CN108470095B - TBM (tunnel boring machine) propulsion prediction method based on data-driven radial basis function model - Google Patents

TBM (tunnel boring machine) propulsion prediction method based on data-driven radial basis function model Download PDF

Info

Publication number
CN108470095B
CN108470095B CN201810186041.2A CN201810186041A CN108470095B CN 108470095 B CN108470095 B CN 108470095B CN 201810186041 A CN201810186041 A CN 201810186041A CN 108470095 B CN108470095 B CN 108470095B
Authority
CN
China
Prior art keywords
propulsion
data
tbm
basis function
radial basis
Prior art date
Legal status (The legal status 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 status listed.)
Active
Application number
CN201810186041.2A
Other languages
Chinese (zh)
Other versions
CN108470095A (en
Inventor
王林涛
李�杰
孙伟
栾鹏龙
杜家楠
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Dalian University of Technology
Original Assignee
Dalian University of Technology
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 Dalian University of Technology filed Critical Dalian University of Technology
Priority to CN201810186041.2A priority Critical patent/CN108470095B/en
Publication of CN108470095A publication Critical patent/CN108470095A/en
Application granted granted Critical
Publication of CN108470095B publication Critical patent/CN108470095B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/15—Correlation function computation including computation of convolution operations

Abstract

The invention discloses a TBM (tunnel boring machine) propulsion prediction method based on a data-driven radial basis function model, which comprises the following steps of: s1, collecting data recorded on site in the TBM tunneling process; s2, determining relevant variable factors influencing the propelling force of the TBM; s3, constructing a radial basis function propulsion model; s4, obtaining factors having main influence on the propelling force; s5, constructing a radial basis function propulsion model again by using factors which mainly affect the propulsion; s6, inputting the prediction sample set into the radial basis function propulsion model to obtain a prediction result; and S7, evaluating the prediction result by adopting the relative error, and obtaining the predicted evaluation criterion and result by applying a statistical analysis method to the relative error. The invention provides a method for processing TBM construction site data, which makes full use of recorded data, obtains the relation between the recorded data and the propelling force, provides reference for TBM regulation and control in the tunneling process and ensures normal and efficient operation of the TBM.

Description

TBM (tunnel boring machine) propulsion prediction method based on data-driven radial basis function model
Technical Field
The invention relates to the technical field of tunnel boring machines, in particular to a TBM (tunnel boring machine) propulsion prediction method based on a data-driven radial basis function model.
Background
Hard rock tunneling has become a common mode of various tunneling, and a full face rock tunneling machine (TBM) is used for tunneling under different geological conditions, and is required to have higher efficiency and reasonable power for tunneling. In the TBM tunneling process, the propelling force in the tunneling load is a heavy parameter, and the propelling force has great influence on the service life of a cutter on a cutter head, the optimized design of the structure, the improvement of the tunneling efficiency and the estimation of the overall economic cost. Meanwhile, with the development of 'informatization and intellectualization' technology in the field of equipment manufacturing, a 'prediction' function is becoming a core technology in equipment manufacturing, and the method can effectively infer and predict key parameters in the equipment, and is an important basis for optimal design and intelligent control. Therefore, the prediction of the propelling force of the TBM is particularly important in the tunneling process.
The sensors and detection equipment in the TBM construction process can provide wide real-time quality measurement data, wherein the data comprise key control variable information in the equipment operation process, such as tunneling speed, penetration degree, shoe supporting pressure and the like. And the inherent relation between the influence factors and the propelling force is determined, and a propelling force model capable of reflecting the inherent regularity is established, so that the method has important significance on the design, control and safe operation of the TBM equipment. And the TBM predicts the propelling force in the tunneling process by using the established propelling force model according to the construction data recorded on site in the construction process, so that the recorded data can be effectively utilized, and the adjustment control is made for the section to be tunneled.
Disclosure of Invention
In light of the above-mentioned technical problem, a method for predicting TBM propulsion based on a data-driven radial basis function model is provided. The technical means adopted by the invention are as follows:
a TBM propulsion prediction method based on a data-driven radial basis function model comprises the following steps:
s1, collecting data recorded on site in the TBM tunneling process;
s2, determining related variable factors influencing the propelling force of the TBM, and establishing a sample data set of the variable factors as a training set;
s3, constructing a radial basis function propulsion model, inputting a training set, and training the radial basis function propulsion model to obtain the weight of the radial basis function propulsion model;
s4, obtaining factors having main influence on the propulsion by estimating a sample mean value and a sample standard deviation method according to the radial basis function propulsion model;
s5, according to the obtained factors and training sets which have main influence on the propulsion, carrying out reconstruction on the radial basis function propulsion model by the factors which have main influence on the propulsion, wherein 80% of data of the factors which have main influence on the propulsion are used as modeling samples, 20% of data of the factors are used as detection samples, carrying out simple cross validation, and determining the precision of the radial basis function propulsion model;
s6, taking a sample set of factors which are to be predicted and have main influence on the propulsion in a tunneling period as a prediction sample set, inputting the sample set into a radial basis function propulsion model, and obtaining a prediction result;
s7, evaluating the prediction result by adopting a relative error, and obtaining a predicted evaluation criterion and result by applying a statistical analysis method to the relative error;
s8, in the TBM tunneling process, generating new data with main influence factors on the propulsion, reconstructing a new training set with the data with the main influence factors on the propulsion obtained in the previous tunneling period, reconstructing a radial basis function propulsion model, and repeating the steps S6 to S8 to enable the radial basis function propulsion model to have certain dynamic learning capacity and form a prediction model for predicting the propulsion while tunneling.
The specific steps of step S1 are as follows: and (3) dividing the TBM tunneling period, identifying effective data in each tunneling period, completing identification of the tunneling period, and carrying out uniform data format on the data in each tunneling period.
Before the step S2, the data collected in the step S1 needs to be processed, which includes the following steps:
screening of data: the following measures are taken for the emergency and data loss encountered in the TBM tunneling process:
a. data elimination: according to the characteristics of the TBM in the tunneling process, the tunneling time and the tunneling distance are taken as references, and the data of the TBM between the last tunneling period and the next tunneling period are removed, so that the aim of continuously changing the propelling force in the tunneling process is fulfilled;
b. and (3) data completion: in a tunneling period, utilizing a mean value replacement method to fill in missing and abnormal data;
normalized values: the value range of the data value of the data with different value ranges of each dimension of the data collected in the step S1 is processed to be between 0 and 1.
In step S2, the relevant variable factors affecting the TBM thrust include a front shield pitch angle, a front shield roll angle, a left side shield pressure, a right side shield pressure, a top shield pressure, a left side rear support pressure, a right side rear support pressure, a shoe support pressure, a cutter head water spray pressure, EP2 (lubricating grease code) inner seal pressure, EP2 outer seal pressure, a cutter head rotation speed detection value, penetration, a thrust speed, a left shoe pitch angle, a left shoe roll angle, a right shoe pitch angle, a right shoe roll angle, a host belt machine rotation speed, a bridge belt machine rotation speed, and a slag belt machine rotation speed.
In the step S3, an MQ kernel function is used to construct and solve the radial basis function propulsion model, so as to obtain the weight of the radial basis function propulsion model.
In step S4, the method of estimating the mean and standard deviation of the sample refers to performing qualitative analysis on the variable factors affecting the thrust of the TBM by Morris' S method, and the factors mainly affecting the thrust include left shield pressure, right shield pressure, boot pressure, penetration and thrust speed. Morris's method aims at estimating the parameters of the distribution of the fundamental effects associated with each variable, in fact determining the magnitude of the effect of each of its parameters on the target variable by estimating the values calculated in different parts of the design space for the sample mean and the sample standard deviation. Firstly, generating a sampling sample, evaluating the influence of two variables participating in basic calculation by an objective function F, sequentially obtaining an estimated sample standard deviation (ssd) and an estimated sample mean value (sm) of each design variable respectively, and further realizing the identification of main influence parameters.
In the step S5With R2To evaluate the accuracy of the radial basis function thrust model:
Figure BDA0001590296170000031
where m is the number of detection points, yiIn order to detect a response from the point,
Figure BDA0001590296170000032
as a result of the calculation of the radial basis function model,
Figure BDA0001590296170000033
for the average of the point responses, R is determined according to the above formula2The closer the value of (a) is to 1, the better the fitting degree of the model is, and the higher the accuracy is.
The evaluation of the prediction result by using the relative error in step S7 refers to comparing the prediction result with the actual propulsion value, and evaluating the prediction result by using the evaluation criterion of the relative error, where the expression of the relative error is as follows:
Figure BDA0001590296170000034
wherein, amAnd acRespectively representing the actual propulsion value and the predicted result, epsilonaIs a relative error.
The method constructs a model by using data in the tunneling process, and solves the problem that the TBM propelling force can only be predicted by using a mechanical analysis method; the method avoids the problem of difficult expression of complex relations, considers the relation between the target variable and the influence factor from the data analysis perspective, increases the learning capacity of the model, improves the utilization of construction data, and simultaneously improves the precision of predicting TBM (tunnel boring machine) propelling force; the invention provides guidance for the optimization design of the structure, the improvement of the tunneling efficiency and the estimation of the overall economic cost.
Based on the reasons, the invention can be widely popularized in the fields of tunnel boring machine technology and the like.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a schematic diagram of radial basis functions in an embodiment of the present invention.
FIG. 2 is a graph of estimated sample mean and sample standard deviation in an embodiment of the present invention.
Fig. 3 is a graph comparing the predicted result and the actual propulsion value in the embodiment of the present invention.
FIG. 4 is a predicted relative error distribution diagram of a radial basis function propulsion model in an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The data processing and analysis are generated on the basis of big data, and the enterprise data reduction data are analyzed and processed by using a big data technical means, and the intrinsic values of the mass data are mined, so that a certain purpose is realized for people.
Principle of radial basis function model:
in the illustration of fig. 1, it is assumed that the input sample variable and the number of variables are x and n, respectively, and the output is y ═ y1,y2,···,yn]TRadial basis function propulsion model can be tabulatedDais type (1)
Figure BDA0001590296170000041
Wherein n iscRepresenting the number of basis functions, c(i)Represents ncThe center of the ith one of the basis functions, denoted by psi, contains the center point c for estimating the input point x(i)N of difference betweencWiki functions, w represents weights, and x represents input variables. The radial basis function has an input x, an implicit psi function and a weight w, linear
Figure BDA0001590296170000042
The method mainly uses the radial basis function to predict the TBM propelling force, and the radial basis function mainly constructs a corresponding relation between a vector with prediction of training data and the radial basis function and predicts the vector to be predicted in test data.
The following types of psi functions are commonly used, as shown in table 1 below:
TABLE 1 several commonly used function types
Figure BDA0001590296170000051
In table 1 above, some parameters can be optionally set, or the parameters are estimated and determined, w can be easily measured for the determined basis functions, and the basis functions can be approximated by a difference condition, as shown in the following formula (2)
Figure BDA0001590296170000052
Unifying data points in a general solution process, i.e. let c(i)＝x(i)，
Figure BDA0001590296170000053
The above equation can be expressed as a matrix equation (3)
ψw＝y (3)
Wherein ψ is called Gram matrix (Gram matrix) and is defined as formula (4)
ψi,j＝ψ(||x(i)-x(j)||)，i,j＝1,...,n (4)
The weight w can thus be expressed as formula (5)
w＝ψ-1y (5)
For the information set, assume that the training set S { (x)1,x2...xk,y1),(x1,x2...xk,y2)···(x1,x2...xk,yl) Therein of
Figure BDA0001590296170000054
k represents the number of features of the sample, l represents the number of samples, and the fitting process is to find the weight w1,w2...wk. The specific fitting process is as follows:
w1ψ(||x1-x1||)+...+wkψ(||x1-xk||)＝y1
w1ψ(||x2-x1||)+...+wkψ(||x2-xk||)＝y2
w1ψ(||xk-x1||)+...+wkψ(||x1-xk||)＝yk
thus can obtain1,w2...wkSo that f (x) can be expressed as shown in the following formula (6)
f(x)＝w1ψ(||x-x1||)+...+wkψ(||x-xk||) (6)
The following is specifically described in connection with a specific TBM in the actual propulsion process:
as shown in fig. 2-4, a method for predicting TBM propulsion based on a data-driven radial basis function model includes the following steps:
s1, collecting data recorded on site in the TBM tunneling process
In the embodiment, a certain water supply tunnel project is mainly used as a research object, field excavation data in the TBM tunneling process needs to be collected firstly, and because data sources recorded in excavation fields are different, formats are different, characteristics and properties of the data are different, the data are necessary in data collection and in a unified data format, and meanwhile, in the data recording process, the missing data and mutation possibly occur aiming at the emergency situation encountered in the tunneling process, and the missing data needs to be supplemented and removed. In the aspect of data supplement and elimination, data processing is carried out according to a tunneling period, missing data is supplemented by an average value in one tunneling period, and the data is supplemented by artificial filling and the average value replacement method to supplement abnormal data. In order to better operate in a computer and eliminate the influence of data characteristic data before modeling initialization on a result, the data is normalized to enable a transformation interval to be a range of [0, 1 ];
s2, determining related variable factors influencing the propelling force of the TBM, and establishing a sample data set of the variable factors as a training set:
the thrust of the TBM during tunneling is the result of multiple resistance interactions. The resistance borne by a TBM in the tunneling process of a certain water supply tunnel engineering is analyzed by combining the excavation current situation of the tunnel engineering, and the tunnel engineering can be divided into three types, namely rock breaking resistance between a cutter head and rock, friction between a shield and the tunnel wall and subsequent rock slag conveying devices. For these resistances, the recorded data of their TBM during the heading can be reflected and by analyzing the recorded data during the on-site heading, 21 relevant influencing variables are determined. As shown in table 2 below. Furthermore, the propulsion oil cylinder of the TBM in the propulsion process has periodic retraction and extension, so that the data record can be divided by the propulsion period in the tunneling process, and a model of the propulsion force is also established periodically. After passing the analysis, a sample data set of the variable factors is established as a training set, wherein the sample size of the training set is 200. Table 3 shows the raw data of the training sample part, and table 4 shows the normalized data result.
Table 221 relevant design variables
Figure BDA0001590296170000071
TABLE 3 Propulsion related variable factor (raw data)
Figure BDA0001590296170000072
Figure BDA0001590296170000081
Partial data of normalized data in table 4
Figure BDA0001590296170000082
S3, constructing a radial basis function propulsion model, inputting a training set, training the radial basis function propulsion model, and obtaining the weight:
the method comprises the steps of constructing a radial basis function, firstly determining a design input variable and a target function variable of the radial basis function, constructing a radial basis function propulsion model by taking 21 parameters as input variables and total propulsion in a tunneling process as a target variable, and constructing the model by taking a multi-element quadratic as a kernel function. The number of training sample sets is 200, wherein the kernel function selects multivariate quadratic (multivariate quadratic) to operate, and the weight of the kernel function is obtained. MQ (multivariate quadratic) kernel function is adopted, so that a Gram matrix (Gram matrix) in operation can be decomposed through LU (Cholesky), and the operation speed is increased. The partial weights are shown in Table 5 below
TABLE 5 partial weight values
Figure BDA0001590296170000083
S4, according to the radial basis function propulsion model, obtaining factors having main influence on the propulsion by using a method of estimating a sample mean value and a sample standard deviation:
the method of using Morris (Morris' method) according to the radial basis function model that has been constructed aims at estimating the parameters of the distribution of the fundamental effects associated with each variable, in fact determining the magnitude of the effect of each of its parameters on the target variable by estimating the values calculated in different parts of the design space for the sample mean and the sample standard deviation. Firstly, generating a sampling sample, evaluating the influence of two variables participating in basic calculation by an objective function F, sequentially obtaining the standard deviation (ssd) of an estimated sample and the mean value (sm) of the estimated sample of each design variable, and further realizing the identification of main influence parameters.
As shown in fig. 2, where points far from the zero point indicate that the influence on the propulsion is large, and points near the zero point indicate that the influence on the propulsion is small, qualitative analysis shows that there are 5 factors having a main influence on the propulsion. As shown in table 6 below
Table 6 5 main influencing factors on propulsion
Figure BDA0001590296170000091
S5, according to the obtained factors and training sets which have main influence on the propulsion, carrying out reconstruction on the radial basis function propulsion model by the factors which have main influence on the propulsion, wherein 80% of data of the factors which have main influence on the propulsion are used as modeling samples, 20% of data are used as detection samples, carrying out simple cross validation, and determining the precision of the radial basis function propulsion model:
the propulsion model of the TBM can be represented by the following formula (7), where ψ (r) ═ r2+σ2)1/2，σ＝1，
f(x)＝w1ψ(||x-x1||)+...+wkψ(||x-xk||) (7)
The weight w is shown in the following table 7
TABLE 7 partial weight values w
Figure BDA0001590296170000092
And according to the factors and the training set which have the main influence on the propulsion and are obtained in the step S4, the factors which have the main influence on the propulsion are used for constructing the radial basis function propulsion model again, and the established model is subjected to simple cross validation. Namely, 80% of the selected sample data is used for training the model, and 20% of the selected sample data is used as a detection sample, so that the verification of the model is completed. The evaluation criterion of the model adopts a correlation coefficient R2(8) To evaluate the accuracy of the radial basis function propulsion model,
Figure BDA0001590296170000101
where m is the number of detection points, yiIn order to detect a response from the point,
Figure BDA0001590296170000102
as a result of the calculation of the radial basis function model,
Figure BDA0001590296170000103
for the average of the point responses, R is determined according to the above formula2The closer the value of (a) is to 1, the better the fitting degree of the model is, and the higher the accuracy is. According to the obtained factors and training set which have main influence on the propulsion, the factors which have main influence on the propulsion are used for constructing the radial basis function propulsion model again, wherein the number of the samples for modeling is 200, the number of the detection sample points is 50, and therefore the precision R of the model is obtained20.8849 according to R2>A criterion of 0.85, it can be determined that the model is known to meet the accuracy requirement.
And S6, taking a sample set of factors which are to be predicted and have main influence on the propulsion in a tunneling period as a prediction sample set, inputting the prediction sample set into the radial basis function propulsion model, and obtaining a prediction result, wherein the prediction result is shown in figure 3.
S7, evaluating the prediction result by adopting a relative error, and obtaining the predicted evaluation criterion and result by applying a statistical analysis method to the relative error:
the predicted result is compared with the actual propulsion value, and the estimated result is evaluated by using the evaluation criterion of the relative error expressed by the following formula (9)
Figure BDA0001590296170000104
Wherein, amAnd acRespectively representing the actual propulsion value and the predicted result, epsilonaIs a relative error.
As for the characteristic analysis of the TBM in the propelling process, the value of the propelling force of the TBM is discontinuously changed and has great randomness, so that the method adopts a statistical method for analyzing the relative error of the TBM, and the TBM has the advantages of intuition and simplicity. The predicted results are shown in table 8 below.
TABLE 8 prediction statistics for radial basis function models
Figure BDA0001590296170000105
It can be seen that when the prediction result is within 20% of the allowable error, the prediction accuracy reaches 83.06, and the feasibility of the method is verified.
S8, in the TBM tunneling process, generating new data with main influence factors on the propulsion, reconstructing a new training set with the data with the main influence factors on the propulsion obtained in the previous tunneling period, reconstructing a radial basis function propulsion model, and repeating the steps S6 to S8 to enable the radial basis function propulsion model to have certain dynamic learning capacity and form a prediction model for predicting the propulsion while tunneling.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present invention.

Claims (8)

1. A TBM propulsion prediction method based on a data-driven radial basis function model is characterized by comprising the following steps:
s1, collecting data recorded on site in the TBM tunneling process;
s2, determining related variable factors influencing the propelling force of the TBM, and establishing a sample data set of the variable factors as a training set;
s3, constructing a radial basis function propulsion model, inputting a training set, and training the radial basis function propulsion model to obtain the weight of the radial basis function propulsion model;
s4, obtaining factors having main influence on the propulsion by estimating a sample mean value and a sample standard deviation method according to the radial basis function propulsion model;
s5, according to the obtained factors and training sets which have main influence on the propulsion, the factors which have main influence on the propulsion are subjected to reconstruction of a radial basis function propulsion model, wherein 80% of data of the factors which have main influence on the propulsion are used as modeling samples, 20% of data of the factors which have main influence on the propulsion are used as detection samples, simple cross validation is carried out, and the precision of the radial basis function propulsion model is determined;
s6, taking a sample set of factors which are to be predicted and have main influence on the propulsion in a tunneling period as a prediction sample set, inputting the sample set into a radial basis function propulsion model, and obtaining a prediction result;
s7, evaluating the prediction result by adopting a relative error, and obtaining a predicted evaluation criterion and result by applying a statistical analysis method to the relative error;
and S8, in the TBM tunneling process, generating new data with main influence factors on the propulsion, reconstructing a new training set with the data with the main influence factors on the propulsion obtained in the previous tunneling period, reconstructing a radial basis function propulsion model, and repeating the steps S6 to S8.
2. The method of claim 1, wherein: the specific steps of step S1 are as follows: and (3) dividing the TBM tunneling period, identifying effective data in each tunneling period, completing identification of the tunneling period, and carrying out uniform data format on the data in each tunneling period.
3. The method of claim 1, wherein: before the step S2, the data collected in the step S1 needs to be processed, which includes the following steps:
screening of data: the following measures are taken for the emergency and data loss encountered in the TBM tunneling process:
a. data elimination: according to the characteristics of the TBM in the tunneling process, the tunneling time and the tunneling distance are taken as references, and the data of the TBM between the last tunneling period and the next tunneling period are removed, so that the aim of continuously changing the propelling force in the tunneling process is fulfilled;
b. and (3) data completion: in a tunneling period, utilizing a mean value replacement method to fill in missing and abnormal data;
normalized values: the value range of the data value of the data with different value ranges of each dimension of the data collected in the step S1 is processed to be between 0 and 1.
4. The method of claim 1, wherein: in step S2, the relevant variable factors affecting the TBM thrust include a front shield pitch angle, a front shield roll angle, a left side shield pressure, a right side shield pressure, a top shield pressure, a left side rear support pressure, a right side rear support pressure, a shoe support pressure, a cutter head water spray pressure, an EP2 inner seal pressure, an EP2 outer seal pressure, a cutter head rotation speed detection value, a penetration degree, a thrust speed, a left shoe pitch angle, a left shoe support roll angle, a right shoe pitch angle, a right shoe roll angle, a host belt machine rotation speed, a bridge belt machine rotation speed, and a slag belt machine rotation speed.
5. The method of claim 1, wherein: in the step S3, an MQ kernel function is used to construct and solve the radial basis function propulsion model, so as to obtain the weight of the radial basis function propulsion model.
6. The method of claim 1, wherein: in step S4, the method of estimating the mean and standard deviation of the samples refers to performing qualitative analysis on the relevant variable factors affecting the TBM thrust by morris, and the factors mainly affecting the thrust include left shield pressure, right shield pressure, boot pressure, penetration and thrust speed.
7. The method of claim 1, wherein: in the step S5, R is used2To evaluate the accuracy of the radial basis function thrust model:
Figure FDA0001590296160000021
where m is the number of detection points, yiIn order to detect a response from the point,
Figure FDA0001590296160000022
as a result of the calculation of the radial basis function model,
Figure FDA0001590296160000023
for the average of the point responses, R is determined according to the above formula2The closer the value of (a) is to 1, the better the fitting degree of the model is, and the higher the accuracy is.
8. The method of claim 1, wherein: the evaluation of the prediction result by using the relative error in step S7 refers to comparing the prediction result with the actual propulsion value, and evaluating the prediction result by using the evaluation criterion of the relative error, where the expression of the relative error is as follows:
Figure FDA0001590296160000024
wherein, amAnd acRespectively representing the actual propulsion value and the predicted result, epsilonaIs a relative error.
CN201810186041.2A 2018-03-07 2018-03-07 TBM (tunnel boring machine) propulsion prediction method based on data-driven radial basis function model Active CN108470095B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810186041.2A CN108470095B (en) 2018-03-07 2018-03-07 TBM (tunnel boring machine) propulsion prediction method based on data-driven radial basis function model

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810186041.2A CN108470095B (en) 2018-03-07 2018-03-07 TBM (tunnel boring machine) propulsion prediction method based on data-driven radial basis function model

Publications (2)

Publication Number Publication Date
CN108470095A CN108470095A (en) 2018-08-31
CN108470095B true CN108470095B (en) 2021-04-02

Family

ID=63264188

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810186041.2A Active CN108470095B (en) 2018-03-07 2018-03-07 TBM (tunnel boring machine) propulsion prediction method based on data-driven radial basis function model

Country Status (1)

Country Link
CN (1) CN108470095B (en)

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
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
CN109543268B (en) * 2018-11-14 2023-05-05 大连理工大学 TBM propulsion main influencing factor identification method based on kriging model
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
CN109854259B (en) * 2019-03-15 2020-09-22 中铁高新工业股份有限公司 Method and system for obtaining optimal value range of construction tunneling parameters of shield tunneling machine
CN111140244B (en) * 2020-01-02 2021-04-23 中铁工程装备集团有限公司 Intelligent support grade recommendation method for hard rock heading machine
CN113033004A (en) * 2021-03-30 2021-06-25 中铁工程装备集团有限公司 Tunnel boring machine propulsion process friction force calculation method based on data driving

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102243672A (en) * 2011-06-22 2011-11-16 浙江大学 Gushing operation condition soft sensing modeling method based on hybrid multiple models in shield tunneling process
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

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CA2696152A1 (en) * 2000-06-29 2002-01-10 Aspen Technology, Inc. Computer method and apparatus for constraining a non-linear approximator of an empirical process

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102243672A (en) * 2011-06-22 2011-11-16 浙江大学 Gushing operation condition soft sensing modeling method based on hybrid multiple models in shield tunneling process
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

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
global sensitivity analysis using a gaussian radial basis function metamodel;Zeping Wu等;《Reliability Engineering and System Safety》;20160608;第154卷;171-179页 *
TBM刀盘驱动系统分层次建模与耦合振动机理;丁鑫;《中国博士学位论文全文数据库 工程科技Ⅱ辑》;20180115(第1期);C034-47页 *
长沙地铁下穿湘江土压平衡盾构隧道掘进参数研究;褚东升;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》;20130215(第2期);C034-315页 *

Also Published As

Publication number Publication date
CN108470095A (en) 2018-08-31

Similar Documents

Publication Publication Date Title
CN108470095B (en) TBM (tunnel boring machine) propulsion prediction method based on data-driven radial basis function model
Liu et al. Prediction model of rock mass class using classification and regression tree integrated AdaBoost algorithm based on TBM driving data
Adoko et al. Bayesian prediction of TBM penetration rate in rock mass
CN106778010B (en) TBM cutter life prediction method based on data-driven support vector regression machine
CN111220387B (en) Vehicle bearing residual life prediction method based on multi-feature-quantity correlation vector machine
Dong et al. Deep transfer learning based on Bi-LSTM and attention for remaining useful life prediction of rolling bearing
Leng et al. A hybrid data mining method for tunnel engineering based on real-time monitoring data from tunnel boring machines
KR20190072652A (en) Information processing apparatus and information processing method
Bashari et al. Estimation of deformation modulus of rock masses by using fuzzy clustering-based modeling
CN108875118B (en) Method and device for evaluating accuracy of prediction model of silicon content of blast furnace molten iron
TWI584134B (en) Method for analyzing variation causes of manufacturing process and system for analyzing variation causes of manufacturing process
CN107368463B (en) Roadway nonlinear deformation prediction method based on fiber bragg grating sensor network data
Wang et al. Determination of the minimum sample size for the transmission load of a wheel loader based on multi-criteria decision-making technology
CN112948932A (en) Surrounding rock grade prediction method based on TSP forecast data and XGboost algorithm
CN113806889A (en) Processing method, device and equipment of TBM cutter head torque real-time prediction model
CN111143768A (en) Air quality prediction algorithm based on ARIMA-SVM combined model
CN103885867A (en) Online evaluation method of performance of analog circuit
CN109543268B (en) TBM propulsion main influencing factor identification method based on kriging model
CN115438897A (en) Industrial process product quality prediction method based on BLSTM neural network
Ma et al. A health indicator construction method based on self-attention convolutional autoencoder for rotating machine performance assessment
Wang et al. Reliability assessment of the vertical roller mill based on ARIMA and multi-observation HMM
Xiao et al. Support evidence statistics for operation reliability assessment using running state information and its application to rolling bearing
CN114818493A (en) Method for quantitatively evaluating integrity degree of tunnel rock mass
CN115700363A (en) Fault diagnosis method and system for rolling bearing of coal mining machine, electronic equipment and storage medium
CN115659271A (en) Sensor abnormality detection method, model training method, system, device, and medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant