Actual performance analysis of a double shield TBM through sedimentary and low to medium grade metamorphic rocks of Ghomrood water conveyance tunnel project (lots 3 and 4)

Abstract

In this paper, after a brief description of engineering geological properties of Ghomrood water conveyance tunnel project (lots 3 and 4 with a total length of 18 km), the actual performance of the machine during the excavation of tunnel through sedimentary and low to medium grade metamorphic rocks has been analyzed. Subsequently, using data collected during pre-construction and construction phases of the project, the relationships between rock mass properties and measured machine performance parameters have been investigated and proposed as a prediction model which can be used in future projects in which similar geological conditions are encountered. Results of analyses indicated that there are strong relationships between TBM performance and engineering geological conditions of the tunnel alignment.

Introduction

Tunnel boring machine (TBM) performance has a great impact on tunneling projects’ schedule and cost. The most important factors that influence on machine performance and particularly on penetration rate are engineering properties of geologic materials, operating parameters, and technical specifications of the employed machine.

Nowadays, with regard to the widespread use of mechanized technology in tunneling projects in different geological conditions and since this technology is still young in many countries, doing research on different aspects of the subject, especially on TBM performance prediction and its controlling factors, is a necessity for tunneling industry.

During the last decades, many researchers have worked on developing various models for predicting the performance of machines in hard rock. Research works done by Rostami (1997), Gehring (1995), Bruland (1998), Barton (2000), Yagiz (2002, 2008), Ramezanzadeh (2005), Gong and Zhao (2009), Hassanpour (2009), Hassanpour et al. (2009, 2010, 2011, 2014, 2015, 2016), Delisio and Zhao (2014), Macias (2016), etc. are some of the most important examples of such studies. Most of the developed models have major limitations and cannot be used directly in projects with different geological conditions and machine types. Inconsistency of the models’ databases with geological conditions of projects and unusual input parameters are two important limitations of the existing models.

Therefore, analysis and interpretation of actual data collected from constructed projects with the aim of establishing relationships between geological conditions of tunnel alignment and machine performance parameters can be a useful experience for clients and contractors in their projects with similar conditions. Experiences show that due to the nature of rock engineering projects, employing local empirical models results in more accurate and reliable predictions.

It must be noted that the occurrence of chipping or grinding and the extent of crack propagation in the rock are key processes in cutter penetration. These processes are very dependent on rock mass characteristics and machine specifications. So, one way for exact prediction of cutter penetration in rock is the identification of all rock mass characteristics and their partial role in cutting mechanism and putting them in a prediction model as input parameters. The other way is wise classification of ground behavior based on cutting mechanism that occurs beneath disc cutters and development of simple equations for each special class of ground (rock mass) behavior.

It is believed that the advantages of the second method are greater because the first method requires performing several tests for developing the model, and many input parameters must be considered in the model (parameters that may show diverse effects in different geological conditions). The second method is more appropriate if the rock mass classes and geological conditions are correctly selected, but its limitation is the need to have access to the information of different projects that are constructed in different geological conditions. Fortunately, information is now available on various tunnels being constructed in Iran as part of a large research project at the University of Tehran. This research is also part of the ongoing project.

This study focuses on low-grade metamorphic rocks with a slightly developed foliation. These types of rocks cover very wide areas in different parts of Iran, including a very important geological zone, namely Sanandaj–Sirjan metamorphic belt, with 100–150 km wide and more than 1500 km length in central parts of Iran. In previous researches conducted by the first author to develop models, different geological conditions and rock mass classes, such as strong to very strong igneous and metamorphic rocks (where the grinding phenomenon is dominant rather than chipping), calcareous and argillaceous sedimentary rocks (which show plastic behaviors during cutting process) and jointed pyroclastic rocks have been studied.

In this paper, after a brief description of actual machine performance and geological conditions of the Ghomrood tunnel project (lots 3 and 4), an effort has been made to develop an empirical relationship between rock mass properties and machine performance parameters with analyzing geological and actual machine operational data. This empirical model will be very useful for predicting TBM performance in future projects which will be constructed in similar geological conditions.

Project description

Ghomrood water conveyance tunnel is a part of water conveying system from Dez river basin to Ghomrood river basin and has been designed to transfer 23 m³/s of water to supply drinking water of some cities in central Iran (Fig. 1). This tunnel with a total length of 36 km and a boring diameter of 4.525 m (and finished diameter of 3.8 m) is a gravity flow tunnel and during construction was divided into 4 equal lots (lots 1 to 4) each 9 km in length. Two northern lots (lots 3 and 4) and parts of lot 2 of this tunnel (totally 24.5 km) were bored and lined (with pre-cast concrete segments with hexagonal arrangement and thickness of 25 cm) using a double shield machine (manufactured by Wirth Company, Germany). The remaining part of the tunnel (lot 1 and parts of lot 2, with a total length of about 11 km), after encountering different adverse hydrogeological and geological conditions and change in construction method, were constructed using an EPB-Hard rock machine (supplied by Herrenknecht, Germany) from the southern portal. Two machines were disassembled in a small cavern and pulled out from the main tunnel using one of the excavated access tunnels.

Fig. 1

The geographical situation of Ghomrood project (Hassanpour 2018)

Geological and engineering geological setting

Ghomrood water conveyance tunnel is located in Sanandaj–Sirjan zone (SSZ) as one of the main geological zones in Iran. This zone is a metamorphic belt and extends almost 1500 km from the northwest (Sanandaj) to southeast (Sirjan) with a width of 150–200 km, parallel to the Zagros Fold Thrust belt. In lithological point of view, tunnel alignment mainly consists of alternation of Jurassic sandstones and shales (which in main parts of the area have been metamorphosed to alternations of metasandstones, slates and phyllites, graphite schists, and quartz schists) along with quartzite veins and Cretaceous carbonate rocks. Figure 2 shows the geological cross-section along the tunnel. As shown in Fig. 2, the total length of the tunnel has passed through sedimentary and low to medium grade metamorphic rocks. Table 1 shows the stratigraphic units and their equivalent engineering geological units.

Fig. 2

Engineering geological cross-section along the tunnel alignment (Hassanpour 2018)

Table 1 Stratigraphic and equivalent engineering geological units identified along the tunnel (Hassanpour 2018)

As can be seen in Table 1, the tunnel length has been divided into 7 engineering geological types (ET) with almost constant rock mass condition in mechanized tunneling. These engineering geological types (ET) which are almost uniform and homogenous in their physical and mechanical properties are subdivisions of four less homogenous lithological types (LT) identified along the tunnel. To obtain the required engineering geological data of each geological unit, a detailed site investigation was carried out to both identify and quantify engineering characteristics (SOI consulting engineers, 2004-2008a, b). Intact rock properties along with rock mass characteristics, both measured or estimated through field or laboratory investigations, are two sets of data that were used to explain different engineering geological units.

To characterize the intact rock, a series of common rock mechanic tests were conducted on a large number of samples collected from different locations such as boreholes, outcrops, and excavated muck. The results of these tests are summarized in Table 2.

Table 2 Typical ranges of geomechanical properties of engineering geological types identified along the tunnel

Additionally, since rock mass characteristics such as structure and the degree of fracturing are also effective in important processes like tool wear, empirical rock mass classification systems such as RQD, RMR, GSI, and Q-system were assessed both from outcrops and boreholes during the pre-construction phase and geological back-mapping of the tunnel face during the construction phase. Rock mass specifications along the tunnel are presented in Table 2.

Structurally, rock masses outcropped along the tunnel alignment and observed in tunnel faces and drilled boreholes show a variation from schistose and foliated rocks (shales, slates, and schists) to blocky and thick-bedded limestones and sometimes thick quartzite veins. Pictures presented in Fig. 3a–f show some of the engineering geological units in surface outcrops, drilled boreholes, and tunnel walls. Shear zones formed in the vicinity of local faults are of the significant structures encountered in the tunnel route which especially have important effects on machine performance.

Fig. 3

Pictures of different engineering geological types: a shale (Sh) in core box of BH7; b outcrop of slates and phyllite (Sph); c sandstone beds (Sn) in tunnel wall behind the shield; d graphite schists (Sg) with thin quartz veins; e quartz schists; f massive limestone (LM) in core box of BH6

According to the data from the borehole drilled along the tunnel, the tunnel level was lower than the groundwater level throughout the entire length of the tunnel, and the water head above the tunnel varied from 50 to 200 m. Due to the low permeability of most of the rocks of the route, including clay rocks (shales) and their metamorphic rocks (slates, phyllites, etc.), the water entering the tunnel was low and the groundwater conditions were “Damp” to “Wet” throughout most of the route. Only in the part of the tunnel that passes through the limestones (about 800 m of tunnel length, as shown in Fig. 2), due to the presence of persistent and open discontinuities, the usual groundwater condition in the tunnel was “dripping” and in some rare sections “flow”. In this part of the tunnel, the total volume of water entering the tunnel never exceeded 30 l/s.

Machine specifications

A double shield (Telescopic) machine manufactured by Wirth Company has been used in this project. In this machine, a flat cutterhead equipped with disc cutters with diameter of 432 mm (17 in.) and the edge width of 19 mm, with load capacity up to 250 kN, has been used. Other features of this machine are listed in Table 3. A view of the machine is shown in Fig. 4.

Table 3 Technical specifications of machine used in Ghomrood project (lots 3 and 4)

Fig. 4

A view of the machine and its cutterhead used for completing the northern section of Ghomrood project

Actual machine performance

Figure 5a–c demonstrate a weekly average of changes in parameters related to machine performance. As can be seen in diagram (a) of this figure, 18 km of this tunnel has been bored and lined during 217 weeks with average advance rate of 83 m/week based on operating times (total project runtime, taking into account long term stops) and 120 m/week based on working times (actually the times that tunneling activity was carried out at project). In this tunnel, the average of weekly penetration rate has fluctuated between 0.7 and 6.9 m/h (Fig. 5b). Also, according to Fig. 5c maximum weekly utilization factor is about 32% and its cumulative averages based on working and operating times are determined 18.9% and 10.0%, respectively. As can be seen in Fig. 5c, the tendency of changes and values of working utilization factor during normal conditions after a few weeks from the beginning of boring has reached about 20%. This number is important in the sense that it represents the range of utilization factors of country similar tunneling projects under normal conditions (in terms of geological conditions). The impact of adverse geological conditions on this machine’s performance has been shown on the operational utilization factor diagram. As can be seen, the operational utilization factor of the machine has drastically reduced after encountering unfavorable geological conditions to about 5%. This point is also important in this regard that if the geological conditions of the tunnel route are not correctly predicted, the total cost and timing of the project will be severely affected.

Fig. 5

Diagram of changes in weekly advance rate, its average and TBM cumulative advance rate in Ghomrood tunnel (lots 3 and 4)

Although the average performance parameters of this machine are not significant due to long stops, records achieved in this project (maximum daily, weekly, and monthly advances) are remarkable (Table 4).

Table 4 Records registered during excavation of Ghomrood tunnel (lots 3 and 4)

Adverse geological conditions

The main geological problems that happened in this project and caused long time delays and, subsequently, drastic drops in machine’s average utilization factor were situations such as squeezing (Fig. 6a, b) and tunnel collapse. These problems mostly occurred in foliated rocks (shale, slate, and phyllite beds) between chainage 3000 and 6000 m, where due to existence of a fault with strike parallel to the tunnel alignment (and poor geotechnical properties of surrounding rocks) and also relatively high overburden (as can be seen in Fig. 2), geological conditions were unfavorable. High convergence of surrounding rock in this section of tunnel caused machine stuck and shield jamming in many points. In some cases, for releasing the machine, it was required to dig the rock around the shield (Fig. 6c, d).

Fig. 6

The occurrence of squeezing and TBM (shield) jamming in chainage 3550 m: a convergence of ground and contact of rock with segment and shield; b making free shield by mining the surrounding rock and stabilizing it temporally by timbers

Apart from stops less than 2 to 3 weeks, in two time periods from week 31 till week 40 (for 10 weeks) and then from week 57 till week 136 (for 80 weeks or almost a year and a half), the machine advance was very low and virtually no progress has been made (Fig. 5a). Meanwhile due to special geological conditions and the occurrence of numerous collapses and exceeded ground convergence and creating numerous executive problems, persisting to continue the activity in the tunnel becomes very difficult and eventually shutdown. This longsome interruption, during which efforts were made to overcome the problem, as is shown in diagram c and d of Fig. 5, caused a sharp drop in the graph of average advance rate and utilization factor. In this project, from 36th week onwards, after making adjustments on the machine and solving the problems, boring has gradually returned to the first favorable state and as noted previously good records in daily, weekly, and monthly advances were achieved.

Data collecting and developing the database

Collecting appropriate data for developing a suitable database is considered to be the main part of this research. In this regard, all geological information obtained in pre-construction and construction phases (including results of surface outcrop studies, geotechnical borehole logs, laboratory tests, tunnel face observations, studying excavated materials and tunnel back mappings) and also machine operational data and actual performance were collected and recorded in appropriate databases. In this research, in addition to databases prepared in the site (which contains tunneling activities during a shift and the average machine operational parameters in each stroke), a special database has also been prepared to verify the relationship between geological parameters and parameters related to actual machine performance and developing an empirical prediction model. This database contains rock mass engineering geological properties and average operational and performance parameters of the machine in 31 different sections of the tunnel in which geological data with appropriate quality was available. Each section has a length of 5–10 m depending on shift advance.

As mentioned previously, this tunnel was bored by a double shield TBM and lined with concrete segmental lining. In mechanized tunneling by shielded machines, there are some limitations for observing the rock around the tunnel and collecting required geological data. These observations and measurements are limited to maintenance shifts when access to the rock face is possible. So, in this research required data were collected only from locations where the rock face was investigated during geological back-mapping or exploration borings extended to the tunnel level.

The developed database consists of two main parts. The first part contains some parameters such as total boring time and length of bored section, the average of operational parameters (including net thrust, total thrust, cutterhead speed, power, torque), and some performance parameters (such as penetration rate and field penetration index or FPI) through the tunnel section. These parameters were obtained from the shift operating records and data acquisition system installed in the machine.

The formulas for the operational parameters and composite parameters are as follows:

$$ PR=\frac{L_b}{t_b} $$

(1)

$$ P=\frac{1000\ PR\ }{RPM\ast 60} $$

(2)

$$ FPI=\frac{F_n\ }{P\ } $$

(3)

$$ SE=\frac{200\ {N}_{TBM}{r}_{mc}}{3\ {d}_{TBM}}\ \frac{F_r}{P} $$

(4)

In these equations, F_n and F_r are average normal cutter load and cutter rolling force (kN/cutter) respectively, L_b is boring length (m), t_b is net boring time (hour), P is the penetration (mm/rev), PR is penetration rate (m/h), RPM is cutterhead rotation speed (rev/min), N_TBM is number of cutters installed on cutterhead, d_TBM is TBM diameter (m), FPI is field penetration index (kN/cutter/mm/rev), and SE is specific energy (MJ/m³).

In the second part of the database, the estimated or averaged value of engineering geological and geotechnical properties of rock masses in each selected section have been recorded. Important intact rock properties (compressive and tensile strength, quartz content, porosity), discontinuity characteristics such as spacing and surface condition and also some rock mass classification systems like RQD, RMR, Q, and GSI as well as rock mass strength parameters like Qc (Barton 2000), rock mass compressive strength or UCS_rm (Hoek 2007), rock mass strength or Sigma_cm (Barton 2000), and rock mass cuttability index or RMCI (Bilgin et al. 1997) are some of the parameters which have been calculated for each section and recorded in database. Equations (5)–(9a, 9b) have been used for calculating rock mass strength parameters:

$$ RMCI=\frac{UCS}{100}{(RQD)}^{2/3} $$

(5)

$$ {Q}_c=Q\cdotp \frac{UCS}{100} $$

(6)

$$ {Sigma}_{cm}=5\cdotp \gamma \cdotp {Q_c}^{1/3} $$

(7)

$$ {UCS}_{rm}= UCS\cdotp {s}^a $$

(8)

where:

$$ s= Exp\frac{100- GSI}{9-3D} $$

(9a)

$$ a=\frac{1}{2}+\frac{1}{6}\left({e}^{- GSI/15}{e}^{-20/3}\right) $$

(9b)

In the above equations, D is the disturbance factor (D = 0 in TBM projects) and g is the density of the rock.

Histograms related to distribution frequency and the outcome of descriptive statistical analyses performed on recorded data are presented in Fig. 7 and Table 5, respectively. Some of these parameter variations are also presented in Fig. 8. These histograms and graphs show relatively wide variation ranges for parameters recorded in the database. Therefore, the established database is appropriate for comprehensive statistical analyses and empirical model development.

Fig. 7

Distribution frequency of geological and TBM performance and operating parameters in the established database: a field penetration index or FPI (kN/cutter/mm/rev); b penetration or P (mm/rev); c cutter load or Fn (kN); d uniaxial compressive strength or UCS (MPa); e rock quality designation or RQD (%); f rock mass cuttability index or RMCI (MPa)

Table 5 Descriptive statistical analysis of developed database

Fig. 8

Variation of some rock mass parameters measured in selected sections of the tunnel

Developing empirical relationships

During the studies conducted by authors, different engineering-geological parameters which are organized in the database (e.g., UCS, joint spacing, rock mass classification and rock mass strength parameters), have been analyzed using statistical techniques and the best correlations between these parameters and machine performance parameters such as PR, P, SE, and FPI are obtained.

The previous studies (Hassanpour et al. 2009, 2011) show that employing composite parameters is very helpful to consider the effect of all effective parameters in analyses. In this research, among all different machine parameters, the composite parameter of field penetration index or FPI, which is introduced by Eq. 3, shows more adaptation with engineering geological parameters. Therefore, in this study, FPI is selected as an appropriate parameter to evaluate the effect of different parameters on machine performance and also development of empirical relationships. Using this parameter penetration rate, thrust and cutterhead rotation speed (RPM) will be involved in the analyses at the same time.

Table 6 and diagrams presented in Fig. 9 show the results of the analyses conducted. As can be seen, in this project a good adaptation between geological parameters and FPI has been observed.

Table 6 Results of the matching between FPI and engineering geological parameters of Ghomrood project

Fig. 9

Correlation between FPI and different geological parameters: a UCS; b RQD₀; c basic RMR; d GSI; e RMCI; and f Sigma_cm

In this study, in addition to the univariate regression analysis, multivariate regression analysis was also used to find empirical relationships between rock mass different parameters and machine performance parameters that are applicable for existing data. To take the advantage of this statistical method, four rock mass parameters including uniaxial compressive strength, joint opening, RQD, and α angle (angle between the discontinuity surface and the tunnel axis) as independent parameters and FPI as dependent parameter have been selected. The impact of each variable on FPI was evaluated according to the forward stepwise regression method using SPSS. Putting aside two parameters of joint opening and α angle, software established the best linear adaptation with a confidence level of 95% between Ln(FPI) as dependent parameter and UCS and RQD as independent parameters, after controlling different combinations of parameters. Finally, Eq. 21 was obtained as a new prediction model for this project’s geological conditions and machine characteristics.

$$ FPI=\mathit{\exp}\left(0.004 UCS+0.023 RQD+1.003\right) $$

(21)

To facilitate using Eq. 21 and making a comparison between different rock masses in terms of boring feasible, the above-mentioned relationship has been drawn as shown in Fig. 10. Using this chart, the variation range of FPI can be estimated according to the variation range of rock mass parameters (UCS and RQD).

Fig. 10

Diagram of FPI changes based on variation range of rock mass parameters (for carbonate rocks and schistose and foliated rocks of Ghomrood tunnel

Comparison with previously developed models

In this research, a comparison was made between this model and the previous models developed for other geological conditions (by Hassanpour et al. 2009, 2010). These models, their characteristics, and the range of their application are listed in Table 7. To reach this goal, sensitivity analyses have been done on input parameters of current and previously developed models. The results have been shown in the graphs presented in Fig. 11. These graphs show that in some geological conditions (jointed and layered pyroclastic rocks), FPI has been more sensitive to UCS than RQD. It means that variations of UCS in these conditions can influence the TBM performance (or cutter penetration) more effectively than the degree of rock fracturing. In other geological conditions (very strong to strong igneous and metamorphic rocks), this trend is reversed and FPI is more sensitive to RQD. These results show that the prediction models are very sensitive to geological conditions and it is necessary to develop a suitable model for each geological condition, separately.

Table 7 Previous equations developed by authors for each tunnel project separately

Fig. 11

Investigation of FPI sensitivity to input parameters in different developed equations: a Eq. 21; b Eq. 22; c Eq. 23; and d Eq. 24

Conclusion

The results of this study indicate that there is a logical and meaningful relationship between engineering geological parameters and parameters related to machine performance. Among all engineering geological parameters, rock mass strength parameters, which represent both the effect of joints and strength at the same time, have better adaptation with parameters related to machine performance such as FPI which is a combined parameter itself; accordingly, they are very suitable for developing empirical relationships and models. Relationships obtained using multivariate regression data analysis method with a relatively good correlation coefficient also take into account two important rock mass parameters of intact rock strength and joint conditions (spacing or RQD).

It should be noted that this model’s application range is limited to the situations with similar geological conditions, in cases where the tunnel is excavated through low to medium grade metamorphic rocks (slate, phyllite, and schist in which foliation and schistosity are slightly developed) and sedimentary rocks (limestone, sandstone, and shale) which in all cases have UCS ranging between 30 and 175 MPa and varying quartz content.