CN106778010B - TBM cutter life prediction method based on data-driven support vector regression machine - Google Patents
TBM cutter life prediction method based on data-driven support vector regression machine Download PDFInfo
- Publication number
- CN106778010B CN106778010B CN201611240128.0A CN201611240128A CN106778010B CN 106778010 B CN106778010 B CN 106778010B CN 201611240128 A CN201611240128 A CN 201611240128A CN 106778010 B CN106778010 B CN 106778010B
- Authority
- CN
- China
- Prior art keywords
- data
- prediction
- support vector
- tbm
- cutter
- 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.)
- Expired - Fee Related
Links
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16Z—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS, NOT OTHERWISE PROVIDED FOR
- G16Z99/00—Subject matter not provided for in other main groups of this subclass
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of tunnel excavators, in particular to a TBM cutter service life prediction method based on a data-driven support vector regression machine, which comprises the following steps: 1. collecting data of a TBM tool excavation site; 2. determining a driving factor influencing the service life of the TBM tool, and establishing a sample data set of the driving factor as a training set; 3. constructing a prediction model of a multi-core support vector regression machine, inputting a training set, and training the prediction model so as to determine an optimal parameter, a penalty function C and an insensitive loss function parameter epsilon corresponding to each kernel function; 4. determining an optimal kernel function of the prediction model; 5. and taking the sample data set of the driving factor of the tool to be predicted as a prediction sample set, and inputting the prediction sample set into the prediction model to obtain a prediction result. The invention provides a TBM tool life prediction method based on a data-driven support vector regression machine, which selects a large amount of field-mined data as a parameter, constructs a model based on the support vector regression machine on the basis, and improves the precision of predicting the tool life.
Description
Technical Field
The invention relates to the technical field of tunnel boring machines, in particular to a TBM cutter service life prediction method based on a data-driven support vector regression machine.
Background
With the rapid development of cities, subways as important components of three-dimensional traffic become effective ways for solving urban congestion and have huge development potential. Urban geological conditions generally show diversity, and a Tunnel Boring Machine (TBM) for hard rock Tunnel construction is a complete set of advanced tunneling equipment integrating excavation, supporting and slag tapping, so that the consumption of cutters is huge in the construction process, the time and the labor for cutter replacement influence the construction period, and whether the time for cutter replacement can be shortened becomes an important factor for efficiently utilizing the TBM. The wear of each cutter is accurately predicted, an acting point for saving cost can be found for a TBM construction party, and the theoretical defects in the aspects of TBM cutter service life prediction and cutter scheduling are overcome.
Factors influencing the abrasion of the cutter are complex and mainly divided into two major aspects, namely static factors and dynamic factors, wherein the static factors comprise the components of the cutter, the shape of the cutter, the installation angle and the like, and the dynamic factors comprise actual geological factors, personnel operation factors and the like. At present, the research on the abrasion of TBM cutters at home and abroad is only researched from the aspects of mechanics and manufacturing materials, and has no practical development on the aspect of dynamic factor research. In practice, once the TBM is manufactured, the static factors cannot be changed, so that the research on actual geological factors and personnel operation factors has very important practical significance on the abrasion of the cutter.
Disclosure of Invention
In order to solve the technical problems, the invention provides a TBM cutter life prediction method based on a data-driven support vector regression machine.
In order to achieve the purpose, the invention adopts the following technical scheme: a TBM cutter life prediction method based on a data-driven support vector regression machine comprises the following steps:
(1) collecting data of a TBM tool excavation site;
(2) determining a driving factor influencing the service life of the TBM tool, and establishing a sample data set of the driving factor as a training set;
(3) constructing a prediction model of a multi-core support vector regression machine, inputting a training set, and training the prediction model so as to determine an optimal parameter, a penalty function C and an insensitive loss function parameter epsilon corresponding to each kernel function;
(4) determining an optimal kernel function of the prediction model;
(5) and taking the sample data set of the driving factor of the tool to be predicted as a prediction sample set, and inputting the prediction sample set into the prediction model to obtain a prediction result.
Further, after the step (1), before the step (2), the method further includes a step (11) of processing the collected data, and the processing includes:
(111) data integration: centralizing data with different sources, formats, characteristics and properties to unify data formats;
(112) data cleaning: the following measures are taken for the sudden situations and data loss encountered by construction,
a. data elimination: if the abnormal data of a certain cutter data is less than 5%, only removing the abnormal data, and if the data loss exceeds 30%, removing all the cutter number data;
b. and (3) data completion: completing partial abnormal data by using a manual filling and mean value replacing method;
(113) data mean normalization: each dimension of the data was normalized to a data set with a mean of 0, a variance of 1, and a normal distribution.
Further, the driving factors influencing the service life of the TBM cutter determined in the step (2) comprise mileage, rock grade, cutter head thrust, blade abrasion, cutter head rotating speed, tunneling speed, radius and cutter damage.
Further, the driving factors influencing the service life of the TBM tool in the step (2) are determined by processing collected data through a grey correlation analysis method.
Further, in the step (4), the training sample size is increased by adopting K-fold cross validation, and parameters of each group of kernel functions are decoupled by using a grid search method, so that the optimal parameters corresponding to each kernel function of the prediction model are determined.
Further, the kernel functions of the prediction model of the multi-kernel support vector regression machine constructed in the step (5) include a polynomial kernel function, a Gauss radial basis kernel function and a multilayer perceptron kernel function.
Further, the optimal kernel function of the prediction model is a multi-layer perceptron kernel function.
The support vector regression model is driven based on data to mine the data to construct the model, and the defect that the tool is damaged due to the fact that normal abrasion can only be predicted from the aspect of mechanical analysis in the past is overcome. The model omits the process of searching the factor relationship in the traditional modeling, avoids the problem that the complex relationship is difficult to express, and improves the precision of predicting the service life of the cutter. Theoretical guidance is provided for operators to schedule the tools, and the construction period is shortened. The recorded data in the excavation of a certain subway in a certain city is taken as a training set, the K-cross inspection and the grid search are combined to find the optimal parameters, the finally established kernel function is taken as a radial basis kernel function, the predicted service life of the cutter at the next stage is compared with the actual data, and the error range can be found through errors and is controlled within 4.5 percent, so that the service life of the cutter can be predicted by using the data which are simple and easy to obtain.
Drawings
FIG. 1 is a contour plot of SVR parameter selection;
FIG. 2 is a 3D view of SVR parameter selection;
FIG. 3 is a graph of actual life and predicted life for test set No. 1 tool test set.
Detailed Description
The invention will be described in detail with reference to the following embodiments:
and (3) data driving:
the data driving is generated on the basis of big data, and the data driving needs to analyze and process massive data of an enterprise by using a technical means of the big data and excavate the intrinsic value of the massive data so as to guide the enterprise to produce, sell, manage and manage.
The principle of the regression machine of the support vector machine is as follows:
support Vector Machines (SVMs), which are first proposed by cornina cortex, equal to 1995, show many advantages in solving small samples, nonlinearity, and high-dimensional recognition. The support vector machine method is established according to VC dimension theory and structure risk minimum theory in statistical learning theory, and seeks an optimal compromise between the complexity (learning precision for a specific training sample) and learning ability (ability to identify any sample without error) of a model according to limited sample information so as to obtain the best generalization ability. The objective of statistical learning is changed from empirical risk minimization to seeking the sum of empirical risk and confidence risk to be minimal, i.e. structural risk is minimal, and the formula of the generalized error bound is:
where R (ω) is the true risk, Remp(ω) is an empirical function of,
is the confidence interval.
Support vector machines can be classified into support vector classifiers and support vector regression machines (SVR) for classification problems and regression problems. The method mainly uses the SVR as a model for processing nonlinear fitting regression to predict the service life of the cutter, and the SVR is mainly used for establishing a corresponding relation between a vector to be predicted of training data and a support vector and performing simulation prediction on the vector to be predicted in test data. For the information set, assume that the training is given S { (x)1,x2...xk,y1),(x1,x2...xk,y2)…(x1,x2...xk,yl) Therein of
k represents the characteristic quantity of the samples, l represents the number of the samples, the SVR maps the data to a high-dimensional space according to the nonlinear transformation defined by the inner product kernel function, and regression fitting is completed in the high-dimensional space as follows:
whereinFor the feature space, ω is the weight coefficient and c is the bias term. According to the structural error minimization principle mentioned above, ω and c can be minimized according to the following function.
In the formula: l f (x)j)-yj| is a loss function, in order to make | ω |2Euler norm minimization and also to avoid accuracy of fit error beyond a predetermined setting, a relaxation variable is added
And
by adjusting, the optimization problem of the formula (2) is transformed into a constraint minimization problem, which is obtained after simplification:
satisfy the requirement of
To solve the problem of equation (3), lagrange multiplier a is introducedj,
ηj,
And constructing a Lagrangian equation. Separately solving omega, b and xi for Lagrange functionj,Partial derivative (value 0). By substituting the obtained result into the lagrange equation, equation (3) becomes:
satisfy the requirement of
After transformation, the problem becomes a problem of solving convex quadratic programming, and according to the method of solving quadratic programming, the final model is as follows:
wherein
Is a kernel function, xiRepresenting a vector of training samples, x being a vector of test samples. For different problems, the precision difference of selecting different kernel functions is very large, so whether to select proper kernel functions becomes a key factor influencing the prediction precision. There is no uniform method for selecting corresponding kernel functions for specific problems, and comparison can be performed only after multiple tests, and the commonly used kernel functions mainly include:
polynomial kernel function:
K(xi,x)＝[γ(xi·x)+coef]d
wherein: d is the order of the polynomial and coef is the bias coefficient.
RBF kernel function (Gauss radial basis kernel function)
K(xi,x)＝exp(-γ‖xi-x‖2)
Wherein: γ represents the radius of the kernel function.
Multilayer perception machine kernel (Sigmoid kernel)
K(xi,x)＝tanh(γ(xi·x)+coef)
Different kernel functions and parameters within the functions have a significant impact on the accuracy of the SVR model. When the feature matrix of the training sample is high-dimensional, d in the polynomial kernel function is large, so that the calculation complexity is high, and a satisfactory result is not easy to obtain; for the RBF kernel function, the larger the radius gamma of the kernel function is, the easier it is to find the difference between local small samples, but excessive magnification gamma can cause the generalization of the hyperplane to be poor; γ and coef in the multi-layer perceptron kernel satisfy Mercer's theorem (a sufficiently non-essential condition).
The TBM cutter life prediction method based on the data-driven support vector regression comprises the following steps:
1. collecting data of a TBM tool excavation site;
the method mainly takes the subway in Qingdao city as a research object, firstly, data of a TBM tool excavation site are collected, and data recorded in the excavation site are different in source, format and characteristics and properties, so that data integration is necessary, and data of different sources, formats, characteristics and properties are collected to unify data formats. The data collection of a construction site is different from the collection of experimental data, aiming at the sudden situation and data loss in construction, the data is further processed, including two aspects of data elimination and data completion, wherein the data elimination means that if the abnormal data part of a cutter data is less than 5%, only the abnormal data part is eliminated, and if the data loss exceeds 30%, the cutter number data is completely eliminated; the data completion means that partial abnormal data is completed by using a method of artificial filling and mean value replacement. And finally, in order to eliminate dimension influence, normalization processing is carried out on the complete data set, the normalization of the mean value with good stability is selected by comparing a min-max normalization method and a mean value normalization method, and each dimension is normalized into a data set which has a mean value of 0, a variance of 1 and normal distribution.
2. Determining a driving factor influencing the service life of the TBM tool, and establishing a sample data set of the driving factor as a training set;
the service life of the TBM cutter is a result of interaction of a plurality of dynamic factors, the invention combines the excavation current situation of subway in Qingdao city, integrates the prior theoretical research at home and abroad, divides the service life driving factors of the cutter into geological factors (rock grade, texture of rock, underground water and uniaxial compressive strength of rock), artificial factors (mileage, cutter head rotating speed, shield machine utilization rate), machines (cutter head thrust, radius, abrasion amount of blades, torque, damage amount of the cutter and tunneling speed), cutter factors (abrasion coefficient and cutting edge half angle), analyzes a main index system influencing the service life of the cutter by using a grey correlation analysis method, finally determines 8 factors with the maximum correlation degree with the service life of the cutter according to the correlation degree, wherein the factors are the mileage, the rock grade, the cutter head thrust, the abrasion amount of the blades, the cutter head rotating speed, the tunneling speed, the radius and the damage amount of the cutter, the degrees of association were 0.985, 0.8482, 0.8181, 0.9374, 0.86381, 0.7892, 0.75238, 0.81684, respectively. The algorithmic process of the gray correlation analysis method is already well developed, and the algorithmic process is not described in detail herein. And establishing a sample data set of the driving factors as a training set.
3. Constructing a prediction model of a multi-core support vector regression machine, inputting a training set, and training the prediction model so as to determine an optimal parameter, a penalty function C and an insensitive loss function parameter epsilon corresponding to each kernel function;
a prediction model of a support vector regression machine is constructed, and a proper kernel function needs to be selected to improve prediction accuracy. As mentioned in the foregoing principle section, the accuracy of selecting different kernel functions is very different for different problems, so whether to select a proper kernel function becomes a key factor affecting the prediction accuracy. And selecting corresponding kernel functions for specific problems without a unified method, and only comparing the kernel functions after multiple tests, so that in order to improve the prediction accuracy, a prediction model of the multi-kernel support vector regression is firstly constructed, and the selected kernel functions comprise polynomial kernel functions, Gauss radial basis kernel functions and multilayer perceptron kernel functions. After the prediction model is built, inputting a training set, and training the prediction model so as to determine parameters corresponding to each kernel function.
And arranging an information set generated when the shield tunneling machine excavates the left line of the No. 2 subway according to the data acquired in the process of building the subway in Qingdao city. For the convenience of modeling, the life condition of the tool is coded and represented by 0 and 1, wherein 0 represents that the tool is not damaged, and 1 represents that the tool is damaged.
As in table 1 below (part of the original data for tool No. 1 in 2015):
table one: partial raw data of No. 1 tool
The normalization process is shown in Table 1 below (part of the data for tool number 1).
Table two: tool I related driving factor standardization data table
Fig. 1 and 2 are diagrams illustrating the process of prediction by the Gauss radial basis kernel function for the tool number one, wherein fig. 1 is a contour diagram of SVR parameter selection, and fig. 2 is a 3D view of SVR parameter selection. On the premise of selecting the radial basic kernel function, the obtained penalty function C of the No. 1 tool is 0.00680, the kernel radius gamma is 256, and the MSE is 0.0233. Under the condition that the structure is determined, the service life of the No. 1 cutter is predicted, and the accuracy rate is 97.8%. As shown in fig. 3.
The parameter selection of the support vector regression machine has a large influence on the prediction accuracy, and on the basis of selecting the kernel function, it is necessary to establish a proper parameter to improve the prediction accuracy. In the method, in order to overcome the condition that the sample capacity is relatively small, the algorithm effect is tested by fully utilizing the data set, and the sample amount is increased by adopting k-fold cross validation. In machine learning, a data set A is divided into a training set (training set) B and a test set (test set) C, and under the condition that the sample size is insufficient, in order to fully utilize the data set to test the algorithm effect, the data set A is randomly divided into k packets, one packet is used as the test set each time, and the remaining k-1 packets are used as the training set to be trained.
And decoupling each group of kernel function parameters by using a grid search method so as to determine parameters corresponding to the kernel functions, a penalty function C and an insensitive loss function parameter epsilon. The basic principle of the grid search method is that the feasible interval (from small to large) of variable values of each parameter is divided into a series of cells, the computer calculates the combination of variable values of each parameter in sequence, and the corresponding error target value (i.e. the deviation square sum of the actually measured and calculated water quality sequence values) is compared and selected one by one, thus obtaining the minimum target value and the corresponding optimal specific parameter value in the interval. The estimation method can ensure that the obtained search solution is basically a global optimal solution, and can avoid significant errors.
And (3) taking data of the line 2 from the left line 15 years to the 1 month 16 years as a sample training set, and performing learning training by respectively using 3 different kernel functions of a polynomial function, a Gauss radial basis kernel function and a multilayer perceptron kernel function to construct a model for predicting the service life condition of the next working interval of the tool. Under the condition that a kernel function is selected, the influence of the adjusting parameters on the precision is large, and the optimal parameters are searched to the maximum extent by using a libsvm toolbox in MATLAB and combining k-fold cross validation and grid search. The optimal parameters of each kernel function are obtained as follows:
table three: optimal parameters corresponding to each kernel function
4. Determining an optimal kernel function for a predictive model
On one hand, in order to determine the optimal kernel function of the prediction model and on the other hand, in order to verify the feasibility of the prediction model, a multiple regression model is introduced for comparison, the service life condition of the next working interval of the No. 41 cutter is predicted, the service life condition is compared with actual statistics, and 12 different models are tested to obtain the relative error and mean square error between the predicted service life and the actual service life.
According to the installation position and the function of the cutter, the cutter is divided into 3 types, namely 1-8 center cutters, 9-25 positive cutters and 26-41 side cutters. The data are divided into 3 groups, 12 structural bodies are constructed based on different kernel functions to predict the service life of the cutter, and the prediction results of the models are shown in the fourth table:
the prediction accuracy of each model in the table four is opposite to
As is apparent from Table IV, the mean relative error and mean square error of the conventional multiple regression model are higher than those of other models. The average relative error and the mean square error of the prediction models based on the kernel functions have large differences, training results of different types of cutters are comprehensively compared, the model based on the radial basis kernel function is excellent in predicting the service life of the 3 types of cutters, and the final prediction model is established as a cutter service life prediction model of a support vector regression machine based on the RBF kernel function.
5. And taking the sample data set of the driving factor of the tool to be predicted as a prediction sample set, and inputting the prediction sample set into the prediction model to obtain a prediction result.
In order to verify a tool life prediction model of a support vector regression machine based on the RBF kernel function, the data of the right line of the No. 2 line is used for prediction. The prediction is carried out according to the model, and the final result is as the following table five:
table five: accuracy and mean square error of prediction results
The statistical result shows that the overall accuracy rate reaches 94.52%, the mean square error is small, and the prediction effect is good. The validity of the model is verified.
Claims (3)
1. The TBM cutter life prediction method based on the data-driven support vector regression is characterized by comprising the following steps of:
(1) collecting data of a TBM tool excavation site;
(11) processing the collected data, wherein the processing process comprises the following steps:
(111) data integration: centralizing data with different sources, formats, characteristics and properties to unify data formats;
(112) data cleaning: the following measures are taken for the sudden situations and data loss encountered by construction,
a. data elimination: if the abnormal data part of a certain cutter data is less than 5%, only the abnormal data part is removed, and if the data loss exceeds 30%, the cutter data is completely removed;
b. and (3) data completion: completing partial abnormal data by using a manual filling and mean value replacing method;
(113) data mean normalization: normalizing each dimension of the data into a data set with a mean value of 0, a variance of 1 and normal distribution;
(2) determining a driving factor influencing the service life of the TBM tool, and establishing a sample data set of the driving factor as a training set; the determined driving factors influencing the service life of the TBM cutter comprise mileage, rock grade, cutter head thrust, blade abrasion, cutter head rotating speed, tunneling speed, radius and cutter damage;
(3) constructing a prediction model of a multi-core support vector regression machine, inputting a training set, and training the prediction model so as to determine an optimal parameter, a penalty function C and an insensitive loss function parameter epsilon corresponding to each kernel function;
(4) determining an optimal kernel function of the prediction model; increasing the amount of training samples by adopting K-fold cross validation, and mutually decoupling the parameters of each group of kernel functions by utilizing a grid search method so as to determine the optimal parameters corresponding to each kernel function of the prediction model; the basic principle of the grid search method is that the feasible interval of each parameter variable value is divided into a series of cells, the computer calculates the variable value combination of each corresponding parameter in sequence, the corresponding error target values are compared one by one and preferred, thus the minimum target value and the corresponding optimal specific parameter value in the interval are obtained;
(5) and taking the sample data set of the driving factor of the tool to be predicted as a prediction sample set, and inputting the prediction sample set into the prediction model to obtain a prediction result.
2. The TBM tool life prediction method based on the data-driven support vector regression machine according to claim 1, wherein: and (3) determining the driving factors influencing the service life of the TBM tool in the step (2) by processing the collected data through a grey correlation analysis method.
3. The TBM tool life prediction method based on the data-driven support vector regression machine according to claim 1, wherein: the kernel functions of the prediction model of the multi-kernel support vector regression machine constructed in the step (3) comprise polynomial kernel functions, Gauss radial basis kernel functions and multilayer perceptron kernel functions.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611240128.0A CN106778010B (en) | 2016-12-29 | 2016-12-29 | TBM cutter life prediction method based on data-driven support vector regression machine |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611240128.0A CN106778010B (en) | 2016-12-29 | 2016-12-29 | TBM cutter life prediction method based on data-driven support vector regression machine |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106778010A CN106778010A (en) | 2017-05-31 |
| CN106778010B true CN106778010B (en) | 2020-01-14 |
Family
ID=58924051
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611240128.0A Expired - Fee Related CN106778010B (en) | 2016-12-29 | 2016-12-29 | TBM cutter life prediction method based on data-driven support vector regression machine |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106778010B (en) |
Families Citing this family (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107392786B (en) * | 2017-07-11 | 2021-04-16 | 中国矿业大学 | Missing data compensation method for mine fiber bragg grating monitoring system based on support vector machine |
| CN107632523B (en) * | 2017-09-30 | 2019-07-23 | 中铁工程装备集团有限公司 | A kind of hard rock TBM digging control parameter intelligent decision-making technique and system |
| CN108470095B (en) * | 2018-03-07 | 2021-04-02 | 大连理工大学 | TBM (tunnel boring machine) propulsion prediction method based on data-driven radial basis function model |
| CN109446470B (en) * | 2018-11-16 | 2020-07-31 | 厦门大学 | Non-contact detection-based die wear prediction method for spiral bevel gear machining |
| TW202026096A (en) | 2019-01-02 | 2020-07-16 | 財團法人工業技術研究院 | Tool life prediction system and method thereof |
| CN112084459A (en) * | 2019-06-12 | 2020-12-15 | 电计贸易(上海)有限公司 | Method and device for predicting battery charge-discharge cycle life, electronic terminal and storage medium |
| CN110378622B (en) * | 2019-07-31 | 2022-07-05 | 华中科技大学 | Service life assessment method and lease management method for reusable steel component |
| CN111461397A (en) * | 2020-02-26 | 2020-07-28 | 山东浪潮通软信息科技有限公司 | Budget prediction method, equipment and medium based on improved support vector regression |
| CN112182973B (en) * | 2020-09-30 | 2022-11-04 | 山东大学 | TBM (Tunnel boring machine) operation parameter optimization method considering cutter vibration and rock slag geometric information |
| CN112378468A (en) * | 2020-11-13 | 2021-02-19 | 四川泛华航空仪表电器有限公司 | Method for measuring fuel consumption of airplane cruising under condition of small sample training data |
| CN113051683B (en) * | 2021-03-26 | 2023-05-26 | 长安大学 | Numerical control machine tool life prediction method, system, equipment and storage medium |
| CN112966355B (en) * | 2021-03-30 | 2023-01-06 | 西安电子科技大学 | Method for predicting residual service life of shield machine cutter based on deep learning |
| CN113222273B (en) * | 2021-05-26 | 2022-06-07 | 中铁十八局集团有限公司 | TBM tunneling rate prediction method based on GWO-FW-MKL-SVR algorithm |
| CN113779694A (en) * | 2021-08-23 | 2021-12-10 | 武汉理工大学 | Manufacturing process reliability modeling method and device based on support vector machine |
| CN114781278B (en) * | 2022-06-17 | 2022-09-27 | 天津理工大学 | Electromechanical equipment service life prediction method and system based on data driving |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103217280A (en) * | 2013-03-18 | 2013-07-24 | 西安交通大学 | Multivariable support vector machine prediction method for aero-engine rotor residual life |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8201646B2 (en) * | 2009-11-20 | 2012-06-19 | Edward Vezirian | Method and apparatus for a true geometry, durable rotating drill bit |
-
2016
- 2016-12-29 CN CN201611240128.0A patent/CN106778010B/en not_active Expired - Fee Related
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103217280A (en) * | 2013-03-18 | 2013-07-24 | 西安交通大学 | Multivariable support vector machine prediction method for aero-engine rotor residual life |
Non-Patent Citations (2)
| Title |
|---|
| 全断面岩石掘进机盘形滚刀寿命预测及管理软件开发;刘吉洋;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》;20091115(第11期);第2.4.2节,3.1节 * |
| 基于支持向量回归机的刀具寿命预测;侍红岩 等;《工具技术》;20151120;第49卷(第11期);第49页第4节 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106778010A (en) | 2017-05-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106778010B (en) | TBM cutter life prediction method based on data-driven support vector regression machine | |
| CN109933577B (en) | Tunnel tunneling prediction method and system based on TBM rock-machine parameter dynamic interaction mechanism | |
| US11307327B2 (en) | Method and system for online monitoring and optimization of mining and mineral processing operations | |
| Koopialipoor et al. | Predicting tunnel boring machine performance through a new model based on the group method of data handling | |
| Aryafar et al. | Application of metaheuristic algorithms to optimal clustering of sawing machine vibration | |
| CN108107838A (en) | A kind of numerical control equipment tool wear monitoring method based on cloud knowledge base and machine learning | |
| KR101967978B1 (en) | Apparatus for predicting net penetration rate of shield tunnel boring machine and method thereof | |
| CN108470095B (en) | TBM (tunnel boring machine) propulsion prediction method based on data-driven radial basis function model | |
| CN109241627A (en) | The dynamic shoring method of probability hierarchical and the device of Automated Design supporting scheme | |
| Mikaeil et al. | An extended multifactorial fuzzy prediction of hard rock TBM penetrability | |
| Zhang et al. | Geological Type Recognition by Machine Learning on In‐Situ Data of EPB Tunnel Boring Machines | |
| CN115017791A (en) | Tunnel surrounding rock grade identification method and device | |
| Li et al. | Prediction of TBM cutterhead speed and penetration rate for high-efficiency excavation of hard rock tunnel using CNN-LSTM model with construction big data | |
| Zhou et al. | Novel approach to evaluate rock mass fragmentation in block caving using unascertained measurement model and information entropy with flexible credible identification criterion | |
| CN116245020A (en) | Cutter head abrasion and energy consumption optimization method and system based on shield tunneling machine | |
| CN115239108B (en) | Weak broken surrounding rock sensing method based on TBM real-time broken rock data | |
| CN114818493A (en) | Method for quantitatively evaluating integrity degree of tunnel rock mass | |
| CN115758515A (en) | TBM tunnel unfavorable geological section intelligent support decision-making method | |
| CN113158562B (en) | TBM rock machine mapping construction method and system based on physical constraint and data mining | |
| CN114066271A (en) | Tunnel water inrush disaster monitoring and management system | |
| Chunliang et al. | The data mining method based on support vector machine applied to predict tool life of TBM | |
| Wang et al. | Sensitivity of TBM’s performance to structural, control and geological parameters under different prediction models | |
| CN113158561A (en) | TBM operation parameter optimization method and system suitable for various rock mass conditions | |
| Ding et al. | Quantitative estimation of TBM disc cutter wear from in-situ parameters by optimization algorithm improved back-propagation neural network: A case study of a metro tunnel in Guangzhou, China | |
| CN113642082B (en) | A-CNN method and system for predicting TBM utilization rate |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20200114 Termination date: 20211229 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |