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

9/17/2013 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 9/17/2013 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 • Delays for tunneling through geological features whether this measureable in days or weeks, as opposed to hours • Advance rate = P x U x T (m/week) 9/17/2013 Operational TBM Parameters $ / Penetration Rate (PR) ( ( / - / TBM Parameters ( ( Advance Rate ( ( ) ( ( ( Solutions for Enzan Koubou System Development ) ( % ( ! & + $ % &'" % 。 * $ ! $ ( "# , ' - " ( 9/17/2013 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 Tunnel Boring Machine (TBM) in tunnel projects % , 4 & * + Fuzzy Reasoning Evaluation model for optimal input parameters ( $ % '&5 5+ $ 4 9/17/2013 Solution in detail ) ( % ( ! & % 。 * $ ! " ( $ "# &'" $ ( % + , ' - Sample results of Penetration rate prediction from http://www.tbmexchange.com/ Sample: PR Prediction • Future 9/17/2013 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 9/17/2013 NN training and testing Errors NN training and testing Errors 0.16 0.14 0.12 0.1 TrainingError TestingError 0.08 0.06 0.04 0.02 10 19 28 37 46 55 64 73 82 91 100 109 118 127 136 145 Simulation Results 1.6 1.4 1.2 Y0_from_NN 0.8 Y0_Desired 0.6 0.4 0.2 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101 9/17/2013 Different Result in Simulation 1.2 0.8 Y0_Desired Y0_from_NN 0.6 0.4 0.2 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101 9/17/2013 Optimizing Factor Structure Fuzzy Reasoning Experts Results in detail 9/17/2013 Optimized Open/ Shield TBM Factor Results EPB performance Prediction / & ( Surface pressure (SP) ( + / - ( ( Data Sets from Bangkok projects 10 9/17/2013 Surface pressure (SP) Prediction 2.5 1.5 Y0_from_NN Y0_Desired 0.5 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Surface pressure (SP) Prediction in diffirent simulation 1.4 1.2 0.8 Y0_Desired Y0_from_NN 0.6 0.4 0.2 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Simulation results 140 120 100 80 Y0_from_NN Y0_Desired 60 40 20 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97 11 9/17/2013 Disaster risks in tunneling Low levels of Tunnel Disasters in simulation results Emergency disasters in simulations 12 9/17/2013 Views in detail of tunneling diasters Disaster risks in day t Disaster risks in day t+1 13 9/17/2013 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 appear in Proceedings of the 2011 IEEE International Symposium on System Integration, IEEE, pp.1149-1154, Kyoto, Japan, December 2011 14 9/17/2013 Research progress 96 : 6- # ' ( 7- , % / $ # $ ( ( $ 7- - $ Future research • To write or journal publication • To open a business in Vietnam • To develop applications to real systems Q&A Thank you for your attentions! 15 ... 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...9/17/2013 TBM utilization and prediction performance ) ( ( ! & $ % " * $ ! "# $ &'" ( + , ' - Factors influencing to TBM performance • The key factors in TBM applications to any... from http://www.tbmexchange.com/ Sample: PR Prediction • Future 9/17/2013 Example TBM performance Prediction Solution in detail • Enzan Koubou has been currently done successfully TBM support