 |
 |
Analysis of Predicted and Actual
Associate Professor, Civil Engineering Department, National University of and Dean Civil Engineering MCE, National University of |
ABSTRACT
Major transport corridors in Pakistan extend in a North-South direction as per distribution of population and economic activities. One of the major transport route is Indus Highway that runs on the west bank of river Indus and stretches approximately 1200 km from Karachi to city of Peshawar bordering Afghanistan. The highway generally runs through the alluvial plains of Indus. About 10 km northeast of Kohat city, however, Bosti Khel mountains intercept smooth traffic flows. A tunnel was therefore proposed at this location to circumvent the traffic congestion. The mountain range at Bostikhel has extremely complex geology. Intensive geotechnical studies, therefore, were undertaken prior to initiation of the tunnel project. These studies included geophysical exploration and other survey techniques. Results of this survey were reviewed by a team of International experts. Predictions were accordingly made pertaining to probable composition of various rock types expected to be encountered during tunneling. The total cost of any tunneling project is highly dependent on rock quality and it is imperative that the predicted and in situ geology match closely. After execution of the project a study was undertaken at National University of Sciences & Technology, Risalpur Campus, to determine success/failure of the techniques used for predicting tunnel geology. This paper presents results of this analysis. It is concluded that the predictions based on limited geophysical exploration and general observations can be in substantial error. Superior survey techniques therefore shall be used for predicting tunnel geology. Furthermore it is concluded that New Austrian Tunneling Method(NATM) can be used successfully on a project of complex geology.
INTRODUCTION
The Kohat tunnel feasibility evaluation was initiated in 1973. Several studies were undertaken and alternative plans for different tunnel routes were studied to determine the most feasible solution. Three alternate routes A , B & C were considered to prepare a firm recommendation. Alternative B for tunnel construction was chosen due to competent and sound geology, straight and aesthetically pleasing alignment and economy. Main factors governing any tunnelling project are the lethology and structural geology of the site. The overall cost of the project is highly dependent on the site geology. Prior to commencement of construction work, therefore, every effort is made to predict the site geology as accurately as possible. At Kohat tunnel Refraction survey was done at project feasibility stage to predict tunnel geology. Internationally renowned geologists and geophysicists were engaged to study geology of Bostikhel mountains. These experts were required to predict tunnel geology based on their observations and records of refraction survey. Based on their conclusions cost estimates were prepared and internal rate of return was computed. The predicted geology indicated that the project is economically feasible and was accordingly taken up for execution. During construction the geology was recorded for each one meter length of the tunnel. The main aim of the study reported herein is to compare predicted and actual geology encountered during tunnel construction. It may be pointed out that the cost estimates are made on the basis of the predicted strata and, therefore, significant deviation of actual geology from the predictions can prove to be economically colossal and wreck the project during execution. Another aim of the paper is to report success of New Austrian Tunneling Method (NATM) on a project of fairly difficult geologic setting.
GENERAL GEOLOGY OF THE TUNNEL AREA
Kohat Tunnel has been constructed in a difficult geological setting. Shale and lime stone are predominant rock types. Tectonically the area is part of Himalayan folded belt. Various active and inactive faults are located in this area. Geologically the range is completely deformed with all types of structural anomalies present. Along the tunnel alignment the oldest rocks exposed pertain to Jurassic era and are mainly limestone. These are followed by Chichali Formations. The lower part of Chichali Formation is of upper Jurassic age and the formation grades into cretaceous. Chichali Formation is represented by sandy shales and glauconitic sandstone. Overlying are the rocks of Lamshiwal Formation and are mainly sandstone and shales. These are followed by Kawagarh Formation of upper cretaceous represented by shale and limestone in this area. Paleocene Is represented by Lockhart Formation that is mainly limestone and of middle Paleocene while upper Paleocene are Patala Shales with subordinate sandstone. Murree Formation represented by shale and sandstone inter-beds represents Miocene and has a thrusted contact with Jurassic limestone in the area. The area has a typical tectonic and geological setting and is completely deformed dividing the range into several tectonic slices that are separated by many North dipping imbricate thrusts. Small scale folds are common in thin bedded limestone. Two phases of deformation are recognizable, the first phase of folding giving rise to large scale recumbent or over turned south verging folds having roughly East-West axes. The other phase appears to be mild and developed open and up right folds. The Main Boundary thrust is a strong member of the fault system and the Northern fault is a south dipping hack thrust Joint. It is concluded that the formation along tunnel alignment has evolved due to the geological events depicted in Figure 1 resulting in the present geological formation.
Figure 1. Evolution of tunnel geology
The limestone near North Portal is massive to thick bedded but has broken down into huge boulders as a result of weathering and also because of the erosion of the underlying shale. The next rock unit is sandstone which is medium to coarse, thin bedded and glauconitic. Most of the sandstone has been eroded but some beds project out. This rock unit forms a depression or saddle and occupies a low position in the profile. One sandy shale layer is also present near the North Portal. This sandy shale layer has three exposures near the south Portal. The rest of the rock consists of limestone which becomes thin bedded towards the top and has been folded into a syncline followed by an anticline. The limestone is jointed and fractured. Almost 50% of these joints are open. The South Portal side is intensely deformed because of the proximity of Main Boundary Thrust Fault. Major rock types at south portal are similar to the ones encountered at North portal.
Seismic survey along Kohat tunnel alignment was conducted by the Oil & Gas Development Corporation (OGDC) in 1990. Reanalysis of seismic survey was carried out in August, 1997. The objective of seismic survey was to provide subsurface information for Kohat tunnel geological features. The exploration was carried out along two lines, main and sub line. The details are given in Table 1.
Table 1. Details of geophysical survey lengths
| |
| Seismic survey section along length of tunnel | |
| |
| a) NS-II line | 230 m long main-line |
| b) NS-II line | 230 m long main-line |
| c) NS-III line | 230 m long main-line |
| d) NS-IV line | 230 m long main-line |
| e) NS-V line | 230 m long main-line |
| U NS-VIII line | 230 m long main-line |
| g) EW-I line | 220 m long sub-line |
| h) EW-I1 line | 460 m long sub-line |
| j) EW-III line | 230 m long sub-line |
| Total | 2,290 m |
| |
It was concluded that limestone with frequent intercalation shales and sandstone is distributed in survey area. Results of time-distance curve indicated that the first stratum (velocity of refraction wave, vp=0.20 - 0.70 km/s) is soil with rock fragment and its thickness ranges between 5.0 - 15.0 meters. The second stratum (vp =1.00-1.80 km/s), represents highly weathered/fractured fault zone. The third stratum (vp =2.10 - 2.70 kmls) is weathered fault zone and its depth ranges between 10 - 30 meters from ground surface. The fourth to sixth strata are base rocks in this area.
Boring cores were tested to obtain Unconfined compression strength (Scs), Rock Quality Designation (RQD), velocity of refraction waves, Absorption, Specific gravity and other pertinent rock parameters. The average value of vp was obtained as 2.7- 5.5 km/s, coefficient of fissure of rock mass is 0.16 - 0.80. The velocity of refracted waves and coefficient of fissure for fracture/fault zone is 1.00 – 2.70 km/s and 0.80 respectively. The test results of unconfined compressive strength of core samples yielded average, maximum and minimum Strength of 66, 145 and 42 Mpa respectively. The tunnel rocks were subsequently classified in accordance with NATM (New Austrian Tunneling Method) Classification system based on coefficient of fissure, Velocity of refraction wave of rock-mass, Semi rock-mass strength (Sc) and Physical characteristics of Rock. The rocks at Kohat tunnel classify as CI, CII & DI. Rock type CI is primarily competent limestone, CII is a mixture of sandstone and limestone with shale intercalations and the weakest of all rocks is classified as D1. It includes shales and very weak sandstone. Following proportions of these rock types were predicted.
Table 2. Predicted distribution of rocks
| |
| Rock type | Predicted length |
| |
| CI | 751.0 m |
| CII | 358.0 m |
| DI | 713.5 m |
| South Portal | 41.5 m |
| North Portal | 2l.0 m |
| Total | 1885.0 m |
| |
Figure 2 presents the distribution of these predicted rocks along the tunnel length.
Figure 2. Predicted distribution of various types of rocks
KOHAT TUNNEL CONSTRUCTION
Japanese standard was adopted in the tunnel design as well as execution. Japanese tunneling practice is based on New Austrian Tunneling Method (NATM). In Japanese construction practice the tunnel design is periodically reviewed and revised in light of actual geological strata. The main factors considered at the design stage are rock type, seismic velocity, compressive strength of rock, overburden pressure, core condition, rock quality designation (RQD). The main factors reviewed at the construction stage are: geological conditions (excavated face conditions), visual observations by striking, crack interval, stability of excavation face and monitoring of ground deformation. It is important in NATM to feed back the information obtained from construction stage to the designer. The design is accordingly reviewed i.e. shotcrete thickness and rock bolt length is revised as per geological conditions and stress measurements. NATM is a widely used method that avoids an extensive supporting system by making effective use of the inherent ground strength and the strengthening of existing ground by shotcrete and rock bolts. There are various excavation methods in NATM according to the split of excavation face such as full face, bench cut, top/bottom drift and sectional excavation. The short bench cut method (standard bench length of 30m between the upper half and lower half excavation) was determined to be suitable for Kohat tunnel due to weak nature of rocks(mostly C and D type). Crown heading with fore poling (Fig. 3) was adopted at Kohat tunnel due to soft geological conditions.
Figure 3. Tunneling methodology adopted at Kohat Tunnel
The arrangement of equipment during construction operation in limited space is depicted in Fig. 4.
Figure 4. Arrangement of equipment during construction operation
It was determined that CI type rock requires support only in the form of rock bolts and shotcrete. Generally 3m deep rockbolts with c/c spacing of 1 to 2m have been used in this type of rock. Supports provided under CII are almost the same as those for CI. In few weak zones depth of rockbolts has been increased to 4m, c/c spacing has been reduced to 1m and thickness of shotcrete has been increased. The supports provided for DI type rock are rockbolts, shotcrete, steel ribs at 1.2 m spacing, wire mesh and fore poles. The thickness of shotcrete is 10 cm for CI and CII whereas it is 15 cm for DI type rock. Due to extensive support system the tunneling cost per meter run for DI type rock is much higher than the corresponding cost for CI or CII type of rocks. Presence of DI in excess of its predicted quantity, therefore, could increase the project cost tremendously. Such an event might have proven to be disastrous for the project due to increase in cost alone.
Based on the Japanese standard a 35 cm thick concrete lining has been used for the portals, length of South portal being 41.5 m and that of North 21 m. The concrete lining is structurally reinforced to resist dynamic forces during an earthquake event. A plain concrete lining of 30 cm thickness has been provided for the main length of tunnel. A water proofing sheet, 1.5 mm thick, has been installed under the final concrete lining all along the tunnel regardless of ground water. The water proofing sheet arrests seepage and the drainage system properly disposes it off. The water proofing sheet also isolates shotcrete and the concrete lining, consequently preventing shrinkage cracks. Based on efficient and successful construction it was concluded that New Austrian Tunneling Method can be successfully used in difficult geologic settings comprising of shale, weak sandstone and fractured lime stone. Various construction phases are depicted below.
Photograph 1. H - Ribs installation
Photograph 2. Drilling by wheel jumbo
Photograph 3. Shotcreting by shotcrete robot
Photograph 4. Waterproofing sheet
Photograph 5. Concrete lining
Photograph 6. Finished concrete lining and concrete lining
in progress with shotcrete robot in far background
ANALYSIS OF PREDICTED AND ACTUAL GEOLOGY
The actual geology of tunnel reported here-in pertains to 1500 m length of tunnel only. The contractor prepared daily reports and a standard Performa was filled by the site geologists for every 1 m length of excavation. This Performa included information such as length of excavation per day, rock classification, major types of rocks encountered, amount of water ingress, cross section of excavated face observed by geologist, recommended rock support pattern and geologic details such as jointing pattern and other anomalies. This record was sifted out and percentage of each type of rock actually encountered for every one meter length of tunnel was determined. The results are reported in Table 3 and Fig. 5.
Table 3. Actual lengths of CI, CII, and
DI type rocks encountered, in 100 m segments
 |
Figure 5. Actual length of CI, CII & DI type rocks encountered in 100 m segments
The predicted percentage of CI, CII and DI (on the basis of geophysical exploration) for every 100m length are reported in Table 3 and Fig. 6 below.
Table 4. Predicted lengths of CI, CII, and
DI type rocks in 100 m segments
 |
Figure 6. Predicted length of CI, CII & DI type rocks (100 m segments)
The deviation from predicted length of each rock type is presented in Fig. 7.
Figure 7. Comparison of predicted and actual rock variation (for every 100 m segment)
It can be observed that for the initial 400 m length the deviation from prediction was significant. DI type rock was substantially less than predicted. Similarly the actual rock composition is much different from predicted geology for the reaches 600 to 700 m, 1000 to 1100m and 1400 to 1500m. A statistical analysis of variation was performed (Fig. 8) and the coefficient of regression is determined to be very small.
Figure 8. Statistical relationship between predicted and actual length of various rocks
It is, therefore, concluded that the correlation between the predicted and the observed geology is extremely poor. It may be mentioned here that the actual length of CI type rock is far in excess of the predicted length, 72% against predicted 41%, and DI type is only 17% against 39% predicted length. Since CI type of rock required less intricate rock support system, the error resulted in considerable cost savings. The cost of project, however, would have ballooned up disproportionately had the length of DI type rock exceeded the predicted length.
CONCLUSIONS
The following conclusions were drawn from this study:
REFERENCES
 | |
| © 2005 ejge | |