wear of shield hob based on the topography of the ballast piece.

SUMMARY OF THE INVENTION

The invention provides a shield hob wear monitoring device and method based on the topography of the ballast piece, the purpose of which is to solve the problem that the traditional construction method cannot accurately and accurately predict the wear amount of the current hob, and cannot meet the requirements of intelligent, safe and efficient shield tunneling. The problem of excavation needs.

In order to achieve the above object, an embodiment of the present invention provides a shield hob wear monitoring device based on the topography of the ballast piece, including:

a data collection box, the data collection box is arranged on one side of the conveyor belt of the shield machine;

Industrial computer, the industrial computer is arranged on one side of the data acquisition box, the first end of the industrial computer is electrically connected to the first end of the data acquisition box, and the second end of the industrial computer is electrically connected to the computer ;

a truss, the first end of the truss is erected on the conveyor belt of the shield machine, and the second end of the truss is erected on the ground;

a lighting device, there are two lighting devices, the two lighting devices are symmetrically arranged on both ends of the truss, and both the lighting devices are detachably connected to the truss;

a camera, the camera is arranged on the top of the truss, the camera is detachably connected to the truss, and the camera is electrically connected to the second end of the data acquisition box;

Infrared thermometer, the infrared thermometer is arranged at the top of the truss, the infrared thermometer is detachably connected to the truss, and the infrared thermometer is electrically connected to the third end of the data acquisition box. connect.

An embodiment of the present invention also provides a method for monitoring the wear of a shield hob based on the topography of a ballast piece, including:

Step 1: Obtain known shield engineering cases in the computer and build a database;

Step 2, using the GRNN neural network to analyze the mapping relationship between the hob structure parameters, engineering geological parameters, shield tunneling parameters, ballast piece topography parameters and temperature parameters and the hob wear amount in the database, and obtain the mapping relationship analysis result;

Step 3, build a hob wear prediction system based on the database, the GRNN neural network and the analysis results of the mapping relationship, and set an alarm threshold for the hob wear amount;

Step 4, collect the hob structure parameters and engineering geological parameters of the current shield machine before construction, and input the hob wear prediction system;

Step 5, collecting the shield tunneling parameters, ballast piece topography parameters and temperature parameters during the construction of the current shield machine in real time and inputting them into the hob wear prediction system;

Step 6, the hob wear prediction system predicts the hob wear amount according to the input hob structure parameters, engineering geological parameters, shield tunneling parameters, ballast piece shape parameters and temperature parameters, and sets the predicted hob wear amount and setting. Compare with the alarm threshold of hob wear amount;

Step 7: When the predicted hob wear exceeds the hob wear alarm threshold, the hob wear prediction system will issue an alarm, the shield machine will stop running, replace the hob, measure the wear of the worn hob and input the hob wear forecasting system;

Step 8, adopt the particle swarm optimization algorithm to adaptively optimize the GRNN neural network according to the wear amount of the worn hob and the predicted wear amount of the hob;

Step 9, restart the shield machine after the tool change is completed, jump to step 4 and continue to execute until the entire shield project is completed, and the monitoring ends;

Step 10, when the predicted hob wear amount does not exceed the hob wear amount alarm threshold, jump to step 5 and continue to execute until the entire shield tunneling work is completed, ending the monitoring.

Wherein, the step 1 specifically includes:

The shield engineering case includes a number of tool change records. The tool change records include hob structural parameters, engineering geological parameters, shield tunneling parameters, ballast shape parameters, temperature parameters and hob wear. Among them, the hob structural parameters include: Blade structure parameters, hob radius, cutter spacing and load weight, engineering geological parameters include rock mechanics parameters, rock material parameters, rock joint and fault parameters and lithology parameters, shield tunneling parameters include shield machine penetration, Shield machine cutter head speed, shield machine thrust, shield machine working torque, shield machine soil bin pressure and shield machine propulsion speed, ballast piece morphology parameters including ballast piece particle size distribution index, ballast piece length and short axis ratio index and ballast sheet texture index, and temperature parameters include ballast sheet temperature.

Wherein, the step 2 specifically includes:

The penetration, propulsion speed, cutter head rotation speed, thrust of the shield machine, working torque of the shield machine, soil pressure of the shield machine, and radius of the hob are used as the 7 neurons of the input layer of the GRNN neural network. , the tool wear amount is used as the neuron of the output layer to form a GRNN neural network;

Divide the data in the database into 100 groups, use random sampling to select 10 groups of data from the 100 groups of data as the test set, and the remaining 90 groups of data as the training set;

The training set is randomly divided into 9 units, each unit includes 10 sets of data, 8 units are randomly selected from the 9 units as the input sample of the training set by the cross-validation method, and the remaining 1 unit is used as the output sample of the training set. The input sample data of the training set is normalized to between [-1, 1], and the verification search is performed with a step size of 0.01 in (0, 1], and the smooth factor σ that minimizes the mean square error between the predicted value and the sample value is found, and Record the best input sample and the best output sample corresponding to the current smooth factor;

The test set data is normalized, and the obtained smooth factor σ, the best input sample and the best output sample are used as input variables to build a 4-layer GRNN neural network, and the output layer outputs the tool wear amount.

Wherein, the step 3 specifically includes:

Based on the GRNN neural network, the mapping relationship between the engineering geological parameters, shield tunneling parameters, shield hob structure, ballast shape size and temperature parameters and the tool wear amount in the database is collected through correlation analysis, and the hob wear prediction system is established:

δ _i =β ₀ p+β ₁ v+β ₂ n+β ₃ F+β ₄ T+β ₅ S+β ₆ L+...+β _m X _m +C (1)

Among them, β ₀ , β ₁ , β ₂ , β ₃ , β ₄ , β _m are the parameters to be estimated, δ _i is the tool wear amount, p is the penetration degree of the shield machine, v is the propulsion speed, n Indicates the rotation speed of the cutter head, F represents the thrust of the shield machine, T represents the working torque of the shield machine, L represents the radius of the hob, S represents the shape and size of the ballast piece, X _m represents other parameters related to the amount of tool wear, and C represents to be determined constant.

Wherein, the step 5 specifically includes:

The main control room of the shield machine is used to collect the current shield tunneling parameters during the construction of the shield machine in real time and input them into the hob wear prediction system. The shape and size of the ballast piece are calculated, and the shape parameters of the ballast piece are obtained and input into the hob wear prediction system.

Wherein, the step 8 specifically includes:

The particle swarm optimization algorithm is used to optimize the CRNN neural network: set the particle swarm calculation parameters, and set the mean square error of the tool wear amount predicted when the hob wear prediction system sends an alarm and the wear amount of the worn hob as the fitness function. The samples and examples are brought into the GRNN neural network, the fitness value F _i is calculated, the fitness values of all positions passed by the i-th particle are compared, and the optimal position P _bi is determined, and the optimal position P _bi of all particles is compared. The fitness value, determine the optimal position G _b of the whole population, adjust the speed and position of the particle according to the position of each particle itself and the optimal particle position, when the iteration termination condition is reached, obtain the optimal position G _b , use the searched optimal position G b . The optimal location G _b optimizes the GRNN neural network.

The above-mentioned scheme of the present invention has the following beneficial effects:

The device and method for monitoring the wear of shield hob based on the topography of the ballast piece described in the above-mentioned embodiments of the present invention can predict the rolling cutter in real time according to the actual engineering geological parameters, the structural parameters of the hob cutter and the real-time excavation parameters and the parameters of the ballast piece. The wear state of the cutter provides an alarm function when the predicted wear state of the hob reaches the set threshold, and can improve the accurate prediction of the hob wear through closed-loop feedback adaptive adjustment according to different projects.

Description of drawings

Fig. 1 is the flow chart of the present invention;

Fig. 2 is the structural representation of the present invention;

3 is a schematic diagram of a CRNN neural network of the present invention;

FIG. 4 is a flowchart of the particle swarm optimization algorithm of the present invention.

[Description of reference numerals]

1-shield conveyor belt; 2-data acquisition box; 3-industrial computer; 4-truss; 5-lighting device; 6-camera; 7-infrared thermometer.

Detailed ways

In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, the following will be described in detail with reference to the accompanying drawings and specific embodiments.

Aiming at the problems that the existing construction method cannot accurately and accurately predict the wear amount of the current hob, and cannot meet the requirements of the shield machine for intelligent, safe and efficient excavation, the invention provides a shield hob wear monitoring device based on the shape of the ballast piece. and methods.

As shown in FIG. 1 to FIG. 4 , an embodiment of the present invention provides a shield hob wear monitoring device based on the topography of the ballast piece, including: a data collection box 2 , and the data collection box 2 is arranged in the shield machine One side of the conveyor belt 1; the industrial computer 3, the industrial computer 3 is arranged on one side of the data acquisition box 2, and the first end of the industrial computer 3 is electrically connected to the first end of the data acquisition box 2, The second end of the industrial control is electrically connected to the computer; the truss 4, the first end of the truss 4 is erected on the conveyor belt 1 of the shield machine, and the second end of the truss 4 is erected on the ground; the lighting device 5, the There are two lighting devices 5, the two lighting devices 5 are symmetrically arranged on both ends of the truss 4, and both the lighting devices 5 are detachably connected to the truss 4; the camera 6, the The camera 6 is arranged at the top of the truss 4, the camera 6 is detachably connected to the truss 4, and the camera 6 is electrically connected to the second end of the data acquisition box 2; the infrared thermometer 7, the The infrared thermometer 7 is arranged at the top of the truss 4 , the infrared thermometer 7 is detachably connected to the truss 4 , and the infrared thermometer 7 is electrically connected to the third end of the data acquisition box 2 .

In the device and method for monitoring the wear of a shield hob based on the topography of the ballast sheet according to the above embodiments of the present invention, the truss 4 is used to support the lighting device 5 , the camera 6 and the infrared thermometer 7 . , the lighting device 5 is used to cooperate with the camera 6 to perform image dataization, and the infrared camera 6 is used to measure the temperature of the ballast in real time.

The embodiment of the present invention also provides a method for monitoring the abrasion of shield hob based on the topography of the ballast piece, including: step 1, obtaining known shield tunneling engineering cases in a computer and building a database; step 2, using a GRNN neural network Analyze the mapping relationship between the hob structure parameters, engineering geological parameters, shield tunneling parameters, ballast sheet topography parameters and temperature parameters in the database and the hob wear amount, and obtain the analysis result of the mapping relationship; step 3, based on the database, GRNN neural network Build a hob wear prediction system with the results of the mapping relationship analysis and set the alarm threshold of the hob wear amount; Step 4, collect the hob structural parameters and engineering geological parameters of the current shield machine before construction and input it into the hob wear prediction system; Step 5 , collect in real time the shield tunneling parameters, ballast shape parameters and temperature parameters during the construction of the current shield machine and input them into the hob wear prediction system; step 6, the hob wear prediction system according to the input hob structure parameters, engineering geological parameters parameters, shield tunneling parameters, ballast shape parameters and temperature parameters to predict the hob wear amount and compare the predicted hob wear amount with the set hob wear amount alarm threshold; Step 7, when the predicted hob wear amount is When the wear amount exceeds the alarm threshold of the hob wear amount, the hob wear prediction system will issue an alarm, the shield machine will stop running, replace the hob, measure the wear amount of the worn hob and input it into the hob wear prediction system; Step 8, according to the The wear amount of the worn hob and the predicted wear amount of the hob are adaptively optimized by the particle swarm optimization algorithm to the GRNN neural network; step 9, restart the shield machine after the tool change is completed, and jump to step 4 to continue execution until the entire The shield tunneling work is completed, and the monitoring is ended; in step 10, when the predicted hob wear amount does not exceed the hob wear amount alarm threshold, jump to step 5 to continue execution until the entire shield tunneling work is completed, ending the monitoring.

The device and method for monitoring the wear of shield hob based on the topography of the ballast piece described in the above-mentioned embodiment of the present invention adopts the CRNN neural network to analyze the mapping relationship between the acquisition parameters and the wear amount of the hob, according to the engineering excavation data and the topography of the ballast piece For fluctuations and changes, set the alarm threshold of the hob wear amount, predict the abnormal state characteristics of the hob wear in advance, and accurately identify various failure conditions of the hob, such as normal wear, partial wear, etc. The main process of using neural network for prediction and judgment is as follows: 1 .Select neural network; 2. Select learning samples; 3. Determine input vector and output vector; 4. Set neural network parameters; 5. Neural network learning, build neural network model; 6. Input target object for prediction and judgment; Network structure, GRNN neural network structure includes input layer, pattern layer, summation layer and output layer, using part of existing data to train GRNN neural network structure; the number of neurons in the input layer is the number of input vectors in the learning sample , the input layer outputs multiple input variables to the model layer, the number of neurons in the model layer is equal to the number of input vectors in the learning sample, each neuron in the model layer corresponds to different input vectors, and the neurons in the model layer transmit function as follows:

Among them, Pi represents the output of the _i -th neuron in the pattern layer, X represents the learning sample, X=[X ₁ , X ₂ ,...,X _n ] ^T , X _a represents the a-th input vector in the learning sample, and σ represents Smoothing factor, i represents the ith neuron.

The summation layer uses two types to sum the outputs of the neurons in the pattern layer:

A class of calculations are:

The output of each neuron in the pattern layer is arithmetically summed, the connection weight between the pattern layer and each neuron is 1, and the summation layer neuron transfer function is as follows:

The calculation method of the second category is:

The weighted summation is performed on each neuron in the pattern layer, and the connection weight between the ith neuron in the pattern layer and the jth molecular summation neuron in the summation layer is the _ith in the ith output sample Yi. i elements, the summation layer neuron transfer function is as follows:

The number of neurons in the output layer is equal to the dimension k of the output vector in the learning sample. Each neuron divides the output of the summation layer, and the output of neuron j is as follows:

In the device and method for monitoring the abrasion of shield hob based on the topography of the ballast piece described in the above-mentioned embodiments of the present invention, parameters such as engineering geology, hob structure and cutter head structure are input into the system as known parameters before construction, and the parameters are input during construction. During the process, real-time monitoring of the tunneling parameters of the shield machine and the shape parameters of the ballast piece are input into the system to predict the real-time wear amount of the hob. The worker enters the warehouse to change the tool and measures the current wear status and wear amount of the tool, and inputs the result into the hob wear prediction system to carry out the feedback adaptive adjustment of the hob wear prediction system, so as to maintain the high prediction accuracy of the current project and improve the hob Robustness of wear prediction system predictions, repeating the cycle after each tool change.

Wherein, the step 1 specifically includes: the shield engineering case includes a plurality of tool change records, and the tool change records include hob structural parameters, engineering geological parameters, shield tunneling parameters, ballast blade morphology parameters, temperature parameters, and hob wear. Among them, the hob structure parameters include blade edge structure parameters, hob radius, cutter spacing and load weight, engineering geological parameters include rock mechanics parameters, rock material parameters, rock joints and fault parameters and lithology parameters, shield tunneling parameters Including the penetration of the shield machine, the rotation speed of the cutter head of the shield machine, the thrust of the shield machine, the working torque of the shield machine, the pressure of the soil silo of the shield machine and the propulsion speed of the shield machine, and the shape parameters of the ballast piece include the ballast piece. The diameter distribution index, the ratio of the long and the short axis of the ballast piece and the texture index of the ballast piece, and the temperature parameters include the temperature of the ballast piece.

Wherein, the step 2 specifically includes: using the penetration of the shield machine, the propulsion speed, the rotational speed of the cutter head, the thrust of the shield machine, the working torque of the shield machine, the pressure of the soil bin of the shield machine and the radius of the hob as the GRNN nerve The 7 neurons in the input layer of the network use the tool wear amount as the neurons in the output layer to form a GRNN neural network;

Divide the data in the database into 100 groups, use random sampling to select 10 groups of data from the 100 groups of data as the test set, and the remaining 90 groups of data as the training set;

The training set is randomly divided into 9 units, each unit includes 10 sets of data, 8 units are randomly selected from the 9 units as the input sample of the training set by the cross-validation method, and the remaining 1 unit is used as the output sample of the training set. The input sample data of the training set is normalized to between [-1, 1], and the verification search is performed with a step size of 0.01 in (0, 1], and the smooth factor σ that minimizes the mean square error between the predicted value and the sample value is found, and Record the best input sample and the best output sample corresponding to the current smooth factor;

The test set data is normalized, and the obtained smooth factor σ, the best input sample and the best output sample are used as input variables to build a 4-layer GRNN neural network, and the output layer outputs the tool wear amount.

Wherein, the step 3 specifically includes: collecting the engineering geological parameters, shield tunneling parameters, shield hob structure, ballast topography size and temperature parameters and the mapping relationship between the tool wear amount and the tool wear amount in the database through correlation analysis based on the GRNN neural network , to establish a hob wear prediction system:

δ _i =β ₀ p+β ₁ v+β ₂ n+β ₃ F+β ₄ T+β ₅ S+β ₆ L+...+β _m X _m +C (1)

Among them, β ₀ , β ₁ , β ₂ , β ₃ , β ₄ , β _m are the parameters to be estimated, δ _i is the tool wear amount, p is the penetration degree of the shield machine, v is the propulsion speed, n Indicates the rotation speed of the cutter head, F represents the thrust of the shield machine, T represents the working torque of the shield machine, L represents the radius of the hob, S represents the shape and size of the ballast piece, X _m represents other parameters related to the amount of tool wear, and C represents to be determined constant.

Wherein, the step 5 specifically includes: collecting the current shield tunneling parameters during the construction of the shield tunneling machine in real time through the main control room of the shield tunneling machine and inputting them into the hob wear prediction system; The picture of the ballast chip is input into the computer to calculate the shape and size of the ballast chip, the shape parameters of the ballast chip are obtained and input into the hob wear prediction system, and the temperature of the ballast chip on the conveyor belt 1 of the shield machine is measured in real time through the infrared thermometer 7 and input Hob wear prediction system.

The device and method for monitoring the wear of shield hob based on the topography of the ballast sheet according to the above embodiments of the present invention, the camera 6 , the infrared thermometer 7 , the data collection box 2 and the industrial computer 3 The connection transmits and stores the image data and temperature data to the hob wear prediction system in the computer through the data line, and analyzes the current hob wear condition in real time through the hob wear prediction system.

Wherein, the step 8 specifically includes: using the particle swarm optimization algorithm to optimize the CRNN neural network: setting the particle swarm calculation parameters, setting the tool wear amount predicted when the hob wear prediction system issues an alarm and the wear amount of the worn hob. The mean square error is the fitness function, and the learning samples and examples are brought into the GRNN neural network to calculate the fitness value F _i , compare the fitness values of all positions passed by the ith particle, determine its optimal position P _bi , compare all The fitness value of the particle at its optimal position P _bi determines the optimal position G _b of the entire population, and adjusts the speed and position of the particle according to the position of each particle and the optimal particle position. When the iteration termination condition is reached, the optimal position is obtained. Position G _b , using the searched optimal position G _b to optimize the GRNN neural network.

For the shield machine tool wear monitoring device and method based on the topography of the ballast piece according to the above-mentioned embodiments of the present invention, the shield machine tool wear monitoring device is built on the shield machine conveyor belt 1 before the shield machine is constructed. , By manually collecting the hob structural parameters and engineering geological parameters of the current shield machine and inputting the hob wear prediction system, when the shield machine is running, the shield machine hob wear monitoring device is used to obtain real-time information on the shield machine during operation. The shape parameters and temperature parameters of the ballast piece are input and stored in the hob wear prediction system. Specifically, the camera 6 is used to take a picture of the ballast piece on the conveyor belt 1 of the shield machine in real time and input it into a computer to carry out the shape of the ballast piece. Calculate the size, obtain the input of ballast piece morphology parameters and store it in the hob wear prediction system, measure the temperature input of the ballast piece on the shield machine conveyor belt 1 in real time by the infrared thermometer 7 and store it in the The hob wear prediction system; the shield tunneling parameters input during the current shield machine construction are collected in real time through the main control room of the shield machine and stored in the hob wear prediction system; the hob wear prediction system is based on the input rolling The tool structure parameters, engineering geological parameters, shield tunneling parameters, ballast sheet shape parameters, temperature parameters and the analysis results of the existing mapping relationship output the hob wear amount corresponding to the current input parameters, and judge whether the output hob wear amount exceeds the set value. Set the hob wear amount alarm threshold, when the output hob wear amount does not exceed the set hob wear amount alarm threshold, continue to predict the hob wear amount of the next group of parameters; when the output hob wear amount exceeds the set hob wear amount When the hob wear amount alarm threshold is set, an alarm will be issued, the shield machine will be stopped, the staff will replace the hob of the shield machine, measure the hob wear amount of the currently worn hob and input the hob wear prediction The system stores and uses the covariance of the hob wear amount of the currently worn hob and the predicted hob wear amount as the fitness function of the particle swarm optimization algorithm to find the optimal smooth factor, and optimizes the CRNN neural network according to the optimal smooth factor. Start the shield machine and continue to predict the hob wear amount of the next set of parameters until the entire shield engineering work is completed and the monitoring is ended.

The device and method for monitoring the wear of shield hob based on the topography of the ballast piece described in the above-mentioned embodiments of the present invention input the wear of the hob by collecting actual engineering geological parameters, structural parameters of the hob and real-time excavation parameters and parameters of the ballast piece. Prediction system, the hob wear prediction system can predict the wear state of the hob in real time. The hob wear prediction system provides an alarm function when the predicted wear state of the hob reaches the set alarm threshold of the hob wear amount. When an alarm occurs, the particle swarm optimization algorithm is used to optimize the parameters of the GRNN neural network according to the predicted and actual hob wear, so as to reduce the influence of human factors on the neural network design, and improve the adaptability of the GRNN neural network. On the basis of the mapping relationship, an adaptive improvement algorithm is added to improve the robustness of the hob wear prediction under the condition of maintaining high prediction accuracy, so that the shield hob wear monitoring device and method based on the topography of the ballast piece can be based on Different projects improve accurate prediction of hob wear through closed-loop feedback adaptive adjustment.

The above are the preferred embodiments of the present invention. It should be pointed out that for those skilled in the art, without departing from the principles of the present invention, several improvements and modifications can be made. It should be regarded as the protection scope of the present invention.

Claims (6)

1. A method for monitoring the wear of shield hob based on the topography of the ballast piece, applied to a device for monitoring the wear of a shield machine hob based on the topography of the ballast piece, comprising: a data acquisition box, the data acquisition box is arranged on the shield. One side of the conveyor belt of the mechanism; an industrial computer, the industrial computer is arranged on one side of the data acquisition box, the first end of the industrial computer is electrically connected to the first end of the data acquisition box, and the industrial computer is connected to the first end of the data acquisition box. The second end is electrically connected to the computer; the truss, the first end of the truss is erected on the conveyor belt of the shield machine, the second end of the truss is erected on the ground; the lighting device, the lighting device is provided with two, two Each of the lighting devices is symmetrically arranged on both ends of the truss, and both of the lighting devices are detachably connected to the truss; a camera, the camera is arranged on the top of the truss, and the camera is connected to the truss. The truss is detachably connected, and the camera is electrically connected to the second end of the data acquisition box; an infrared thermometer is arranged at the top of the truss, and the infrared thermometer is connected to the truss. Detachable connection, the infrared thermometer is electrically connected to the third end of the data acquisition box, characterized in that it includes: Step 1: Obtain known shield engineering cases in the computer and build a database; Step 2, using the GRNN neural network to analyze the mapping relationship between the hob structure parameters, engineering geological parameters, shield tunneling parameters, ballast piece topography parameters and temperature parameters and the hob wear amount in the database, and obtain the mapping relationship analysis result; Step 3, build a hob wear prediction system based on the database, the GRNN neural network and the analysis results of the mapping relationship, and set an alarm threshold for the hob wear amount; Step 4, collect the hob structure parameters and engineering geological parameters of the current shield machine before construction, and input the hob wear prediction system; Step 5, collecting the shield tunneling parameters, ballast piece topography parameters and temperature parameters during the construction of the current shield machine in real time and inputting them into the hob wear prediction system; Step 6, the hob wear prediction system predicts the hob wear amount according to the input hob structure parameters, engineering geological parameters, shield tunneling parameters, ballast piece shape parameters and temperature parameters, and sets the predicted hob wear amount and setting. Compare with the alarm threshold of hob wear amount; Step 7: When the predicted hob wear exceeds the hob wear alarm threshold, the hob wear prediction system will issue an alarm, the shield machine will stop running, replace the hob, measure the wear of the worn hob and input the hob wear forecasting system; Step 8, adopt the particle swarm optimization algorithm to adaptively optimize the GRNN neural network according to the wear amount of the worn hob and the predicted wear amount of the hob; Step 9, restart the shield machine after the tool change is completed, jump to step 4 and continue to execute until the entire shield project is completed, and the monitoring ends; Step 10, when the predicted hob wear amount does not exceed the hob wear amount alarm threshold, jump to step 5 and continue to execute until the entire shield tunneling work is completed, ending the monitoring. 2. The method for monitoring the abrasion of shield hob based on the topography of the ballast piece according to claim 1, wherein the step 1 specifically comprises: The shield engineering case includes a number of tool change records. The tool change records include hob structural parameters, engineering geological parameters, shield tunneling parameters, ballast shape parameters, temperature parameters and hob wear. Among them, the hob structural parameters include: Blade structure parameters, hob radius, cutter spacing and load weight, engineering geological parameters include rock mechanics parameters, rock material parameters, rock joint and fault parameters and lithology parameters, shield tunneling parameters include shield machine penetration, Shield machine cutter head speed, shield machine thrust, shield machine working torque, shield machine soil bin pressure and shield machine propulsion speed, ballast piece morphology parameters including ballast piece particle size distribution index, ballast piece length and short axis ratio index and ballast sheet texture index, and temperature parameters include ballast sheet temperature. 3. The method for monitoring the abrasion of shield hob based on the topography of the ballast piece according to claim 2, wherein the step 2 specifically comprises: The penetration, propulsion speed, cutter head rotation speed, thrust of the shield machine, working torque of the shield machine, soil pressure of the shield machine, and radius of the hob are used as the 7 neurons of the input layer of the GRNN neural network. , the tool wear amount is used as the neuron of the output layer to form a GRNN neural network; Divide the data in the database into 100 groups, use random sampling to select 10 groups of data from the 100 groups of data as the test set, and the remaining 90 groups of data as the training set; The training set is randomly divided into 9 units, each unit includes 10 sets of data, 8 units are randomly selected from the 9 units as the input sample of the training set by the cross-validation method, and the remaining 1 unit is used as the output sample of the training set. The input sample data of the training set is normalized to between [-1, 1], and the verification search is performed with a step size of 0.01 in (0, 1], and the smooth factor σ that minimizes the mean square error between the predicted value and the sample value is found, and Record the best input sample and the best output sample corresponding to the current smooth factor; The test set data is normalized, and the obtained smooth factor σ, the best input sample and the best output sample are used as input variables to build a 4-layer GRNN neural network, and the output layer outputs the tool wear amount. 4. The method for monitoring the abrasion of shield hob based on the topography of the ballast piece according to claim 3, wherein the step 3 specifically comprises: Based on the GRNN neural network, the mapping relationship between the engineering geological parameters, shield tunneling parameters, shield hob structure, ballast shape size and temperature parameters and the tool wear amount in the database is collected through correlation analysis, and the hob wear prediction system is established: δ _i =β ₀ p+β ₁ v+β ₂ n+β ₃ F+β ₄ T+β ₅ S+β ₆ L+...+β _m X _m +C (1) Among them, β ₀ , β ₁ , β ₂ , β ₃ , β ₄ , β _m are the parameters to be estimated, δ _i is the tool wear amount, p is the penetration degree of the shield machine, v is the propulsion speed, n Indicates the rotation speed of the cutter head, F represents the thrust of the shield machine, T represents the working torque of the shield machine, L represents the radius of the hob, S represents the shape and size of the ballast piece, X _m represents other parameters related to the amount of tool wear, and C represents to be determined constant. 5. The method for monitoring the abrasion of shield hob based on the topography of the ballast piece according to claim 4, wherein the step 5 specifically comprises: The main control room of the shield machine is used to collect the current shield tunneling parameters during the construction of the shield machine in real time and input them into the hob wear prediction system. The shape and size of the ballast piece are calculated, and the shape parameters of the ballast piece are obtained and input into the hob wear prediction system. 6. The method for monitoring the abrasion of shield hob based on the topography of the ballast piece according to claim 5, wherein the step 8 specifically comprises: The particle swarm optimization algorithm is used to optimize the CRNN neural network: set the particle swarm calculation parameters, and set the mean square error of the tool wear amount predicted when the hob wear prediction system sends an alarm and the wear amount of the worn hob as the fitness function. The samples and examples are brought into the GRNN neural network, the fitness value F _i is calculated, the fitness values of all positions passed by the i-th particle are compared, and the optimal position P _bi is determined, and the optimal position P _bi of all particles is compared. The fitness value, determine the optimal position G _b of the whole population, adjust the speed and position of the particle according to the position of each particle itself and the optimal particle position, when the iteration termination condition is reached, obtain the optimal position G _b , use the searched optimal position G b . The optimal location G _b optimizes the GRNN neural network.

Images