- Email: [email protected]
Delineating subterranean water conduits using hydraulic testing and machine performance parameters in TBM tunnel post-grouting
International Journal of Rock Mechanics & Mining Sciences 70 (2014) 308–317
Contents lists available at ScienceDirect
International Journal of Rock Mechanics & Mining Sciences journal homepage: www.elsevier.com/locate/ijrmms
Delineating subterranean water conduits using hydraulic testing and machine performance parameters in TBM tunnel post-grouting Ghassem Jalilian Khave n Lar Consulting Engineers, 30 Sharifi St., N. Gandi Ave., Vanak Sq., 19699-44311 Tehran, Iran
art ic l e i nf o
a b s t r a c t
Article history: Received 20 November 2012 Received in revised form 28 June 2013 Accepted 25 April 2014
Zagros Water Conveyance tunnel in western Iran crosses a vast unconfined aquifer. This TBM burrowed tunnel has long drained the region’s groundwater which is unusually rich in hydrogen sulfide (H2S) gas. The gas reacts with the tunnel humidity and produces an acidic fume that penetrates the tunnel lining and causes its decay. The exact locations of the discharging conduits and their morphology are not known, since they are concealed by TBM segmental lining. An elaborate post-grouting plan is on the drawing board to reclaim the aquifer, but the cost of a systematic grouting is particularly high and at best this scenario not conclusive. The scope of this paper is to discuss a series of carefully controlled field experiments in a 100 m pilot study area along the tunnel. The scheme manipulates the TBM performance parameters recorded during the tunnel excavation. Based on this information, suitable field models are established that may be interpreted as being associated with either water or air-filled solution channels. & 2014 Elsevier Ltd. All rights reserved.
Keywords: TBM tunnelling Machine performance parameters H2S gas Ecological impact Tunnel post-grouting
1. Introduction Zagros water conveyance tunnel in Kermanshah province of Iran is near the town of Pol-e Zahab. This part of the tunnel ( 26 km long) is regarded as the second lot of a broader plan, which is schemed approximately 48 km long and is 6.73 m in diameter. It has been under construction using a Herrenknecht hard rock double-shield TBM since March 2005. So far, 16 km (61%) of this tunnel has been completed. In the course of tunneling, the machine encountered nearly many extraordinary situations related to continuous intrusion of H2S gas, all of which resulted in a significant reduction in TBM utilization rate and an increase in construction delays, as well as high cost. The encountered circumstances contradicted the preconstruction subsurface investigation results, which had forecasted low to moderate amount of fresh water seepage at a maximum rate of 5–10 l/s/km [1]. One of the major factors affecting the performance of tunnel boring machines is the degree of fracturing of the rock [2]. During excavation, machine performance parameters are continuously displayed and recorded on the control cabin monitors. Fig. 1 shows the TBM display screens. The screens display TBM performance parameters and register TBM mechanical behavior against the excavated material [3]. These screens disclose an assorted set of
n
Tel.: þ 98 21 22000811; fax: þ 98 21 88883905; Cell: þ 98 912 5035910. E-mail address: [email protected]
http://dx.doi.org/10.1016/j.ijrmms.2014.04.013 1365-1609/& 2014 Elsevier Ltd. All rights reserved.
data; e.g. penetration rate, boring time, total thrust, torque power, cutter speed and etc. Based on the careful analysis of these parameters, uniform patterns were established as a model to identify concealed joints and to delineate water-carrying conduits along the tunnel path. The objective is to improvise a model to pinpoint water conduits and fill in their leaks and cracks by postcurtain-grouting. This is in contrast to a systematic grouting approach where grouting is done in a continuum fan pattern regardless of water points of entries [4].
2. Geologic setting According to the structural geology zonations of Iran, Zagros tunnel is located in the core of the Zagros Mountain Range. This region includes simple structures of reverse faulting and symmetrical anticlines and synclines known as Folded Zagros province. The tunnel horizon is situated within the folded zone of Zagros Mountain Range, consisting of sedimentary rock formations at an average depth of 400 m (Hmax 950 m) [1]. The geological unit along the Zagros tunnel in the pilot study area consists of brownish gray limestone of Garou Formation. Overlaying this unit is Gurpi Formation, which is comprised of alternating thin to thick-bedded shale and argillaceous limestone. Some layers of Gurpi Formation are rich in pyrites in the form of nodules [5]. The assembled rock strata changed frequently from hard rock to soft, dry to wet, stable to instable, stick to nonsticky ground (and vice versa), more often than anticipated. Essentially,
G.J. Khave / International Journal of Rock Mechanics & Mining Sciences 70 (2014) 308–317
309
Fig. 1. TBM control cabin main display screens showing excavation data and machine mechanical behaviour against the excavated material.
Fig. 2. Longitudinal geological profile of lot 2 Zagros tunnel along the study area showing the regional GWL. [after 8].
Gurpi sections are aquiclude layers, so there is no substantial groundwater accumulation in these parts of the tunnel. However, Garou sections in the form of anticline serve as an unconfined “water table” aquifer [1]. The geological profile of the project line is illustrated in Fig. 2. The pilot study area is located in the core of a major geological structure known as Ezgelleh Anticline (chainage 04 þ359 to 04 þ458). The tunnel overburden thickness along this area is roughly 155 m. Generally, the phreatic zone is congruent to the areal topography and had stood 105 m above the tunnel crown before its excavation. It was later drained out and the water table plunged below the invert level after completion of the tunnel. The water seepage in low amounts was first experienced at TM 3700. A significant water ingress in the range of Q4110 l/s was later intruded at TM 4157, which rapidly accumulated to Q 315 l/s with further advancement to TM 4435. Advancing deeper into the core of aquifer at TM 8256, the accumulative water seepage totaled Q 730 l/s, with further advancement to TM 13846, 155 more liters of gas bearing water seeped into the tunnel, totaling the tunnel discharge flow at the portal outlet to 900 l/s. This amount of water released 700 ppm hydrogen sulfide gas into tunnel atmosphere. Fig. 3 shows the water discharge rate and the H2S gas concentration at the pilot study area where is marked as Ezgelleh Anticline on the graph. As is evident by the piezometric levels of boreholes 26 and 27 in Fig. 4, the groundwater level on the average has dropped 27 cm/ m of TBM’s advancement before it subsided below the tunnel invert level. This observation is roughly equivalent to the secondary permeability of the medium (hydraulic conductivity reaching up to 0.1–0.2 m/s), which is of similar value with most permeable rocks [6]. The conclusion became a useful criterion to predict the hydraulic characteristics of the aquifer and to estimate the host rock permeability at any given location along the tunnel. Generally, Garou Formation is known to be as the host rock in many major oil (gas) bearing basins further down in the south
provinces of Iran. As a result, some hydrocarbon materials may have migrated or leached out during uplift movements in this area, due to lack of a suitable cap rock and a favorable geological structure to trap the hydrocarbons. Therefore, only traces of residual hydrocarbons have remained in the rock formations in the form of black tarry liquids [7]. These liquids have been frequently observed along the tunnel path, seeping through holes and gaps of the tunnel lining. Hence, the hydrogen sulfide (H2S) gas is most likely an associated component of the remaining gas and oil in existing traps. Hydrogen sulfide liberates from the groundwater; so, its amount depends on the quantity of water inflow. The groundwater discharges through the open joints and solution channels. Obviously, part of water circulates along the space between segments and rock and the water source is difficult to track and handle properly. Prior to this experiment, an elaborate joint survey was carried out from the surface to geometrically define the dominant discontinuities along the Ezgelleh Anticline and to produce a reliable model for the hydraulic characteristics of the rock mass. Overall, four major joint sets identified along the pilot study area (axis N2151). It was also determined that the prevailing joints dipped at an angle (α) of 301 to 801 with dip directions (β) varying from 441 to 2641. Their surface characteristics are undulating, moderately to highly weathered, emax ¼25 mm, at 35–225 cm spacing, and are elongated 1–10 m. Based on the information adopted from over 150 borehole water pressure tests (Lugeon Test) during pre-construction subsurface geotechnical investigation—interpolated most compatible with actual geological conditions (elevation, structural and lithological) along the bored tunnel, average permeability coefficient of the intact rock is calculated between 10 5 to 10 6 cm/s [1]. In general, 17% of the Lugeon tests were reported as impervious and 42% resulted in “laminar” flow. Turbulent flow that generally represents open joints and “washout” that refers to washing the joint filling were tallied at 21% and 12%, respectively. Dilation flow
310
G.J. Khave / International Journal of Rock Mechanics & Mining Sciences 70 (2014) 308–317
Histogram of Portal Water Discharge and TBM Cutterhead Gas Concentration 700
900 Total water Discharge
650
Gas Concentration
600 550
700 600
500
Pilot Study Area Ezgelleh Anticline
450 400
500
350 400
300 250
300
200 200
150
H2S Gas Concentration(ppm)
Total Water Discharge (lit/s)
800
100
100
50 0
3, 7 3, 68 8 3, 50 8 3, 63 8 3, 63 86 4, 3 1 4, 15 3 4, 06 4 4, 46 5 4, 02 8 5, 40 19 5, 6 9 7, 25 0 7, 74 8 8, 64 2 8, 31 3 8, 12 59 8, 9 9 9, 30 10 353 , 10 021 , 11 738 , 12 536 , 13 394 , 13 222 , 13 334 , 13 782 , 13 858 , 13 923 , 14 941 , 14 118 , 14 233 , 14 309 , 14 393 , 14 643 ,7 34
0
Tunnel Chainage Fig. 3. Tunnel accumulative water discharge rate histogram at portal showing the hydological condition and approximate location of the pilot study area.
Zagros Tunnel Areal Aquifer Fluctuation Corresponding to TBM Advancement
DrawdownTable
750
Groundwater Elevation (m)
730 710 690 670 650 630 610 590 570 550 21F
22F
23
26
27
Project Borehole Piezometric Levels on November 2003 Prior to Tunnel Excavation Project Borehole Piezometric Levels on July 2007 and TBM Location Project Borehole Piezometric Levels on December 2008 and TBM Location Project Borehole Piezometric Levels on June 2009 and TBM Location
28
29F
30F
30
Borehole ID. No.
Fig. 4. Tunnel construction impact by January 2008 on Ezgelleh anticline groundwater hydrostatic level and its drawdown rate as elucidated by project borehole piezometers.
as governed by elastic behavior of joint openings, which is not ordinarily considered a normal rock behavior at typical groundwater pressure but is expected to take effect at seasonal fluctuations of groundwater hydrostatic head, accounted for 8% of test results. Thus, it is interpreted, the host rock has low permeability properties and any occurring water seepage will be transmitted through a secondary network of interconnecting open joint systems. Nevertheless, approximately 33% of all joints encountered could provide a suitable medium for water migration [1]. The cores of Garou Limestone, as observed in borehole 27, exhibit bedding planes spaced from 100 mm to 1 m. The orientation of the bedding planes varied from perpendicular to the core to approximately 251 from this. Inclined joints within the core varied about 401 to 701 to the core axis and, typically, there were four or five inclined joints in each of the core boxes. The condition of the core was not uniform, with frequent calcite lenses and stylolitic surfaces [8].
3. Tunnelling impact on ecological conditions Excavation of tunnels usually affects the natural surroundings to a great extent. Its intensity depends on the location of the tunnel in relation to the region’s aquifer and the hydrogeological behavior of the rock mass in respect to tunnel. In the case of Zagros tunnel, the excavation work has adverse affects on all aspects of the project construction because of repeated intrusion of hydrogen sulfide gas contaminated groundwater. To make matters more complex, the project is located in a semi-arid area, and constant drainage of the aquifer (416 l/s) by the advancing tunnel is rapidly exhausting the region’s water resources. The implemented mitigation measures have considerably reduced the adverse affects. Nevertheless, the interference with the ecological setting is so severe that demands an immediate and comprehensive remedial action. In addition, droughts in recent years have strained the already imbalanced regional groundwater reserves
G.J. Khave / International Journal of Rock Mechanics & Mining Sciences 70 (2014) 308–317
311
Zagros Tunnel Regional Water Table Drawdown Radius of Influence Permanently dried springs Initial Static Water Table Prior to Tunnel Excavation Water Table Drawdown Abdallan spring influenced by seasonal precipitations
3.5km Zone with Springs Influenced By Water Table Drawdown Draining Zagros Tunnel Fig. 5. Regional groundwater table drawdown showing tunnel radius of influence (cone of depression) with springs effected by water table drawdown. [after14].
and have dried many local springs. Hence, a large number of permanent springs have dried up and in a few areas that the tunnel has not fully affected the groundwater table; water discharge has trickled to 50–400 l/min. As a result, many villages have suffered serious water shortage and their habitants reluctantly have relocated to other areas. Most areal springs are directly or indirectly connected to the joint systems that have developed to a network of subterranian channels by solutional weathering. As it was also concluded by monitoring of the local springs, the tunnel maximum radius of influence is calculated approximately 3.5 km (Fig. 5). For instance, Abdallan spring, which is one of the prominent springs in the region, is roughly the same distance from the tunnel axis. Its discharge rate fluctuations are consistent with the seasonal precipitations and the tunnel advancing rate in the aquifer. For example, most springs in the area show a gradual increase in discharge rate within 15 days from the beginning of every wet season and a rapid decline after similar time lap when the wet season expires. Fig. 6 illustrates Abdallan spring monthly discharge histogram. The overall depletion of the aquifer is evident by the histogram declinig trend.
4. Tunnel post-excavation grouting It is fundamentally true that successful contact grouting (liner backfilling by pea-gravelþcement grout injection) in TBM tunnelling is crucial in reducing groundwater infiltration [9]. Hence, in Zagros tunnel case a successful contact grouting could dampen the intrusion of water into the tunnel and decrease the intensity of gas emission into atmosphere. As a result, less health risk for the tunnel crew and more life expectancy for the tunnel lining. On the other hand, taking the hexagonal geometry (flexible) of the Zagros tunnel lining into account, the importance of segment installation tolerances and imperfections demanded particular attention. Hence, the process of quality assurance that would control both the geometrical dimensions and fitting of segments into lining rings has an important role to compensate for some of the inherent deficiency of hexagonal lining in restraining water intrusion into the tunnel [10]. Ordinarily, to block the leakage between segments, efforts were made to backfill the space between segments and rock with as much as pea-gravel, and fill the voids with enough cement slurry to form a secondary concrete lining. Despite the best efforts, filling
the overt and invert areas in high water seepage zones was not possible due to gravitational purge of grout mix from holes in overt area and continuous leakage of grout mix from submergedunsealed gaps in the invert area. Fig. 7 shows three representative core boxes sampled during water pressure tests. They reveal the extent of contact grouting success in high water areas. The alternative measures that were carried out were polyurethane based foam grouting and resin injection. Pre-drainage and plugging the segment joints with hydrophilic sealants, which all proved incredibly time consuming and in particular, costly. To reclaim the aquifer and to rehabilitate the local springs, and to also prevent tunnel lining from further decay by H2S destructive forces; an elaborate post construction curtain-grouting program was schemed. The notion was to supplement the existing contact (back-fill) grouting, to seal off the remaining spot leakages most especially where the contact grouting had not satisfied the required average tightness of the tunnel lining, e.g. 6 l/s/km [11]. It was originally perceived that a sequential (systematic) grouting of the tunnel from eight holes in the circumference was the best approach. The program comprised radially fan pattern, perpendicular to the centerline. The holes were to drill systematically to 5 m depth and grout with cement mix, up to 5 bar pressure at 1.6 m lateral spacing. To complete the consecutive rings, drilling in segment center-points and injecting micro cement to depth of 4 m at higher pressure (pmax ¼7.5 bar) to penetrate difficult to reach areas was deemed necessary [12]. To be able to react on disintegrated rock mass and deficiencies of the bedding plain, an immediate consilidation borehole grouting was to be applied in the zone where the P-gravel is applied under normal condition. The immediate consolidation grouting was to be carried out via short boreholes ( 1.0 m) by a single mobile grout pump. The estimated cost of the designed curtain grouting in a systematic method for the first 16 km of the tunnel was enormous and the time delay could not be justified within the contract time-frame. To achieve a more reasonable timetable and attain a more assuring result for the water-tightening plan, a swift approach and an accurate procedure to distinguish the location of the water conduits and to individually plug them off, was deemed the sole alternative. The procedure must to avoid lengthy exposure of crew to tunnel gassy condition and to facilitate grouting work in tunnel constraining work condition, e.g. limited workspace, high pressure water inrush, hydrogen sulfide gas pollution and logistic problems. Voluminous information generated during tunnel excavation in the form of machine performance parameters (excavation data
312
G.J. Khave / International Journal of Rock Mechanics & Mining Sciences 70 (2014) 308–317
Abdallan Village Spring Discharge Histogram 1800
Before Tunnel Excavation After Tunnel Excavation 1600
Seasonal Precipitations Marker (+) Beginning (-) Ending
1400
Q (lit/min)
1200 1000 800 600 400 200
A
ug Se -06 p O -06 c N t-06 ov D -06 ec Ja -06 n Fe -07 b M -07 ar A -07 p M r-0 ay 7 Ju -07 nJu 07 A l-0 ug 7 Se -07 p O -07 c N t-07 ov D -07 ec Ja -07 n Fe -08 b M -08 a A r-08 p M r-0 ay 8 Ju -08 nJu 08 A l-0 ug 8 Se -08 p O -08 c N t-0 ov 8 D -08 ec Ja -08 n Fe -09 b M -09 ar A -09 p M r-0 ay 9 Ju -09 nJu 09 A l-0 ug 9 Se -09 p O -09 c N t-09 ov D -09 ec Ja -09 n Fe -10 b M -10 a A r-10 p M r-1 ay 0 Ju -10 nJu 10 A l-1 ug 0 Se -10 p O -10 c N t-1 ov 0 D -10 ec -1 0
0
Date (Month-Year) Fig. 6. Histogram of Abdallan spring monitoring during the tunnel construction, showing the monthly discharge rate and the duration of seasonal precipitation periods as reference.
5. Experimental grout method
Fig. 7. Cores showing segment concrete and P-gravel zone infilling behind tunnel lining that often left partially to completely empty due to grout mix washouts in high water pressure zones. The top photo represents a complete recovered core. The middle photo shows a gap in core between segment concrete and bedrock left by the grout washout, and the bottom picture shows a partially filled P-gravel zone between rock and tunnel segmental lining.
and machine mechanical behaviour against the excavated ground curves) were determined a viable approach to be explored. Therefore, the machine performance data was gathered and carefully examined to delineate some logical correlations with the actual ground conditions. Hence, identify the open joints and the cracks that are capable of transmitting water. The analysis took into account all relevant parameters on TBM data displayers, but only two parameters in particular; Excavation Penetration Rate and Cutter-head Torque Power parameters exhibited pronounced changes that also conformed to the corresponding water pressure test (WPT) results, which incorporated the location of tunnel known water intrusion points. The information analysis required great care to avoid very complicated ambiguities associated with interferences from softer rock in interlayer and probable errors in data keeping during the excavation phase. Fig. 8 illustrates the schematic drawing of the preferred array and angle of incidence adapted in methodic (localized) tunnel grouting in the pilot study area.
As mentioned before, three prominent master joints and a bedding plane were recognized along the pilot study area. The angle of incidence of most prominent joint set (J2) dips about 791 in a southeasterly direction of N1741. The J2 average aperture is measured 15–25 mm (moderately narrow) and their spacing is between 150 cm to 225 cm (very widely spaced). The secondary and tertiary master joints with respect to their percentage of pole concentrations are classified as J3 and J4. They are generally spaced at 150–225 and 45–100 cm (e¼10–20 and 0.1–1.0 mm), respectively. Table 1 diplayes the summary of the indigenous rock mass discontinuity characteristics along the pilot study area [8]. The main tunnel extends in a NE-SW direction (N2151) and the areal bedding plane intersects the tunnel axis at an angle of (α ¼641). In the first stage of experiment, 141 boreholes were systematically drilled and water pressure tested based on the original pattern. The results were plotted in an X–Y graph. Based on the obtained data, the lugeon values were classified in four categories (Fig. 9). To summarize, a total of 51% of the lugeon tests attained low permeability. Twenty percent classified as medium permeability, 12% exhibited high permeability and 17% of the tested holes displayed very high permeability. The graphical analysis revealed that in a systematic approach, when test holes are aligned regardless of the joint orientation, the possibility of intercepting all of the dominating discontinuities is more likely incidental. To enhance the method, a series of boreholes at different azimuths (þ301 to 301) were drilled. The goal was to intercept as many cracks and fissure planes as possible and perform water pressure tests. From the lugeon values, an optimum azimuth angle which apperaed most appropriate to detect air or water filled open joints was determined. In short, the borehole azimuth that intercepted most number of open joints and active water conduits was α ¼ 301. This particular azimuth had resulted in the highest number of medium to very high lugeon values. Fig. 10 illustrates the concentration of boreholes at different angles (azimuths) and their corresponding lugeon values.
G.J. Khave / International Journal of Rock Mechanics & Mining Sciences 70 (2014) 308–317
313
Fig. 8. Plan and profile of the preferred array and the pattern adapted in post- grouting procedure in pilot area. The Plan shows locations of different trials for favorable WPT azimuthal angle. Table 1 Summary of the pilot area rock mass discontinuity characteristics [8]. Parameter
Type Joint set
Bedding joint Conjugate shear joint
J1 J2 J3 J3
Attitude (degrees)
Roughness
βJ
αJ
JRC
Jr
061 174 044 264
30 79 71 51
4–6 6–8 6–8 6–8
3.0 1.5 1.5 2.0
Weathering (Ja)
Aperture e(mm)
Spacing (cm)
Length (m)
3.0 2.0 2.0 1.0
1–10 15–25 10–20 0.1–1.0
35–65 150–225 150–225 45–100
3–10 1–3 3–10 1–3
Tunnel Rockmass Lugeon Values in Systematic Approach 100 90 80
VERY HIGH= 17%
LU Value
70 60 50 40
HIGH= 12%
30 20
MEDIUM= 20%
10 LOW=51%
0 0
10
20
30
40
50
60
70
80
90
100
110
120
130
140
150
No. of Tests Fig. 9. In-situ water pressure test results in a systematic approach with majority of the test results are classified as low to moderate values, indicating an incidental interception of the discontinuities.
Monitoring a new round of WPT with the selected azimuth (α ¼ 301), it was concluded that lugeon values have significantly increased by almost 40%. The assumption is based on 118 lugeon tests, conducted. In general, 26% of the tests showed low perneability,
26% displayed medium permeability, 26% had high permeability and finally 22% indicated a very high permeability. Thus, the method had good merits and proven to be fairly reliable. Fig. 11 illustrates the graphical analysis of the lugeon values in the methodic approach.
314
G.J. Khave / International Journal of Rock Mechanics & Mining Sciences 70 (2014) 308–317
Incorporating flow rate, pressure and total volume of cement consumption during the actual grouting stage, pressure-flow diagrams were developed. The diagrams also indicated if leaks back to the tunnel perimeter or connections to adjacent holes had occurred or if excessive pressure had caused damage to the rock mass. The drilling for grouting was made in 6.5 m linear stages and a down-stage procedure was utilized. The drilling holes with huge water leakages and material transport were performed in ascending stage with collar. To avoid local empty rooms and ‘air pillows” above backfilling concrete close to crown, contact grouting with mortar and curtain grouting were done, simultaneously as post construction program. Thus, the grout hose with simple valves type “Tube-a’-manchette” (sleeve port pipes) were mounted, as well. In following stage, rock just behind segments was grouted with a low pressure (5 bar) to avoid damage of segments and grout mix leakage into the tunnel. For deeper stages, recommended pressure was 7.5 bar. In some cases, the pressure was raised to 10 bar to overcome prevailing pressure. The ambient pressure varied from 7 to 9 bar depending on depth below water surface
and volume of water penetration. And injection pressure did not surpass the ambient pressure by more than one bar. The injection equipment used in Zagros tunnel post-grouting as illustrated in Fig. 12 included two double piston injection pump KOS 1000, two double piston injection pump KOV 1000, two electric power rotary drill rigs, two ultra mixers UM 10, six flat rail cars with separate containers and lift frames, a cement mortar mixer, an electro-hydraulical power units totaling 400 kW, a reverse tanker water tanker with a capacity of 2000 l, a cement holding silo with a capacity of 2.5 t, injection pressure and electronic flow measuring units with a central operating panel, and injection pressure and electronic flow measuring units with central operating panel.
6. Linking machine performance parameters to ground anomalies Wanner and Aeberli [13] cited to penetration rates between 50 and 100% higher than the daily average when open joints were
Tunnel Rockmass In-Situ Water Pressure Test Results at Different Azimuths 100 90 80 VERY HIGH
70
LU Value
60 50 HIGH
40 30 20
MEDIUM
10 LOW
0 -40
-30
-20
-10
0
10
20
30
40
Borehole Alignment Azimuths (Degree) Fig. 10. Tunnel in-situ WPT results at different azimuths in respect to tunnel longitudinal axis. The optimum azimuth proved to be α ¼ 301 signifying most number of open jonts intercepted.
Tunnel Rockmass Lugeon Value In Methodic Approach 100 90 80
VERY HIGH= 22%
Lu Value
70 60 50 HIGH= 26%
40 30 20 MEDIUM= 26%
10 LOW= 26%
0 0
10
20
30
40
50
60
70
80
90
100
110
120
130
No. of Tests Fig. 11. In-situ water pressure test results in methodic approach with majority of the results are classified as moderate to very high values, exhibiting a high precision in intercepting major discontinuities.
G.J. Khave / International Journal of Rock Mechanics & Mining Sciences 70 (2014) 308–317
315
Fig. 12. Zagros tunnel post-grouting equipment setup. [5].
present. They also suggested that rough-walled tension joints were least helpful to penetration rate, especially if they were healed. To examine the validity of their proposed criterion, a 100 m span along the Zagros tunnel with known high water discharge between TM 4359 to TM 4458 was appointed as a pilot study area (Fig. 2). Field reconnaissance in conjunction with local rockmass discontinuity survey and TBM performance parameters supplemented by tunnel daily water discharge rate were compiled to individual curves to establish a fixed pattern for the location of known high water bearing zones. Recognition of subsurface materials or conditions from TBM performance parameters was ambiguous at times and distinguishing solution channels from their surrounding rocks based on machine parameters required a comprehensive knowledge of TBM mechanism. This is due to influential effects of contingent machine parameters that are under direct TBM operator control and most significantly, the TBM’s overall performance status and equipment efficiency, e.g. disc cutters, hydraulic jacks, electromotor, power usage, etc. In order to test the validity of technique and establish a field model for further scrutiny, the approximate location of five known conduits with discharge rates ranging from 5 to 120 l/s in the study area were transposed on the TBM “Advance Speed” and “Cutter-head Torque” curves. Upon examination of each data curve, it was learned that the cutter-head torque power tend to drop significantly as TBM neared the known water transmitting points. It was also apparent from the TBM advance speed that the same points tend to peak as the TBM cutter-head crossed the center-point of the known solution channels at a right angle. The close approximation of the groundwater percolation points to the slope changes in both cases was very encouraging, considering the vague relationship between the cutter-head torque power and the TBM advance speed rate. In general, points of slope change in data curves indicated face rock strength (resistive) contrast, whereas linear data were interpreted as evidence of uniform strength. In addition, sharp declines in cutter-head torque power were marked as air or water-filled solution channels, where as steep increase in penetration rate interpreted as the same circumstances. Moreover, the tension cracks that had showed high permeability (Q¼13–23 l/s at TM 4440-4458) displayed rather indistinctive signals in both penetration rate and cutter-head torque power data. Thus, pinpointing the exact location of tension cracks will merely depend on the engineer’s judgment. Although not particularly evaluated in a great extend, it is likely that similar anomalies could prevail by TBM’s total trust power, which is rather proportionate to TBM operator performance and skill.
After establishing the applicability of the technique over the known solution channels, the effort was directed toward detecting other suspected networks of open joints along the same span. Eighteen additional grout-holes were drilled and tested using the set criteria. Figs. 13 and 14 are the graphical presentations of the data points obtained from the machine parameters, the consolidation (curtain) grout cement take and the approximate location of points of water entry into the tunnel. Thus, the contrast in machine performance curves combined with grout cement takes and water infiltration fluctuations at the tunnel portal revealed anomalies correlating with the individual water conduits. This demonstrated that the excavation data and machine mechanical behaviour against the excavated material can differentiate between solid rock and the open joints and fissure zones in fractured rock. Subsequently, 32 additional check-holes were drilled to assess the efficiency of the devised method. They were drilled on both sides of the grout-holes. Water pressure tests were conducted before and after grouting procedure at each hole. Fig. 15 illustrates the degree of ground improvement at grout check holes after conducting the methodic pattern. It clearly manifests that ground improvement in the range of 50% can be accomplished by methodic grouting, which is conceived as an optimum method with respect to the particular hydrogeological setting of the study area.
7. Conclusions In recent years, along with the advancement of the machine manufacturing in many aspects, and rapid advances in technology as a whole, especially information and computer technologies, somewhat smart TBMs have been introduced in the market. As a result, more accurate and comprehensive machine performance parameters are now available for analyzing ground condition and rock properties. By incorporating these parameters into the tunnel as-built geology, the research identified two types of anomalies prevailing throughout the investigation: (a) anomalies associated with cutter-head torque power variations, and (b) anomalies associated with TBM penetration rate variations. Both types of anomalies occurred with considerable frequency along the pilot study area. Although the research was intended to assess the validity of machine performance parameters in detecting open joints and solution channels, it is evident from the data curves that the bedrock fractures and fissures, in which the solution features develop, are also measured. The universalization of such models could lead to erroneous interpretations, since many variants may affect the measurements as previously discussed. One condition that is unique in
316
G.J. Khave / International Journal of Rock Mechanics & Mining Sciences 70 (2014) 308–317
Lot 2 Zagros Tunnel TBM Cutterhead Torque Power 1200
3,000
Cutterhead Torque (KN/m)
Q= 5 lit/s
900 800
Crushed Zone
2,500
2,000
700 600
1,500
500 400
1,000
300 200
500
Post-Grouting Cement Consumption (kg)
Suspected Open Joints
Q=13it/s
Grout Cement Consumption
1000
Q=23 lit/s
1100
Q=120 lit/s
Q=71 lit/s
Cutterhead Torque Power
100 0 4457
4449
4453
4442
4445
4436
4440
4432
4428
4424
4416
4420
4408
4412
4400
4404
4396
4388
4392
4384
4376
4380
4372
4367
4358
4363
0
Tunnel Marking (m) Fig. 13. TBM cutter-head torque power changes coinciding with high cement-take of the grout-holes and actual water discharge measurement at percolation points along the pilot area.
Lot 2 Zagros Tunnel TBM Excavation Penetration Rate
Penetration Rate (mm/min)
35
Crushed Zone
2,500
2,000
30 1,500
25 20
1,000 15 10
500
5
Post-Grouting Cement Consumption (kg)
Q= 5 lit/s
40
3,000
Q=13 lit/s
Grout Cement Consumption
45
Suspected Open Joints
Q=23 lit/s
Q=71 lit/s
TBM Penetration Rate
Q=120lit/s
50
0 4457
4453
4449
4445
4442
4440
4436
4432
4424
4428
4416
4420
4412
4408
4404
4396
4400
4392
4388
4384
4380
4376
4372
4367
4363
4358
0
Tunnel Marking (m) Fig. 14. TBM penetration rate changes coinciding with high cement-take of the grout-holes and actual water discharge measurement at percolation points along the pilot area.
interpretation of the data in this research work is the presence of independent quantitative data at the points of water entries that enhances the evaluation of machine parameters. This data is in the form of actual water discharge monitored at the portal. Although, the reference points for modeling water conduits were empirically verified by an elaborate in-situ testing program, a more comprehensive drilling and grouting plan would be an effective way to substantiate the validity of the modeling in a wider range of geological and hydrogeological conditions. It would also resolve the problem associated with the ambiguities associated with interferences of tightly fractured and fissured zones and erroneous results in softer interlayer, as both penetration rate and torque power parameters, where exhibited inconclusive results at tunnel markings between 4440 and 4458. In spite of the fact that systematic grouting at fixed spacing in tunnel full perimeter is considered to be a routine procedure in conventional post-construction grouting, the economic burden
and time delays are not so readily justified. A properly conceived pre-excavation grouting at TBM face should be adapted and become an integrated component of the design and construction to ensure the project will be successfully and economically completed. As mentioned earlier, exact points and characteristics of rock mass in discharging areas are not morphologically delineated since they are always covered by segments, and water jets were registered along the contacts between segments. Most probably groundwater discharges through the open joints or cavities with limited apertures. Obviously, part of water circulates along the space between segments and rock. Considering all mitigation measures implemented in Zagros Tunnel project, a successful contact grouting and sealing segment joints with hydrophilic mastics are the most reliable measure to prevent initial ground water and H2S gas intrusion into the tunnel. However, water infiltration from the tunnel face is always inevitable.
G.J. Khave / International Journal of Rock Mechanics & Mining Sciences 70 (2014) 308–317
317
Rockmass RQD and Lu Values in Check-holes Before and After Grouting 100 Lu Value Before Grouting Lu Value After Grouting
90
Poly. (Lu Value After Grouting)
80
Poly. (Lu Value Before Grouting)
LU Value
70 60 50 40 30 20 10 0 70
73
75
78
80
83
85
88
90
93
95
98
100
RQD Index Fig. 15. Water pressure test results in check-holes before and after methodic grouting program in pilot study area exhibiting a drastic decline in rock mass Lugeon values.
The study was subject to several limitations, which included; lack of visual control of water infiltration points, ambiguities associated with interferences from softer interlayer, and the inaccuracy of the technique in pinpointing narrow joints within the fissure zones. Despite such limitations, the investigation verified that TBM penetration rate values are valuable indicators of open joints and/or solution channels. However, cutter-head torque power is certainly as equally reliable in delineating jointed and fractured rocks.
Acknowledgement The author would like to express his sincere appreciation to his colleagues on the LAR Consulting project supervision team who assisted in compiling an ample amount of complex data and gave their undivided support. I also dedicate this paper to Mitra, my affectionate consort in my life and work, and to our children, Laya and Kasra.
Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.ijrmms.2014.04. 013. These data include Google maps of the most important areas described in this article.
References [1] Lar Consulting Engineers Co. Report of investigation of tunnel engineering geology. In: Technical reports of phase I Zagros tunnel project; 1991. [2] Barton N. TBM tunnelling in jointed and faulted rock. Rotterdam: Balkema; 2000. [3] Barnett P, Hosoi M, Kheng NB. TBM excavation control system. In: Proceedings of the international conference on tunnelling and trenchless technology. Subang, Malaysia; 2006. pp 89–98. [4] Barton N, Quadros E. Improved understanding of high-pressure grouting effects for tunnels in hard rock. In: Proceedings of ISRM 2003—technology road map for rock mechanics. South African Institute for Mining and Metallurgy; 2004. pp 41–44. [5] Ieman-Sazan Consulting Engineers Co. Report of investigation of technical provisions to confront H2S gas. In: Technical reports of phase II Zagros tunnel project; 2011. [6] Weeks EP. Field methods for determining vertical permeability and aquifer anisotropy. U.S. Geological Survey Professional paper 501-D 1964:196–214. [7] Pars Kaneh Kish Consulting Engineers Co. Report of investigation of tunnel gas and water seepage. In: Technical reports of phase II Zagros tunnel project; 2010. [8] Sahel Consulting Engineers Co. Complementary report of investigation of tunnel engineering geology. In: Technical reports of phase II Zagros tunnel project; 2007. [9] Barla G. Lessons learnt from the excavation of a large diameter TBM tunnel in complex hydrogeological conditions. Politecnico di Torinto, Department of Structural and Geotechnical Engineers; 2000. [10] Kolic D, Mayerhofer A. Segmental lining tolerances and imperfections 2011. [11] Vigil A, Barwart C. Sirvan water conveyance tunnel-segmental lining design. In: Technical reports of phase II Zagros Tunnel no.VC-04-015/006. SELI Technologie; 2005. [12] Henn E, Ganse P, Bandimere S, Smoak G, Warner G. Comparison of penetration test results of grouts made with various ultrafine cement products. In: Proceedings of RETC (SME/AIME); 2001. pp 345–361. [13] Wanner H, Aeberli HU. Tunnelling machine performance in jointed rock. In: Proceedings of 4th ISRM congress, Montreux, Rotterdam, Balkema; 1979. pp 573–579. [14] Glover RE. Studies of groundwater movements. Bur Reclam Tech Memorandum 1960;657.
Contents lists available at ScienceDirect
International Journal of Rock Mechanics & Mining Sciences journal homepage: www.elsevier.com/locate/ijrmms
Delineating subterranean water conduits using hydraulic testing and machine performance parameters in TBM tunnel post-grouting Ghassem Jalilian Khave n Lar Consulting Engineers, 30 Sharifi St., N. Gandi Ave., Vanak Sq., 19699-44311 Tehran, Iran
art ic l e i nf o
a b s t r a c t
Article history: Received 20 November 2012 Received in revised form 28 June 2013 Accepted 25 April 2014
Zagros Water Conveyance tunnel in western Iran crosses a vast unconfined aquifer. This TBM burrowed tunnel has long drained the region’s groundwater which is unusually rich in hydrogen sulfide (H2S) gas. The gas reacts with the tunnel humidity and produces an acidic fume that penetrates the tunnel lining and causes its decay. The exact locations of the discharging conduits and their morphology are not known, since they are concealed by TBM segmental lining. An elaborate post-grouting plan is on the drawing board to reclaim the aquifer, but the cost of a systematic grouting is particularly high and at best this scenario not conclusive. The scope of this paper is to discuss a series of carefully controlled field experiments in a 100 m pilot study area along the tunnel. The scheme manipulates the TBM performance parameters recorded during the tunnel excavation. Based on this information, suitable field models are established that may be interpreted as being associated with either water or air-filled solution channels. & 2014 Elsevier Ltd. All rights reserved.
Keywords: TBM tunnelling Machine performance parameters H2S gas Ecological impact Tunnel post-grouting
1. Introduction Zagros water conveyance tunnel in Kermanshah province of Iran is near the town of Pol-e Zahab. This part of the tunnel ( 26 km long) is regarded as the second lot of a broader plan, which is schemed approximately 48 km long and is 6.73 m in diameter. It has been under construction using a Herrenknecht hard rock double-shield TBM since March 2005. So far, 16 km (61%) of this tunnel has been completed. In the course of tunneling, the machine encountered nearly many extraordinary situations related to continuous intrusion of H2S gas, all of which resulted in a significant reduction in TBM utilization rate and an increase in construction delays, as well as high cost. The encountered circumstances contradicted the preconstruction subsurface investigation results, which had forecasted low to moderate amount of fresh water seepage at a maximum rate of 5–10 l/s/km [1]. One of the major factors affecting the performance of tunnel boring machines is the degree of fracturing of the rock [2]. During excavation, machine performance parameters are continuously displayed and recorded on the control cabin monitors. Fig. 1 shows the TBM display screens. The screens display TBM performance parameters and register TBM mechanical behavior against the excavated material [3]. These screens disclose an assorted set of
n
Tel.: þ 98 21 22000811; fax: þ 98 21 88883905; Cell: þ 98 912 5035910. E-mail address: [email protected]
http://dx.doi.org/10.1016/j.ijrmms.2014.04.013 1365-1609/& 2014 Elsevier Ltd. All rights reserved.
data; e.g. penetration rate, boring time, total thrust, torque power, cutter speed and etc. Based on the careful analysis of these parameters, uniform patterns were established as a model to identify concealed joints and to delineate water-carrying conduits along the tunnel path. The objective is to improvise a model to pinpoint water conduits and fill in their leaks and cracks by postcurtain-grouting. This is in contrast to a systematic grouting approach where grouting is done in a continuum fan pattern regardless of water points of entries [4].
2. Geologic setting According to the structural geology zonations of Iran, Zagros tunnel is located in the core of the Zagros Mountain Range. This region includes simple structures of reverse faulting and symmetrical anticlines and synclines known as Folded Zagros province. The tunnel horizon is situated within the folded zone of Zagros Mountain Range, consisting of sedimentary rock formations at an average depth of 400 m (Hmax 950 m) [1]. The geological unit along the Zagros tunnel in the pilot study area consists of brownish gray limestone of Garou Formation. Overlaying this unit is Gurpi Formation, which is comprised of alternating thin to thick-bedded shale and argillaceous limestone. Some layers of Gurpi Formation are rich in pyrites in the form of nodules [5]. The assembled rock strata changed frequently from hard rock to soft, dry to wet, stable to instable, stick to nonsticky ground (and vice versa), more often than anticipated. Essentially,
G.J. Khave / International Journal of Rock Mechanics & Mining Sciences 70 (2014) 308–317
309
Fig. 1. TBM control cabin main display screens showing excavation data and machine mechanical behaviour against the excavated material.
Fig. 2. Longitudinal geological profile of lot 2 Zagros tunnel along the study area showing the regional GWL. [after 8].
Gurpi sections are aquiclude layers, so there is no substantial groundwater accumulation in these parts of the tunnel. However, Garou sections in the form of anticline serve as an unconfined “water table” aquifer [1]. The geological profile of the project line is illustrated in Fig. 2. The pilot study area is located in the core of a major geological structure known as Ezgelleh Anticline (chainage 04 þ359 to 04 þ458). The tunnel overburden thickness along this area is roughly 155 m. Generally, the phreatic zone is congruent to the areal topography and had stood 105 m above the tunnel crown before its excavation. It was later drained out and the water table plunged below the invert level after completion of the tunnel. The water seepage in low amounts was first experienced at TM 3700. A significant water ingress in the range of Q4110 l/s was later intruded at TM 4157, which rapidly accumulated to Q 315 l/s with further advancement to TM 4435. Advancing deeper into the core of aquifer at TM 8256, the accumulative water seepage totaled Q 730 l/s, with further advancement to TM 13846, 155 more liters of gas bearing water seeped into the tunnel, totaling the tunnel discharge flow at the portal outlet to 900 l/s. This amount of water released 700 ppm hydrogen sulfide gas into tunnel atmosphere. Fig. 3 shows the water discharge rate and the H2S gas concentration at the pilot study area where is marked as Ezgelleh Anticline on the graph. As is evident by the piezometric levels of boreholes 26 and 27 in Fig. 4, the groundwater level on the average has dropped 27 cm/ m of TBM’s advancement before it subsided below the tunnel invert level. This observation is roughly equivalent to the secondary permeability of the medium (hydraulic conductivity reaching up to 0.1–0.2 m/s), which is of similar value with most permeable rocks [6]. The conclusion became a useful criterion to predict the hydraulic characteristics of the aquifer and to estimate the host rock permeability at any given location along the tunnel. Generally, Garou Formation is known to be as the host rock in many major oil (gas) bearing basins further down in the south
provinces of Iran. As a result, some hydrocarbon materials may have migrated or leached out during uplift movements in this area, due to lack of a suitable cap rock and a favorable geological structure to trap the hydrocarbons. Therefore, only traces of residual hydrocarbons have remained in the rock formations in the form of black tarry liquids [7]. These liquids have been frequently observed along the tunnel path, seeping through holes and gaps of the tunnel lining. Hence, the hydrogen sulfide (H2S) gas is most likely an associated component of the remaining gas and oil in existing traps. Hydrogen sulfide liberates from the groundwater; so, its amount depends on the quantity of water inflow. The groundwater discharges through the open joints and solution channels. Obviously, part of water circulates along the space between segments and rock and the water source is difficult to track and handle properly. Prior to this experiment, an elaborate joint survey was carried out from the surface to geometrically define the dominant discontinuities along the Ezgelleh Anticline and to produce a reliable model for the hydraulic characteristics of the rock mass. Overall, four major joint sets identified along the pilot study area (axis N2151). It was also determined that the prevailing joints dipped at an angle (α) of 301 to 801 with dip directions (β) varying from 441 to 2641. Their surface characteristics are undulating, moderately to highly weathered, emax ¼25 mm, at 35–225 cm spacing, and are elongated 1–10 m. Based on the information adopted from over 150 borehole water pressure tests (Lugeon Test) during pre-construction subsurface geotechnical investigation—interpolated most compatible with actual geological conditions (elevation, structural and lithological) along the bored tunnel, average permeability coefficient of the intact rock is calculated between 10 5 to 10 6 cm/s [1]. In general, 17% of the Lugeon tests were reported as impervious and 42% resulted in “laminar” flow. Turbulent flow that generally represents open joints and “washout” that refers to washing the joint filling were tallied at 21% and 12%, respectively. Dilation flow
310
G.J. Khave / International Journal of Rock Mechanics & Mining Sciences 70 (2014) 308–317
Histogram of Portal Water Discharge and TBM Cutterhead Gas Concentration 700
900 Total water Discharge
650
Gas Concentration
600 550
700 600
500
Pilot Study Area Ezgelleh Anticline
450 400
500
350 400
300 250
300
200 200
150
H2S Gas Concentration(ppm)
Total Water Discharge (lit/s)
800
100
100
50 0
3, 7 3, 68 8 3, 50 8 3, 63 8 3, 63 86 4, 3 1 4, 15 3 4, 06 4 4, 46 5 4, 02 8 5, 40 19 5, 6 9 7, 25 0 7, 74 8 8, 64 2 8, 31 3 8, 12 59 8, 9 9 9, 30 10 353 , 10 021 , 11 738 , 12 536 , 13 394 , 13 222 , 13 334 , 13 782 , 13 858 , 13 923 , 14 941 , 14 118 , 14 233 , 14 309 , 14 393 , 14 643 ,7 34
0
Tunnel Chainage Fig. 3. Tunnel accumulative water discharge rate histogram at portal showing the hydological condition and approximate location of the pilot study area.
Zagros Tunnel Areal Aquifer Fluctuation Corresponding to TBM Advancement
DrawdownTable
750
Groundwater Elevation (m)
730 710 690 670 650 630 610 590 570 550 21F
22F
23
26
27
Project Borehole Piezometric Levels on November 2003 Prior to Tunnel Excavation Project Borehole Piezometric Levels on July 2007 and TBM Location Project Borehole Piezometric Levels on December 2008 and TBM Location Project Borehole Piezometric Levels on June 2009 and TBM Location
28
29F
30F
30
Borehole ID. No.
Fig. 4. Tunnel construction impact by January 2008 on Ezgelleh anticline groundwater hydrostatic level and its drawdown rate as elucidated by project borehole piezometers.
as governed by elastic behavior of joint openings, which is not ordinarily considered a normal rock behavior at typical groundwater pressure but is expected to take effect at seasonal fluctuations of groundwater hydrostatic head, accounted for 8% of test results. Thus, it is interpreted, the host rock has low permeability properties and any occurring water seepage will be transmitted through a secondary network of interconnecting open joint systems. Nevertheless, approximately 33% of all joints encountered could provide a suitable medium for water migration [1]. The cores of Garou Limestone, as observed in borehole 27, exhibit bedding planes spaced from 100 mm to 1 m. The orientation of the bedding planes varied from perpendicular to the core to approximately 251 from this. Inclined joints within the core varied about 401 to 701 to the core axis and, typically, there were four or five inclined joints in each of the core boxes. The condition of the core was not uniform, with frequent calcite lenses and stylolitic surfaces [8].
3. Tunnelling impact on ecological conditions Excavation of tunnels usually affects the natural surroundings to a great extent. Its intensity depends on the location of the tunnel in relation to the region’s aquifer and the hydrogeological behavior of the rock mass in respect to tunnel. In the case of Zagros tunnel, the excavation work has adverse affects on all aspects of the project construction because of repeated intrusion of hydrogen sulfide gas contaminated groundwater. To make matters more complex, the project is located in a semi-arid area, and constant drainage of the aquifer (416 l/s) by the advancing tunnel is rapidly exhausting the region’s water resources. The implemented mitigation measures have considerably reduced the adverse affects. Nevertheless, the interference with the ecological setting is so severe that demands an immediate and comprehensive remedial action. In addition, droughts in recent years have strained the already imbalanced regional groundwater reserves
G.J. Khave / International Journal of Rock Mechanics & Mining Sciences 70 (2014) 308–317
311
Zagros Tunnel Regional Water Table Drawdown Radius of Influence Permanently dried springs Initial Static Water Table Prior to Tunnel Excavation Water Table Drawdown Abdallan spring influenced by seasonal precipitations
3.5km Zone with Springs Influenced By Water Table Drawdown Draining Zagros Tunnel Fig. 5. Regional groundwater table drawdown showing tunnel radius of influence (cone of depression) with springs effected by water table drawdown. [after14].
and have dried many local springs. Hence, a large number of permanent springs have dried up and in a few areas that the tunnel has not fully affected the groundwater table; water discharge has trickled to 50–400 l/min. As a result, many villages have suffered serious water shortage and their habitants reluctantly have relocated to other areas. Most areal springs are directly or indirectly connected to the joint systems that have developed to a network of subterranian channels by solutional weathering. As it was also concluded by monitoring of the local springs, the tunnel maximum radius of influence is calculated approximately 3.5 km (Fig. 5). For instance, Abdallan spring, which is one of the prominent springs in the region, is roughly the same distance from the tunnel axis. Its discharge rate fluctuations are consistent with the seasonal precipitations and the tunnel advancing rate in the aquifer. For example, most springs in the area show a gradual increase in discharge rate within 15 days from the beginning of every wet season and a rapid decline after similar time lap when the wet season expires. Fig. 6 illustrates Abdallan spring monthly discharge histogram. The overall depletion of the aquifer is evident by the histogram declinig trend.
4. Tunnel post-excavation grouting It is fundamentally true that successful contact grouting (liner backfilling by pea-gravelþcement grout injection) in TBM tunnelling is crucial in reducing groundwater infiltration [9]. Hence, in Zagros tunnel case a successful contact grouting could dampen the intrusion of water into the tunnel and decrease the intensity of gas emission into atmosphere. As a result, less health risk for the tunnel crew and more life expectancy for the tunnel lining. On the other hand, taking the hexagonal geometry (flexible) of the Zagros tunnel lining into account, the importance of segment installation tolerances and imperfections demanded particular attention. Hence, the process of quality assurance that would control both the geometrical dimensions and fitting of segments into lining rings has an important role to compensate for some of the inherent deficiency of hexagonal lining in restraining water intrusion into the tunnel [10]. Ordinarily, to block the leakage between segments, efforts were made to backfill the space between segments and rock with as much as pea-gravel, and fill the voids with enough cement slurry to form a secondary concrete lining. Despite the best efforts, filling
the overt and invert areas in high water seepage zones was not possible due to gravitational purge of grout mix from holes in overt area and continuous leakage of grout mix from submergedunsealed gaps in the invert area. Fig. 7 shows three representative core boxes sampled during water pressure tests. They reveal the extent of contact grouting success in high water areas. The alternative measures that were carried out were polyurethane based foam grouting and resin injection. Pre-drainage and plugging the segment joints with hydrophilic sealants, which all proved incredibly time consuming and in particular, costly. To reclaim the aquifer and to rehabilitate the local springs, and to also prevent tunnel lining from further decay by H2S destructive forces; an elaborate post construction curtain-grouting program was schemed. The notion was to supplement the existing contact (back-fill) grouting, to seal off the remaining spot leakages most especially where the contact grouting had not satisfied the required average tightness of the tunnel lining, e.g. 6 l/s/km [11]. It was originally perceived that a sequential (systematic) grouting of the tunnel from eight holes in the circumference was the best approach. The program comprised radially fan pattern, perpendicular to the centerline. The holes were to drill systematically to 5 m depth and grout with cement mix, up to 5 bar pressure at 1.6 m lateral spacing. To complete the consecutive rings, drilling in segment center-points and injecting micro cement to depth of 4 m at higher pressure (pmax ¼7.5 bar) to penetrate difficult to reach areas was deemed necessary [12]. To be able to react on disintegrated rock mass and deficiencies of the bedding plain, an immediate consilidation borehole grouting was to be applied in the zone where the P-gravel is applied under normal condition. The immediate consolidation grouting was to be carried out via short boreholes ( 1.0 m) by a single mobile grout pump. The estimated cost of the designed curtain grouting in a systematic method for the first 16 km of the tunnel was enormous and the time delay could not be justified within the contract time-frame. To achieve a more reasonable timetable and attain a more assuring result for the water-tightening plan, a swift approach and an accurate procedure to distinguish the location of the water conduits and to individually plug them off, was deemed the sole alternative. The procedure must to avoid lengthy exposure of crew to tunnel gassy condition and to facilitate grouting work in tunnel constraining work condition, e.g. limited workspace, high pressure water inrush, hydrogen sulfide gas pollution and logistic problems. Voluminous information generated during tunnel excavation in the form of machine performance parameters (excavation data
312
G.J. Khave / International Journal of Rock Mechanics & Mining Sciences 70 (2014) 308–317
Abdallan Village Spring Discharge Histogram 1800
Before Tunnel Excavation After Tunnel Excavation 1600
Seasonal Precipitations Marker (+) Beginning (-) Ending
1400
Q (lit/min)
1200 1000 800 600 400 200
A
ug Se -06 p O -06 c N t-06 ov D -06 ec Ja -06 n Fe -07 b M -07 ar A -07 p M r-0 ay 7 Ju -07 nJu 07 A l-0 ug 7 Se -07 p O -07 c N t-07 ov D -07 ec Ja -07 n Fe -08 b M -08 a A r-08 p M r-0 ay 8 Ju -08 nJu 08 A l-0 ug 8 Se -08 p O -08 c N t-0 ov 8 D -08 ec Ja -08 n Fe -09 b M -09 ar A -09 p M r-0 ay 9 Ju -09 nJu 09 A l-0 ug 9 Se -09 p O -09 c N t-09 ov D -09 ec Ja -09 n Fe -10 b M -10 a A r-10 p M r-1 ay 0 Ju -10 nJu 10 A l-1 ug 0 Se -10 p O -10 c N t-1 ov 0 D -10 ec -1 0
0
Date (Month-Year) Fig. 6. Histogram of Abdallan spring monitoring during the tunnel construction, showing the monthly discharge rate and the duration of seasonal precipitation periods as reference.
5. Experimental grout method
Fig. 7. Cores showing segment concrete and P-gravel zone infilling behind tunnel lining that often left partially to completely empty due to grout mix washouts in high water pressure zones. The top photo represents a complete recovered core. The middle photo shows a gap in core between segment concrete and bedrock left by the grout washout, and the bottom picture shows a partially filled P-gravel zone between rock and tunnel segmental lining.
and machine mechanical behaviour against the excavated ground curves) were determined a viable approach to be explored. Therefore, the machine performance data was gathered and carefully examined to delineate some logical correlations with the actual ground conditions. Hence, identify the open joints and the cracks that are capable of transmitting water. The analysis took into account all relevant parameters on TBM data displayers, but only two parameters in particular; Excavation Penetration Rate and Cutter-head Torque Power parameters exhibited pronounced changes that also conformed to the corresponding water pressure test (WPT) results, which incorporated the location of tunnel known water intrusion points. The information analysis required great care to avoid very complicated ambiguities associated with interferences from softer rock in interlayer and probable errors in data keeping during the excavation phase. Fig. 8 illustrates the schematic drawing of the preferred array and angle of incidence adapted in methodic (localized) tunnel grouting in the pilot study area.
As mentioned before, three prominent master joints and a bedding plane were recognized along the pilot study area. The angle of incidence of most prominent joint set (J2) dips about 791 in a southeasterly direction of N1741. The J2 average aperture is measured 15–25 mm (moderately narrow) and their spacing is between 150 cm to 225 cm (very widely spaced). The secondary and tertiary master joints with respect to their percentage of pole concentrations are classified as J3 and J4. They are generally spaced at 150–225 and 45–100 cm (e¼10–20 and 0.1–1.0 mm), respectively. Table 1 diplayes the summary of the indigenous rock mass discontinuity characteristics along the pilot study area [8]. The main tunnel extends in a NE-SW direction (N2151) and the areal bedding plane intersects the tunnel axis at an angle of (α ¼641). In the first stage of experiment, 141 boreholes were systematically drilled and water pressure tested based on the original pattern. The results were plotted in an X–Y graph. Based on the obtained data, the lugeon values were classified in four categories (Fig. 9). To summarize, a total of 51% of the lugeon tests attained low permeability. Twenty percent classified as medium permeability, 12% exhibited high permeability and 17% of the tested holes displayed very high permeability. The graphical analysis revealed that in a systematic approach, when test holes are aligned regardless of the joint orientation, the possibility of intercepting all of the dominating discontinuities is more likely incidental. To enhance the method, a series of boreholes at different azimuths (þ301 to 301) were drilled. The goal was to intercept as many cracks and fissure planes as possible and perform water pressure tests. From the lugeon values, an optimum azimuth angle which apperaed most appropriate to detect air or water filled open joints was determined. In short, the borehole azimuth that intercepted most number of open joints and active water conduits was α ¼ 301. This particular azimuth had resulted in the highest number of medium to very high lugeon values. Fig. 10 illustrates the concentration of boreholes at different angles (azimuths) and their corresponding lugeon values.
G.J. Khave / International Journal of Rock Mechanics & Mining Sciences 70 (2014) 308–317
313
Fig. 8. Plan and profile of the preferred array and the pattern adapted in post- grouting procedure in pilot area. The Plan shows locations of different trials for favorable WPT azimuthal angle. Table 1 Summary of the pilot area rock mass discontinuity characteristics [8]. Parameter
Type Joint set
Bedding joint Conjugate shear joint
J1 J2 J3 J3
Attitude (degrees)
Roughness
βJ
αJ
JRC
Jr
061 174 044 264
30 79 71 51
4–6 6–8 6–8 6–8
3.0 1.5 1.5 2.0
Weathering (Ja)
Aperture e(mm)
Spacing (cm)
Length (m)
3.0 2.0 2.0 1.0
1–10 15–25 10–20 0.1–1.0
35–65 150–225 150–225 45–100
3–10 1–3 3–10 1–3
Tunnel Rockmass Lugeon Values in Systematic Approach 100 90 80
VERY HIGH= 17%
LU Value
70 60 50 40
HIGH= 12%
30 20
MEDIUM= 20%
10 LOW=51%
0 0
10
20
30
40
50
60
70
80
90
100
110
120
130
140
150
No. of Tests Fig. 9. In-situ water pressure test results in a systematic approach with majority of the test results are classified as low to moderate values, indicating an incidental interception of the discontinuities.
Monitoring a new round of WPT with the selected azimuth (α ¼ 301), it was concluded that lugeon values have significantly increased by almost 40%. The assumption is based on 118 lugeon tests, conducted. In general, 26% of the tests showed low perneability,
26% displayed medium permeability, 26% had high permeability and finally 22% indicated a very high permeability. Thus, the method had good merits and proven to be fairly reliable. Fig. 11 illustrates the graphical analysis of the lugeon values in the methodic approach.
314
G.J. Khave / International Journal of Rock Mechanics & Mining Sciences 70 (2014) 308–317
Incorporating flow rate, pressure and total volume of cement consumption during the actual grouting stage, pressure-flow diagrams were developed. The diagrams also indicated if leaks back to the tunnel perimeter or connections to adjacent holes had occurred or if excessive pressure had caused damage to the rock mass. The drilling for grouting was made in 6.5 m linear stages and a down-stage procedure was utilized. The drilling holes with huge water leakages and material transport were performed in ascending stage with collar. To avoid local empty rooms and ‘air pillows” above backfilling concrete close to crown, contact grouting with mortar and curtain grouting were done, simultaneously as post construction program. Thus, the grout hose with simple valves type “Tube-a’-manchette” (sleeve port pipes) were mounted, as well. In following stage, rock just behind segments was grouted with a low pressure (5 bar) to avoid damage of segments and grout mix leakage into the tunnel. For deeper stages, recommended pressure was 7.5 bar. In some cases, the pressure was raised to 10 bar to overcome prevailing pressure. The ambient pressure varied from 7 to 9 bar depending on depth below water surface
and volume of water penetration. And injection pressure did not surpass the ambient pressure by more than one bar. The injection equipment used in Zagros tunnel post-grouting as illustrated in Fig. 12 included two double piston injection pump KOS 1000, two double piston injection pump KOV 1000, two electric power rotary drill rigs, two ultra mixers UM 10, six flat rail cars with separate containers and lift frames, a cement mortar mixer, an electro-hydraulical power units totaling 400 kW, a reverse tanker water tanker with a capacity of 2000 l, a cement holding silo with a capacity of 2.5 t, injection pressure and electronic flow measuring units with a central operating panel, and injection pressure and electronic flow measuring units with central operating panel.
6. Linking machine performance parameters to ground anomalies Wanner and Aeberli [13] cited to penetration rates between 50 and 100% higher than the daily average when open joints were
Tunnel Rockmass In-Situ Water Pressure Test Results at Different Azimuths 100 90 80 VERY HIGH
70
LU Value
60 50 HIGH
40 30 20
MEDIUM
10 LOW
0 -40
-30
-20
-10
0
10
20
30
40
Borehole Alignment Azimuths (Degree) Fig. 10. Tunnel in-situ WPT results at different azimuths in respect to tunnel longitudinal axis. The optimum azimuth proved to be α ¼ 301 signifying most number of open jonts intercepted.
Tunnel Rockmass Lugeon Value In Methodic Approach 100 90 80
VERY HIGH= 22%
Lu Value
70 60 50 HIGH= 26%
40 30 20 MEDIUM= 26%
10 LOW= 26%
0 0
10
20
30
40
50
60
70
80
90
100
110
120
130
No. of Tests Fig. 11. In-situ water pressure test results in methodic approach with majority of the results are classified as moderate to very high values, exhibiting a high precision in intercepting major discontinuities.
G.J. Khave / International Journal of Rock Mechanics & Mining Sciences 70 (2014) 308–317
315
Fig. 12. Zagros tunnel post-grouting equipment setup. [5].
present. They also suggested that rough-walled tension joints were least helpful to penetration rate, especially if they were healed. To examine the validity of their proposed criterion, a 100 m span along the Zagros tunnel with known high water discharge between TM 4359 to TM 4458 was appointed as a pilot study area (Fig. 2). Field reconnaissance in conjunction with local rockmass discontinuity survey and TBM performance parameters supplemented by tunnel daily water discharge rate were compiled to individual curves to establish a fixed pattern for the location of known high water bearing zones. Recognition of subsurface materials or conditions from TBM performance parameters was ambiguous at times and distinguishing solution channels from their surrounding rocks based on machine parameters required a comprehensive knowledge of TBM mechanism. This is due to influential effects of contingent machine parameters that are under direct TBM operator control and most significantly, the TBM’s overall performance status and equipment efficiency, e.g. disc cutters, hydraulic jacks, electromotor, power usage, etc. In order to test the validity of technique and establish a field model for further scrutiny, the approximate location of five known conduits with discharge rates ranging from 5 to 120 l/s in the study area were transposed on the TBM “Advance Speed” and “Cutter-head Torque” curves. Upon examination of each data curve, it was learned that the cutter-head torque power tend to drop significantly as TBM neared the known water transmitting points. It was also apparent from the TBM advance speed that the same points tend to peak as the TBM cutter-head crossed the center-point of the known solution channels at a right angle. The close approximation of the groundwater percolation points to the slope changes in both cases was very encouraging, considering the vague relationship between the cutter-head torque power and the TBM advance speed rate. In general, points of slope change in data curves indicated face rock strength (resistive) contrast, whereas linear data were interpreted as evidence of uniform strength. In addition, sharp declines in cutter-head torque power were marked as air or water-filled solution channels, where as steep increase in penetration rate interpreted as the same circumstances. Moreover, the tension cracks that had showed high permeability (Q¼13–23 l/s at TM 4440-4458) displayed rather indistinctive signals in both penetration rate and cutter-head torque power data. Thus, pinpointing the exact location of tension cracks will merely depend on the engineer’s judgment. Although not particularly evaluated in a great extend, it is likely that similar anomalies could prevail by TBM’s total trust power, which is rather proportionate to TBM operator performance and skill.
After establishing the applicability of the technique over the known solution channels, the effort was directed toward detecting other suspected networks of open joints along the same span. Eighteen additional grout-holes were drilled and tested using the set criteria. Figs. 13 and 14 are the graphical presentations of the data points obtained from the machine parameters, the consolidation (curtain) grout cement take and the approximate location of points of water entry into the tunnel. Thus, the contrast in machine performance curves combined with grout cement takes and water infiltration fluctuations at the tunnel portal revealed anomalies correlating with the individual water conduits. This demonstrated that the excavation data and machine mechanical behaviour against the excavated material can differentiate between solid rock and the open joints and fissure zones in fractured rock. Subsequently, 32 additional check-holes were drilled to assess the efficiency of the devised method. They were drilled on both sides of the grout-holes. Water pressure tests were conducted before and after grouting procedure at each hole. Fig. 15 illustrates the degree of ground improvement at grout check holes after conducting the methodic pattern. It clearly manifests that ground improvement in the range of 50% can be accomplished by methodic grouting, which is conceived as an optimum method with respect to the particular hydrogeological setting of the study area.
7. Conclusions In recent years, along with the advancement of the machine manufacturing in many aspects, and rapid advances in technology as a whole, especially information and computer technologies, somewhat smart TBMs have been introduced in the market. As a result, more accurate and comprehensive machine performance parameters are now available for analyzing ground condition and rock properties. By incorporating these parameters into the tunnel as-built geology, the research identified two types of anomalies prevailing throughout the investigation: (a) anomalies associated with cutter-head torque power variations, and (b) anomalies associated with TBM penetration rate variations. Both types of anomalies occurred with considerable frequency along the pilot study area. Although the research was intended to assess the validity of machine performance parameters in detecting open joints and solution channels, it is evident from the data curves that the bedrock fractures and fissures, in which the solution features develop, are also measured. The universalization of such models could lead to erroneous interpretations, since many variants may affect the measurements as previously discussed. One condition that is unique in
316
G.J. Khave / International Journal of Rock Mechanics & Mining Sciences 70 (2014) 308–317
Lot 2 Zagros Tunnel TBM Cutterhead Torque Power 1200
3,000
Cutterhead Torque (KN/m)
Q= 5 lit/s
900 800
Crushed Zone
2,500
2,000
700 600
1,500
500 400
1,000
300 200
500
Post-Grouting Cement Consumption (kg)
Suspected Open Joints
Q=13it/s
Grout Cement Consumption
1000
Q=23 lit/s
1100
Q=120 lit/s
Q=71 lit/s
Cutterhead Torque Power
100 0 4457
4449
4453
4442
4445
4436
4440
4432
4428
4424
4416
4420
4408
4412
4400
4404
4396
4388
4392
4384
4376
4380
4372
4367
4358
4363
0
Tunnel Marking (m) Fig. 13. TBM cutter-head torque power changes coinciding with high cement-take of the grout-holes and actual water discharge measurement at percolation points along the pilot area.
Lot 2 Zagros Tunnel TBM Excavation Penetration Rate
Penetration Rate (mm/min)
35
Crushed Zone
2,500
2,000
30 1,500
25 20
1,000 15 10
500
5
Post-Grouting Cement Consumption (kg)
Q= 5 lit/s
40
3,000
Q=13 lit/s
Grout Cement Consumption
45
Suspected Open Joints
Q=23 lit/s
Q=71 lit/s
TBM Penetration Rate
Q=120lit/s
50
0 4457
4453
4449
4445
4442
4440
4436
4432
4424
4428
4416
4420
4412
4408
4404
4396
4400
4392
4388
4384
4380
4376
4372
4367
4363
4358
0
Tunnel Marking (m) Fig. 14. TBM penetration rate changes coinciding with high cement-take of the grout-holes and actual water discharge measurement at percolation points along the pilot area.
interpretation of the data in this research work is the presence of independent quantitative data at the points of water entries that enhances the evaluation of machine parameters. This data is in the form of actual water discharge monitored at the portal. Although, the reference points for modeling water conduits were empirically verified by an elaborate in-situ testing program, a more comprehensive drilling and grouting plan would be an effective way to substantiate the validity of the modeling in a wider range of geological and hydrogeological conditions. It would also resolve the problem associated with the ambiguities associated with interferences of tightly fractured and fissured zones and erroneous results in softer interlayer, as both penetration rate and torque power parameters, where exhibited inconclusive results at tunnel markings between 4440 and 4458. In spite of the fact that systematic grouting at fixed spacing in tunnel full perimeter is considered to be a routine procedure in conventional post-construction grouting, the economic burden
and time delays are not so readily justified. A properly conceived pre-excavation grouting at TBM face should be adapted and become an integrated component of the design and construction to ensure the project will be successfully and economically completed. As mentioned earlier, exact points and characteristics of rock mass in discharging areas are not morphologically delineated since they are always covered by segments, and water jets were registered along the contacts between segments. Most probably groundwater discharges through the open joints or cavities with limited apertures. Obviously, part of water circulates along the space between segments and rock. Considering all mitigation measures implemented in Zagros Tunnel project, a successful contact grouting and sealing segment joints with hydrophilic mastics are the most reliable measure to prevent initial ground water and H2S gas intrusion into the tunnel. However, water infiltration from the tunnel face is always inevitable.
G.J. Khave / International Journal of Rock Mechanics & Mining Sciences 70 (2014) 308–317
317
Rockmass RQD and Lu Values in Check-holes Before and After Grouting 100 Lu Value Before Grouting Lu Value After Grouting
90
Poly. (Lu Value After Grouting)
80
Poly. (Lu Value Before Grouting)
LU Value
70 60 50 40 30 20 10 0 70
73
75
78
80
83
85
88
90
93
95
98
100
RQD Index Fig. 15. Water pressure test results in check-holes before and after methodic grouting program in pilot study area exhibiting a drastic decline in rock mass Lugeon values.
The study was subject to several limitations, which included; lack of visual control of water infiltration points, ambiguities associated with interferences from softer interlayer, and the inaccuracy of the technique in pinpointing narrow joints within the fissure zones. Despite such limitations, the investigation verified that TBM penetration rate values are valuable indicators of open joints and/or solution channels. However, cutter-head torque power is certainly as equally reliable in delineating jointed and fractured rocks.
Acknowledgement The author would like to express his sincere appreciation to his colleagues on the LAR Consulting project supervision team who assisted in compiling an ample amount of complex data and gave their undivided support. I also dedicate this paper to Mitra, my affectionate consort in my life and work, and to our children, Laya and Kasra.
Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.ijrmms.2014.04. 013. These data include Google maps of the most important areas described in this article.
References [1] Lar Consulting Engineers Co. Report of investigation of tunnel engineering geology. In: Technical reports of phase I Zagros tunnel project; 1991. [2] Barton N. TBM tunnelling in jointed and faulted rock. Rotterdam: Balkema; 2000. [3] Barnett P, Hosoi M, Kheng NB. TBM excavation control system. In: Proceedings of the international conference on tunnelling and trenchless technology. Subang, Malaysia; 2006. pp 89–98. [4] Barton N, Quadros E. Improved understanding of high-pressure grouting effects for tunnels in hard rock. In: Proceedings of ISRM 2003—technology road map for rock mechanics. South African Institute for Mining and Metallurgy; 2004. pp 41–44. [5] Ieman-Sazan Consulting Engineers Co. Report of investigation of technical provisions to confront H2S gas. In: Technical reports of phase II Zagros tunnel project; 2011. [6] Weeks EP. Field methods for determining vertical permeability and aquifer anisotropy. U.S. Geological Survey Professional paper 501-D 1964:196–214. [7] Pars Kaneh Kish Consulting Engineers Co. Report of investigation of tunnel gas and water seepage. In: Technical reports of phase II Zagros tunnel project; 2010. [8] Sahel Consulting Engineers Co. Complementary report of investigation of tunnel engineering geology. In: Technical reports of phase II Zagros tunnel project; 2007. [9] Barla G. Lessons learnt from the excavation of a large diameter TBM tunnel in complex hydrogeological conditions. Politecnico di Torinto, Department of Structural and Geotechnical Engineers; 2000. [10] Kolic D, Mayerhofer A. Segmental lining tolerances and imperfections 2011. [11] Vigil A, Barwart C. Sirvan water conveyance tunnel-segmental lining design. In: Technical reports of phase II Zagros Tunnel no.VC-04-015/006. SELI Technologie; 2005. [12] Henn E, Ganse P, Bandimere S, Smoak G, Warner G. Comparison of penetration test results of grouts made with various ultrafine cement products. In: Proceedings of RETC (SME/AIME); 2001. pp 345–361. [13] Wanner H, Aeberli HU. Tunnelling machine performance in jointed rock. In: Proceedings of 4th ISRM congress, Montreux, Rotterdam, Balkema; 1979. pp 573–579. [14] Glover RE. Studies of groundwater movements. Bur Reclam Tech Memorandum 1960;657.