Elsevier

Measurement

Volume 126, October 2018, Pages 46-57
Measurement

State-of-the-art predictive modeling of TBM performance in changing geological conditions through gene expression programming

https://doi.org/10.1016/j.measurement.2018.05.049Get rights and content

Highlights

  • The new attempt to hard rock TBM performance prediction.

  • The use of GEP as one of the best approaches for complex nonlinear modeling.

  • The highest accuracy of predictive modeling of TBM performance up until now.

  • Comprehensive practical verification of the models using models from the literature.

Abstract

Hard rock TBM performance prediction is of great interest to the tunneling community on account of its importance in time and cost risk management of underground projects. Continuous development of new empirical models in recent decades reveals the importance of accurate prediction of this factor in diverse ground and machine conditions. The great number of different parameters influencing TBM performance and the high variability linked to specific field conditions cause the problem to be very complex. Gene expression programming (GEP) models, a robust variant of genetic programming, are developed in this study to correlate hard rock TBM performance with routine ground properties for project design applications. The developed models are compared with those from statistical and soft computing-based models in the literature. Overall, GEP models show good performance and are proven to be much better than the previous models. The proposed models of this study can be remarked as an ultimate stage to one decade of researchers’ attempts to improve the accuracy of predictive equations developed through a well-known database of TBM performance in one of the most complex tunneling projects in the world.

Introduction

In our time, tunnel boring machines (TBMs) are frequently utilized in excavation of long infrastructural tunneling projects in a wide range of geological conditions. It is to ensure that the construction is completed with a reasonable time and cost, hence making the prediction tasks more crucial for planning and risk management purposes [96]. Among several technical and managerial parameters, TBM performance is of vital importance since it highly affects the time and cost of any tunneling project.

During the last four decades, several prediction models were proposed to estimate hard rock TBM performance using tunnel path geological parameters as inputs. Main output parameters suggested by prediction models are the rate of penetration (ROP), that is the penetration per revolution or per unit of time (this index is believed to be the most significant representative of the machine performance together with the other indicators; see, e.g., [52], the utilization factor (UF) (percentage of time in which the machine is excavating) and more recently the field penetration index (FPI) (the ratio of thrust per cutter and penetration rate per cutterhead revolution). Basically, two major approaches have already been followed to develop predictive models: (i) Analytical models which estimate the cutting force by studying the rock fragmentation process with indentation of mechanical tools. The most well-known developed models in this category are the Gehring’s model [34] and the Colorado School of Mines (CSM) model [80], [81]. The analytical models usually combine intact rock and the cutter characteristics, and they do not take into account rock mass structural properties that already proved to essentially contribute to TBM performance, such as rock discontinuities frequency and orientation; (ii) Empirical models based on databases of records of previous projects. Due to multiple number of influencing parameters on TBM performance and variability of field conditions, numerous models have been developed in this category. However, the most widespread empirical models are Norwegian University of Sciences and Technology (NTNU) model developed by Bruland [17] and recently updated by Macias [62], and the QTBM [11]. Other models have also been developed ranging from single factor models [25], [41] to multi-factor models [2], [4], [8], [16], [20], [21], [22], [26], [30], [31], [40], [45], [46], [47], [48], [49], [51], [55], [59], [61], [62], [64], [69], [71], [72], [74], [75], [76], [77], [78], [79], [82], [83], [84], [89], [90], [91], [92], [93], [94], [98], [99], [100], [101]. In addition, some researchers have tried to apply soft-computing solutions to the problem of TBM performance prediction mostly based on artificial intelligence (AI) techniques [1], [5], [6], [7], [15], [33], [37], [38], [63], [65], [67], [68], [87], [88] which resulted in development of kinds of black boxes without a practical potential for utilization in future applications, hence referred to as only “predictions” instead of model development attempts.

The present study aims to the prediction of TBM performance in changing rock mass environment with a relatively higher confidence compared to the previous similar models. Therefore, gene expression programming (GEP) [27], as an extension to genetic algorithm (GA) and genetic programming (GP) [10], [58], is utilized as an AI technique for this purpose. It is highly capable of developing functional relationships for predicting a specific output using multiple related inputs. In fact, based on the numerical experiments by Ferreira [29], the GEP approach is considered as an efficient alternative to traditional GP. Recently, there have been several scientific efforts directed at applying GEP to various civil and geotechnical engineering problems such as prediction of soil–water characteristic curve [53], estimation of compressive and tensile strength of different rock types [7], [13], [18], [23], [73], modeling of plate load test moduli of soil [70], predicting settlement induced by tunneling [3], estimation of air flow in a single rock joint [57], modeling of shear strength of concrete columns [14] and so on.

In this paper, the GEP approach is utilized to develop predictive equations for hard rock TBM performance with routine ground properties of tunnel path including intact rock and rock mass properties. The main aim of this recent modeling attempt is to increase the accuracy of estimations. Therefore, for the sake of comparison, a well-known database of TBM performance (the Queens Water Conveyance Tunnel #3, Stage 2, New York City, USA), which has previously been used to develop equations by other authors [2], [89], [90], [91], [92], [93], is selected and utilized in the analyses. Performance of proposed models of this research is directly compared to those of other researchers and the models based on mentioned database are eventually ranked.

Section snippets

Methodology

Genetic Algorithm (GA) was first presented by Holand [50] as an abstraction of biological evolution in the form of a stochastic optimization technique. The GA has then been further developed by Goldberg [39]. Chromosomes as symbolic strings of fixed length are the unique solutions in GA; they are usually evaluated using a fitness function and the algorithm stops once the output requirement is met [19]. Koza [58] presented the Genetic Programming (GP) which is derived from the extended version

Database of TBM performance

In this research, predictive modeling of hard rock TBM performance is to be conducted considering multiple influencing characteristics of the ground. To this end, a database compiled by Yagiz [89] from a hard rock mechanized tunneling project (the Queens Water Conveyance Tunnel #3, Stage 2) is employed which has already been widely used by other researchers for this purpose. Among other reasons for using this database such as the quality of measurements and good sort of influencing parameters,

Verification and discussion of the results

In order to have an idea about the predictive power of the developed GEP models, their performance indicators were compared with that of powerful statistical and soft computing-based models of TBM performance analysis. For the sake of a practical verification, models were chosen from the literature which have developed using the same database of this paper (The Queens Tunnel, No. 3, stage 2, NY, USA) with the exact number of observations (151 data points). This way, precision of the previous

Conclusions

Gene Expression Programming (GEP) is one of the best approaches for complex, nonlinear modeling. As the newest type of circulative algorithm, GEP gets the most of the capabilities of both GA and GP. This approach was utilized in this research to deal with complexity of modeling hard rock TBM performance in changing geological conditions in tunnel path. Enormous number of runs was conducted within an extensive program of modeling procedure in order to achieve the best models ever to describe TBM

Acknowledgements

The authors wish to express their deep sense of gratitude to Prof. Dr. Hanifi Copur, Head of Department of Mining Engineering, Istanbul Technical University (Turkey), for his meticulous suggestions and astute criticism in the revision phase of this article.

References (102)

  • H. Copur et al.

    Predicting performance of EPB TBMs by using a stochastic model implemented into a deterministic model

    Tunn. Undergr. Space Technol.

    (2014)
  • A. Delisio et al.

    A new model for TBM performance prediction in blocky rock conditions

    Tunn. Undergr. Space Technol.

    (2014)
  • A. Delisio et al.

    Analysis and prediction of TBM performance in blocky rock conditions at the Lötschberg Base Tunnel

    Tunn. Undergr. Space Technol.

    (2013)
  • I. Ebtehaj et al.

    Gene expression programming to predict the discharge coefficient in rectangular side weirs

    Appl. Soft Comput.

    (2015)
  • E. Farrokh et al.

    Study of various models for estimation of penetration rate of hard rock TBMs

    Tunn. Undergr. Space Technol.

    (2012)
  • O. Frough et al.

    An application of rock engineering systems for estimating TBM downtimes

    Eng. Geol.

    (2013)
  • A.H. Gandomi et al.

    A new prediction model for the load capacity of castellated steel beams

    J. Constr. Steel Res.

    (2011)
  • Q.M. Gong et al.

    Development of a rock mass characteristics model for TBM penetration rate prediction

    Int. J. Rock Mech. Min. Sci.

    (2009)
  • H. Güllü

    Prediction of peak ground acceleration by genetic expression programming and regression: a comparison using likelihood-based measure

    Eng. Geol.

    (2012)
  • H. Güllü

    Function finding via genetic expression programming for strength and elastic properties of clay treated with bottom ash

    Eng. Appl. Artif. Intell.

    (2014)
  • J. Hassanpour et al.

    Evaluation of common TBM performance prediction models based on field data from the second lot of Zagros water conveyance tunnel (ZWCT2)

    Tunn. Undergr. Space Technol.

    (2016)
  • J. Hassanpour et al.

    A new hard rock TBM performance prediction model for project planning

    Tunn. Undergr. Space Technol.

    (2011)
  • S.M.E. Jalali et al.

    Development of a new laboratory apparatus for the examination of the rotary-percussive penetration in tunnel boring machines

    Tunn. Undergr. Space Technol.

    (2013)
  • A. Keshavarz et al.

    New Gene Expression Programming models for normalized shear modulus and damping ratio of sands

    Eng. Appl. Artif. Intell.

    (2015)
  • J. Khademi Hamidi et al.

    Performance prediction of hard rock TBM using Rock Mass Rating (RMR) system

    Tunn. Undergr. Space Technol.

    (2010)
  • Q. Liu et al.

    A case study of TBM performance prediction using a Chinese rock mass classification system – Hydropower Classification (HC) method

    Tunn. Undergr. Space Technol.

    (2017)
  • S. Mahdevari et al.

    A support vector regression model for predicting tunnel boring machine penetration rates

    Int. J. Rock Mech. Min. Sci.

    (2014)
  • R. Mikaeil et al.

    Multifactorial fuzzy approach to the penetrability classification of TBM in hard rock conditions

    Tunn. Undergr. Space Technol.

    (2009)
  • A. Mollahasani et al.

    Empirical modeling of plate load test moduli of soil via gene expression programming

    Comput. Geotech.

    (2011)
  • M. Namli et al.

    A model to predict daily advance rates of EPB-TBMs in a complex geology in Istanbul

    Tunn. Undergr. Space Technol.

    (2017)
  • A. Ozbek et al.

    Estimating uniaxial compressive strength of rocks using genetic expression programming

    J. Rock Mech. Geotech. Eng.

    (2013)
  • E. Paltrinieri et al.

    Analysis and estimation of gripper TBM performances in highly fractured and faulted rocks

    Tunn. Undergr. Space Technol.

    (2016)
  • M. Rasouli Maleki

    Rock Joint Rate (RJR); a new method for performance prediction of tunnel boring machines (TBMs) in hard rocks

    Tunn. Undergr. Space Technol.

    (2018)
  • A. Salimi et al.

    Application of non-linear regression analysis and artificial intelligence algorithms for performance prediction of hard rock TBMs

    Tunn. Undergr. Space Technol.

    (2016)
  • M. Sapigni et al.

    TBM performance estimation using rock mass classifications

    Int. J. Rock Mech. Min. Sci.

    (2002)
  • M. Saridemir

    Effect of specimen size and shape on compressive strength of concrete containing fly ash: application of genetic programming for design

    Mater. Des.

    (2014)
  • S. Yagiz

    Utilizing rock mass properties for predicting TBM performance in hard rock condition

    Tunn. Undergr. Space Technol.

    (2008)
  • S. Yagiz et al.

    Application of two non-linear prediction tools to the estimation of tunnel boring machine performance

    Eng. Appl. Artif. Intell.

    (2009)
  • S. Yagiz et al.

    Prediction of hard rock TBM penetration rate using particle swarm optimization

    Int. J. Rock Mech. Min. Sci.

    (2011)
  • S. Yagiz et al.

    Application of various optimization techniques and comparison of their performances for predicting TBM penetration rate in rock mass

    Int. J. Rock Mech. Min. Sci.

    (2015)
  • M. Alber, Prediction of penetration and utilization for hard rock TBMs, in: ISRM International Symposium Eurock ’96,...
  • D.J. Armaghani et al.

    Uniaxial compressive strength prediction through a new technique based on gene expression programming

    Neural Comput. Appl.

    (2017)
  • H.M. Azamathulla et al.

    Gene-expression programming for the development of a stage-discharge curve of the Pahang river

    Water Resour. Manage.

    (2011)
  • W. Banzhaf et al.

    Genetic Programming: An Introduction

    (1998)
  • N. Barton

    TBM Tunnelling in Jointed and Faulted Rock

    (2000)
  • C.A. Baskerville et al.

    The separation of the Hartland Formation and Ravenswood Granodiorite from the Fordham Gneiss at Cameron’s Line in the New York City area

    Northeast. Geol.

    (1989)
  • A. Bruland, Hard rock Tunnel Boring, Norwegian Institute of Technology (NTNU), Trondheim, Norway,...
  • H. Çanakcı et al.

    Prediction of compressive and tensile strength of Gaziantep basalts via neural networks and gene expression programming

    Neural Comput. Appl.

    (2009)
  • D.L. Chambers

    The Practical Handbook of Genetic Algorithms Applications

    (2001)
  • S.R. Dindarloo et al.

    Estimating the unconfined compressive strength of carbonate rocks using gene expression programming

    Eur. J. Sci. Res.

    (2015)
  • Cited by (56)

    • Measurement and enhancing prediction of EPBM torque using actual Machine data

      2023, Measurement: Journal of the International Measurement Confederation
    View all citing articles on Scopus
    View full text