Case study xây dựng hệ cơ sở tri thức ứng dụng, TBM case study (2)

Case Study for Knowledge-based system Research on Tunnel Boring Machine TBM Utilization and Prediction Performance under Complex Ground Conditions in Tunnel Projects Feb 2012 Hai V.. P

Case Study for Knowledge-based system

Research on Tunnel Boring Machine (TBM)

Utilization and Prediction Performance under

Complex Ground Conditions in Tunnel Projects

Feb 2012 Hai V Pham

Soft Intelligence Lab, Ritsumeikan University

Email: hai@spice.ci.ritsumei.ac.jp

FUJITA Yuji System and Development Research Dept.,

Enzan Koubou CO., LTD

Introduction

• Ｔunneling in difficult ground conditions is one of the most

challenging tasks in tunnel engineering

• Tunnel Boring Machine (TBM) applications have been

implemented in tunneling projects to predict accurately

TBM performance and reduce cutter costs

• Geological effects and operational states of TBM

machine performance prediction are closely related to

predict TBM performance

• Prediction of the TBM utilization performance, especially

in long-term projects, has become very important,

considering the machine parameters and ground

conditions

Research backgrounds

Tunnel engineers and experts need to make realistic estimates

of TBM performance as a basis for project planning, choice of

tunneling methods and scheduling

TBM utilization and prediction

performance

! "#

$ " $ &'"

(

) (

! & $

( * ( * ( * ( *

+ , ''''

Factors influencing to TBM

performance

• The key factors in TBM applications to any tunneling

project, which classify into categories of factors,

influencing TBM performance as follows: operational

parameters, machine specifications, rock properties,

geological conditions, and cutting geometry

Utilization Performance and

Performance Prediction for TBM

• The main TBM utilization performance is as follows:

• Instantaneous penetration rates (PR) measured in

mm/rev or m/hr for the time of TBM spends cutting

ground

• TBM utilization (U): proportion of time spent cutting

expressed as an average of the total available working

time (T)

• Cutter rate consumption and cutter costs., Disk force

penetration index

Operational TBM Parameters

Penetration Rate (PR)

$

(

/ 0 1

( /

TBM Parameters

Advance Rate

(

(

1

( ) ( ) ( ) ( ) ( 2 ( ( 1

(

Solutions for Enzan Koubou

System Development

! "#

$ % "

$ &'" (

) (

! & $$ $ $

。

3 % ( %

3 % ( %

3 % ( %

3 % ( %

( * ( * ( * ( *

+ , ''''

Research Plan

• August – September, 2011 (1 month): Research

surveys, detail plan, and System solutions

• October- November, 2011 (1,5 months): A

proposal of project for optimal TBM utilization.

• November- December, 2011 (1 month): A

proposal of project for TBM performance

prediction.

• December, 2011 – February, 2012 (2 months):

Integrated systems and system evaluation.

Proposed model

0 % , 4 4

0 % , 4 4

0 % , 4 4

0 % , 4 4

Tunnel Boring Machine (TBM) in tunnel projects

1

+

2

*

3

3

3

3

&

Fuzzy Reasoning Evaluation model

for optimal input parameters

(

$ '&5 0 4 4

$ 5+

Solution in detail

! "# $

" $ &'" (

) (

! & $$ $ $

。

3 % (

3 % 3 % ( (

3 % (

( * ( ( * * ( *

+ , ''''

Sample results of Penetration rate prediction

from http://www.tbmexchange.com/

Sample: PR Prediction

Example TBM performance Prediction

Solution in detail

• Enzan Koubou has been currently done successfully

TBM support systems with advanced systems

• In order to solve full solutions for tunneling, the company

should be established new applications which focus on

TBM Utilization Evaluation, Performance Prediction and

Rock Mass classification Evaluation

• Furthermore, Penetration Rate Prediction is also

important in TBM performance evaluation We hope to

give PR standard namely Enzan PR in the future

• Intelligent system will apply for selection of optimal

projects when tunnel project may have several solutions

In addition, business intelligent needs to find potential

customers in Asia and the world to extend the partners in

global

• Written science and engineering research papers and

publications are also to improve expertise reputations of

the company in the near future

TBM Data sets from Asian Countries

NN training and testing Errors

NN training and testing Errors

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

1 10 19 28 37 46 55 64 73 82 91 100 109 118 127 136 145

TrainingError TestingError

Simulation Results

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101

Y0_from_NN Y0_Desired

Different Result in Simulation

0

0.2

0.4

0.6

0.8

1

1.2

1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101

Y0_Desired Y0_from_NN

Optimizing Factor Structure

Fuzzy Reasoning Experts

Results in detail

Optimized Open/ Shield TBM

Factor Results

EPB performance Prediction

Surface pressure (SP)

.

1

1

-(

1

(

1

(

&

Data Sets from Bangkok projects

Surface pressure (SP) Prediction

0

0.5

1

1.5

2

2.5

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Y0_from_NN Y0_Desired

Surface pressure (SP) Prediction

in diffirent simulation

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1 2 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Y0_Desired Y0_from_NN

Simulation results

0

20

40

60

80

100

120

140

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97

Y0_from_NN Y0_Desired

Disaster risks in tunneling

Low levels of Tunnel Disasters in

simulation results

Emergency disasters in simulations

Views in detail of tunneling diasters

Disaster risks in day t

Disaster risks in day t+1

If we have 19 days

Combined results in simulations

Conclusion & Future work

• The proposed approach can be predicated

on real-timeTBM performance

• Hybrid NN models is good for

improvement of the system performance

• Enzan Koubou has been currently done

successfully TBM support systems and

Intelligent System with advanced systems

• For any reference, please visit us

http://www.enzan-k.com

Publications in this research

• [1 ] Hai V Pham, Fujita Yuji and Kamei Katsurari, Neural

Networks Integrated with Fuzzy Reasoning Evaluation

Model for TBM Performance Prediction in Uncertain

Underground Conditions, To appear in Proceedings of

the 2012 International Conference on Embedded

Systems and Intelligent Technology (ICESIT 2012),

January 2012, Nara, Japan

• [2] Hai V Pham, Fujita Y and Kamei K., Hybrid Artificial

Neural Networks for TBM Utilization and Performance

Prediction in Complex Underground Conditions, To

Research progress

6-7- ,

2

7- $

% 5.

# $ 8

4 - ( (

$

-96 :

4 # ' (

Future research

• To write 1 or 2 journal publication

• To open a business in Vietnam

• To develop applications to real systems

Q & A Thank you for your attentions!