Selecting a contract type that provides suitable risk allocation may be crucial for the success of any subsurface project. What a suitable risk allocation is depends on many factors i.e. the complexity and knowledge of the geology, the owner’s experience, the method of financing, and even the public’s involvement.
Subsea tunnels will normally have, or are perceived to have, a higher risk profile than other tunnels. Thus, it becomes especially important to identify the actual risks, and to find out how these can be allocated to either the owner or the contractor, or whether the parties should share the risks in a defined manner.
The experiences from the construction of the Hvalfjördur tunnel in Iceland and the Vága tunnel in the Faroe Islands, both excavated by drill and blast, are discussed. These two examples of ‘adjustable fixed price’ contracts illustrate the principles of risk allocation.
Subsea risks
Subsea tunnel risks revolve around the limited knowledge of ground conditions due to practical difficulties and higher costs of undertaking site investigations prior to construction. Often poor rock in fault zones typically occurs on the deepest points with the least rock overburden, whilst ever present is the inherent hazard of tunnelling below the sea.
Several of the problems apply also to tunnelling on land, but there the risks may be smaller as the consequences if something goes wrong may be less. The lessons learned may still apply.
In Norway, about 40 subsea tunnels in rock have been built during the last 25 years. They have all been excavated by drill and blast. The risks during construction are reduced by extensive use of probe drilling and pre-grouting. The dominating contract type has been the typical ‘unit price’ (or unit rate) contract developed during decades of construction of underground hydropower projects, especially in the 1960’s through the 1980’s. They involved the application of tendered unit prices (or unit rates) and, notably, regulation mechanisms for completion time according to performed quantities, by using pre-set ‘standard capacities’. This keeps the bulk of the risk for the ground conditions with the owner. Accordingly, the responsibility for the site investigations naturally rests with the owner, often being a public body.
Different conditions
When the construction of subsea tunnels spread from Norway to two neighbouring Nordic countries i.e. the Hvalfjördur tunnel in Iceland and the Vága tunnel in the Faroe Islands, the need to adapt the contracts to suit the specific circumstances arose. This included considerations for different geological conditions, other types of owners and other financing methods. In both countries, the respective tunnel was the first of its kind.
The site investigations prior to construction followed Norwegian practice with respect to methods. Due to the different geological conditions (young basalt formations rather than Precambrian gneisses), the extent of seismics performed was significantly higher. For the Hvalfjördur tunnel the refraction seismic profiles were used in a wide grid to confirm a suitable corridor, due to the presence of many faults and areas of thick deposits above the bedrock. For the Vága tunnel, the reflection seismics showed an even sea floor with thin sediments, allowing the refraction seismics to be more concentrated along the alignment.
Hvalfjördur tunnel
The 5.8km long (3.6km subsea), 55-65m² cross section Hvalfjördur tunnel is 30km north of Iceland’s capital Reykjavik. It improves access to an important industrial area and cuts the travel time on the west coast stem road significantly. A private enterprise, Spölur pf, was granted concession to design, build and operate the project based on toll revenues(1,2). The contractor Fossvirki h/f was established as a JV between Skanska (Sweden), Pihl & Søn (Denmark) and Istak (Iceland).
As the project was the first of its kind in Iceland, there was a rather strong public opposition. Besides the normal initial scepticism to subsea tunnels, the negative opinion was partly due to experience on another road tunnel (on land) which encountered water inflows of 3m³/sec during construction. The fact that the area is seismically active did not ease the public’s fear of potential collapse and flooding.
The bedrock is basaltic lava, with gently dipping beds 2m -10m thick, making mixed face situations common. The lava flows are inter-bedded with scoria and sediments, and are cut by numerous faults and dikes.
Spölur needed to provide a high degree of predictability to the cost. At the same time it was necessary to allocate the risks in a balanced manner to avoid unnecessarily high contingencies being included in the bids, as can happen with fully fixed price contract. It appeared that an ‘adjustable fixed price’ contract would work. Spölur’s advisers prepared the conceptual design, including the tunnel alignment, and technical specifications based on functional requirements as far as possible, also for the technical installations.
The contract committed the contractor to perform detailed design; additional site investigation, as necessary; construct the tunnel within 39 months; commission installations and operate the tunnel two months before take over; transfer the tunnel to the owner, with a guarantee period of 40 year.
The contract had a fixed price portion of 85%, the remainder allocated to rock support and grouting. If the actual quantity of rock support and grouting measures were within +/- 10% of the BoQ (Bill of Quantities), no regulation would occur. If higher quantities occurred, lower unit prices would apply and vice versa. Rock support was based on pre-defined rock support classes, supported by the Q-system, and the ground treatment on a pre-defined programme of continuous probe drilling and pre-grouting according to pre-set trigger criteria.
The contract had a ‘cash cap’ of +/-5% of the fixed price. If this was exceeded, loan notes would be issued to the contractor with a lower priority than the long-term loans, which would have brought the contractor to partial ownership. The contractor financed the project during construction, followed by the long-term lenders take over. A tunnelling ‘board of experts’ was established with representatives from both the short- and long-term lenders which followed the design and construction to ensure a safe and sound process.
The aim of the remuneration model was to establish a clear incentive for the contractor to utilise efficient, but safe techniques for the tunnelling. A bonus for early completion was included. This necessitated a close follow-up from the owner’s side, including specific requirements to the documentation the contractor had to provide before, during and after construction. Thorough mapping of geological conditions and follow-up of probe drilling and pre-grouting was conducted.
The encountered geological conditions were better than expected. The large number of faults and dikes were mostly at a favourable angle to the tunnel. Rock bolting and sprayed concrete were sufficient for rock support. The pre-grouting amounted to 35% of expected quantities. Only 13% of the tunnel needed pre-grouting, and the remaining water seepage was 20% of the requirement (see Table 1). One of the portals had to be elongated some 200m by concrete cut and cover due to unforeseen depth of soil deposits, not detected by the ordinary percussive rock head contour drilling from the surface.
The construction started in May 1996 and the tunnel was opened for public traffic in August 1998, eight months ahead of schedule, earning the contractor a significant bonus, and the owner the advantages of a successful project. The costs ended on ~US$49M, ~5% above budget, mainly due to the elongation of the concrete portal. The early completion was very beneficial also to the owner, as it allowed an earlier start on toll collection and reduced the financial costs. The public scepticism was turned into confidence and the traffic volume reached twice the expected 20 years’ AADT of 1700, in the fifth year after opening AADT is already 3540. The ‘adjustable fixed price’ contract with various types of incentives worked according to the intentions and satisfaction of both parties.
Vága tunnel
In the Faroe Islands the 4.9km long (2.5km subsea), 65m² cross section Vága tunnel below the Vestmanna strait connects the two main islands, providing a fixed link between the capital and the main airport. Although this is the first subsea tunnel on the islands, Faroe Islands can boast a very high rate of tunnels per capita with more than 60km of tunnels built for the 50,000 inhabitants. A private enterprise, Vágatunnilin pf, was established and was granted the concession to design, build and operate the tunnel. The project is financed by private loans and government contributions. The invested capital will be repaid by toll revenues. The Scandinavian contractor NCC was contracted for the construction, with Byggitek as local partner.
The bedrock consists of thin basaltic lava flows, with typical thickness of a few decimetres, ranging up to 1m-2m. The Mid-Basalt series is known from other projects as generally competent rocks. Abundant vesicles or pores, sometimes with up to 20% pore volume, are frequently concentrated on the top of each flow. Stability problems and water inflows are mainly connected to fault zones and dikes, and to occasional volcanic ‘chimneys’. A number of weakness zones and some tuff horizons were expected(3).
The tender was based on a standard unit price contract, and included incentives for the contractor intended to encourage efficient, safe tunnelling. There was an early completion bonus, with a penalty for late completion, a sharing of cost savings on a 50/50 basis for alternative solutions, and compensation of part of ‘lost profit’ in case of reduced quantities of rock support and treatment. Increased quantities would be remunerated by reduced unit prices.
The rock support classes would follow the recommendations of the Q-system after adaptation to the basaltic rock mass conditions. The grouting, to be performed basically as pre-grouting, should ensure a remaining water ingress of less than 300lit/min/km.
During negotiations, certain items were agreed to be transformed to fixed price elements. This included mobilisation/demobilisation, rock excavation, rock support and rock mass grouting, whilst all installations including frost insulated shielding against water dripping were kept as unit price items. By this, the fixed price portion of the contract constituted ~70% of the contract. This transferred the risk of variations in quantity of rock support and grouting to the contractor, however, the fixed price included only measures specified in the BoQ. The need for support measures other than those specified in the BoQ would be at the risk of the owner. An agreement entitled the contractor to a bonus if the remaining water inflow was reduced to less than 150lit/min/km, with an interpolation between 150lit/min/km and 100lit/min/km.
The conditions were as expected. Few weakness zones were encountered, and the tuff horizons did not cause significant stability problems. Rock support was performed by rock bolting and sprayed concrete only, with some more rock bolts, but less sprayed concrete than in BoQ. Except for some very permeable zones, that required extensive pre-grouting, water was not a problem. However, the water drip shielding increased due to numerous areas with small seepages. Grouting quantities ended somewhat above expectations. The remaining water inflow is below 100lit/min/km, i.e. 1/3 of the project requirement.
The contractor’s QA system was monitored closely by day-to-day follow-up. In fixed price contracts it can be important to ensure that the contractor is not becoming ‘over-confident’ compromising the overall safety. A ‘panel of experts’ was established, consisting of three experienced tunnelling engineers, which had inspections and meetings on site every three to four months. The panel monitored the progress and procedures and gave recommendations to the parties, being instrumental in solving conflicts at site. The owner and the contractor were both contractually obliged to follow the recommendations of the ‘panel of experts’. This worked according to expectations.
The project was opened in December 2002, approx. 28 months after construction started in September 2000 and six months ahead of schedule. Three months after opening, the traffic volume is higher than expected at AADT=800. The expected AADT after 20 years is 2000. The costs after extra works ended on ~US$32M (according to today’s exchange rate) after an increase by ~13%; approximately half of which was due to the increased water drip shielding.
The contract was originally a unit price contract, but was converted to an ‘adjustable fixed price’ contract. It worked according to intentions and to the satisfaction of both parties, and contributed to the success of the project. The contractual aspects have been dealt with without any need for litigation.
Conclusions
It is possible to customise a contract to the specifics of a project, considering cost predictability for the owner, while keeping a reasonable risk on his part for the more uncertain elements like rock support and water control.
This can be achieved by applying elements from unit price contracts to account for variations in the geological conditions, which may help avoid that unnecessary high contingencies are built into the bids. Unless the completion time is also adjustable, it must not be set too tight. Such ‘adjustable fixed price’ contracts may include incentives for the contractor to use innovative methods to improve efficiency without sacrificing safety. Such incentives may also be built into unit price contracts.
The potential benefits of ‘adjustable fixed price’ contracts were identified at the planning stage, and helped the realisation of both projects. As the encountered conditions were more or less as expected, the contracts were not put into test in unexpected or unforeseen conditions.
Related Files
Longitudinal section of the Hvalfjördur subsea tunnel in Iceland
