Contractor Intekar Yapi Turizm Elektrik Insaat San. ve Tic. Ltd. Sti chose a Robbins Single Shield TBM, 8 m in diameter (Fig. 1), to excavate one section of tunnel. Mixed ground conditions prevail on the project, and range from abrasive, interbedded sandstone and mudstone with quartzite veins to highly weathered shale and dolomitic limestone. The TBM has thus far encountered some of the hardest rock ever tunneled in Turkey, measuring between 136 and 327 MPa UCS.
Location and Geological Studies
The 17 km long double-track railway project for owner Turkish State Railways (TCDD) will include the longest railway tunnels in Turkey once complete (10.1 km total length). The railway will bridge a gap of one of the busiest railway sections between Toprakkale and Malatya, where trains can only proceed with the support of an additional locomotive. With the new project, totalling 65 million euros, the route between Bahçe and Nurdağı will be 15 km shorter, with gentler gradients, wider curves (minimum radius will increase from 500 m to 1500 m) and proceed at faster speeds (from 40 km/h to 120 km/h). It is planned to complete the tunnelling portion of the project by the end of 2019 and to open the new section of line to traffic by 2023.
The excavation of the tunnel was planned to start from chainage 13+450 m and to terminate at chainage 3+700 m. The chainage from 13+450 m to 12+400 consists of Karadag Limestone of Mesozoic age, which is affected by the East Anatolian Fault (EAF), fracturing the rock formation to a great extent. High water ingress was expected in this area. Karadag Limestone discharges the water at the foot of the mountain on the Nurdagi side. Several springs are present along the EAF, which is a large fault zone about 550 km long, extending from the Gulf of Iskenderun to the North of Anatolia.
This fault takes up most of the motion between the Turkish and Arabian plates. Due to technical difficulties and time necessary to provide the TBM, the first 1050 m in limestone was planned to be excavated by the New Austrian Tunnelling Method (NATM)  since a risk analysis showed that this area was risky to very risky for TBM excavation. However, once the tunnel construction was underway, only 850 m of this tunnel could be excavated by NATM.
The geological formation from chainage 12+400 m to 4+850 m consists of middle-Ordovician-aged Kızlaç Formation of very massive interbedded meta-sandstone, meta-quartzite and meta-mudstone with very high strength and abrasivity characteristics. This section is being excavated with a single shield hard rock TBM. However, the massive characteristic of the rock formation changes from chainage 4+850 m to 3+700 m, being affected by local faults and shear zones; in this zone RQD (Rock Quality Designation) values are very low and high water ingress is also expected. This section is planned again to be excavated with NATM.
Laboratory Testing of Rock Characteristics
The strength characteristics and the ratio of compressive strength to tensile strength play an important role in determining the optimum penetration of disc cutters, and thus, net penetration rates of TBMs.
A series of Schmidt Hammer tests (N-type) were carried out in the field during a site visit to the Bahce–Nurdagi tunnel area. The mean values obtained in the field for meta-sandstone are 56.4 ± 5.6, for meta-mudstone 34.8 ± 9.7 and for limestone is 56.1 ± 4.8. See Tables 1 and 2 for UCS (Uniaxial Compressive Strength) and UTS (Uniaxial Tensile Strength) values.
Basic parameters affecting tool consumption of a TBM are the abrasivity, strength and geological discontinuities of the formation to be excavated. After petrographic analysis carried out on the rock samples, the meta-sandstone and the meta-siltstone have been classified as very abrasive with quartz content varying from 48 to 68 %.
The sandstone is composed mainly of quartz with minor feldspar, opaque and rock fragments. The grains are cemented by chlorite. A thin chlorite cement holds the grains together. The sandstone can be classified as quartz arenite. The mean quartz content is 68 % with grain sizes varying between 0 and 0.3 mm.
The mudstone (siltstone) consists of angular quartz grains embedded in a voluminous matrix of sericitic muscovite, opaque, chlorite, feldspar and quartz. It shows a distinct cleavage defined by the parallel orientation of the mica grains. The mean quartz content is around 48 % with grain sizes between 0.05 and 0.1 mm.
Rock Cutting Experiments
Two block samples of meta-sandstone and meta-mudstone were subjected to full-scale laboratory rock cutting experiments . A constant cross-section disc cutter with a tip width of 1.2 cm and ring diameter of 13 inches (330 mm) was used in the experiments (Fig. 2). A testing program with a constant cutter (line) spacing of 80 mm including relieved (adjacent cutters) and unrelieved (single cutter) cutting patterns was set, and depth of cut was varied (3, 5, 7 mm). The sample surface was cut several times to condition the surface similar to the real case of a tunnel face. Normal, rolling and side forces were recorded, muck samples were collected and their weights were measured during each cut. Force values were reduced by a custom-made macro program. The meta-mudstone sample was cut parallel to the bedding planes. Based on the experimental results, optimum line spacing to penetration ratio was determined for the TBM to be used in the tunnel. As a basic rule of rock cutting mechanics, specific energy – defined as the energy consumed per unit volume of the excavated rock – is optimum for a given s/d (cutter spacing/cutting depth) ratio. In optimum conditions, the energy spent to excavate a unit volume of rock is minimal. Cutter spacing is a constant value of a TBM cutterhead, which dictates that for a given rock formation the predetermined cutting depth in the laboratory will determine the optimum thrust values of the excavating machine. In the light of these main rules, an intensive laboratory full-scale cutting test program was planned to obtain the relationships between cutting depth and cutter force values, specific energy and s/d ratio for two rock samples. The cutting test results are discussed in  and summarized in Table 3.
Machine Design & Assembly
The design of the Bahce–Nurdag Single Shield TBM was optimized for variable geology and for the results obtained by laboratory testing. A high-speed segment erector and hoist with mechanical pickup were designed to build segment rings of 350 mm thickness in a 5+1 arrangement. The system allowed for pea gravel injection and grouting through the segments for backfill. 360 degree probe drill coverage was provided to allow for systematic probing of ground conditions and grouting if necessary.
In the case of a large inrush of water or mud, the uniquely designed Single Shield TBM can engage an emergency sealing system consisting of muck chute closure doors (Fig. 3). Since the conveyor is a belt conveyor and is not enclosed like a screw conveyor, it must be sealed off at the front. The bulkhead has a large sealing gate just above the belt conveyor. These are pressure relieving gates. These gates can also be used in a semi-EPB mode: As the pressure builds in the cutting chamber, the gate is opened by the pressure, and material spills onto the belt. As the pressure is relieved, the gates then automatically close, again sealing off the chamber. In extreme cases, the gates can be sealed and the probe/grout drills can be used to forward drill and grout for ground consolidation and to seal off the water.
Assembly on Location
Onsite First Time Assembly (OFTA, Fig. 4) was deemed the most efficient method of TBM assembly due to the remoteness of the site and proximity to a conflict area (the jobsite is about 48 km from the Syrian border). OFTA is a method developed by Robbins that allows for the TBM to be initially assembled at the jobsite rather than in a manufacturing facility, saving the contractor up to three months on the delivery schedule and millions in US-dollars.
While several villages were nearby and roadways gave good access, the logistics of shipping internationally was challenging. Crews often had to wait for small items such as cables, lights, and hydraulic fittings that could not be sourced in Turkey. Customs clearances were difficult to obtain and resulted in delays of international shipments. Despite this, the machine was completed in early 2016, and following a ceremony in January, was walked forward to the entrance of a 500 m long starter chamber. Once the tail shield was flush with the portal entrance, an invert thrust frame was installed that allowed the machine to build full starter rings, complete with pea gravel and grout backfill, up to the launch face.
Challenging Site Conditions
The TBM advanced successful in very hard and abrasive rock formations after chainage 11+400 m (Bild 5), with mean daily advance rates changing from 10 to 16 m. Typical values in hard interbedded meta sandstone and meta mudstone were 11.8 m per day. As of May 2018, the TBM had just 3000 m left to bore, and had achieved a best month of 456 m. Cutter ring consumption varies widely depending on the quartzite content of the rock: anywhere from 4.83 m to 129.75 m/ring. The TBM has encountered metaquarzite seams containing 80 to 90 % quartzite, which account for the lower numbers. These seams are expected in the remaining tunnel as well. In May 2018, the TBM was boring in mudstone and sandstone with a lower 30 to 40 % quartzite content.
Difficult to extreme geology, limited access to the project area and challenging working conditions on site make the Bahce–Nurdag Railway Tunnel Project a remarkable and special one in the industry. An intelligent combination of two tunnelling methods, excavation by NATM and fully mechanized tunnelling with a Hard Rock Single Shield TBM, was selected to provide for the most efficient and safe tunnel excavation. For the sake of optimal TBM design, intense studies of the rock properties were taken in advance, and this has proven the correct approach. Well proven features were added to the design of the TBM for crossing difficult and varying ground formations. Delivery to site did not allow shipment of big bulk items, and onsite first assembly was successfully employed to overcome this given restriction. The machine’s performance in the extreme and difficult ground formations right after start and where TBM employment was qualified as risky to very risky demonstrated the approach taken was the right one. TBM excavation was worth the efforts made in special design and detailed studies of the rock formations. This performance is proof of modern TBM design and machine performance in ground formations where in the past employment of mechanized excavation methods was not possible or would not be practical over NATM. The industry is well on its way to design and deliver TBM for almost all ground formations – limitations are steadily decreasing and very soon will have disappeared completely.
 Bilgin, N., Copur, H., Balci, C. 2016. TBM excavation in difficult ground conditions, case studies from Turkey, Berlin: Ernst & Sohn. 350 p.
 Bilgin, N. (2016). An appraisal of TBM performances in Turkey in difficult ground conditions and some recommendations, Tunnelling and Underground Space Technology 57, 265–276.
 Bilgin, N., Copur, H., Balci, C., Tumac, D. (2017). TBM Performance prediction using laboratory cutting tests in very hard and abrasive rock formations, International Conference on Tunnel Boring Machines in Difficult Grounds (TBM DiGs), Wuhan, 16–18 November 2017