Add us to your taskbar by dragging this icon RGI logo to the bottom of your screen.

Share |

Balancing the risks in København

01 Jan 2003

Bids were received last month for construction of Phase 3 of the København Metro. Andrew Grantham finds that the tendering process has drawn on the experience of the many different parties involved in the first two phases of the project

TENDERING closed on December 3 for the contracts to build the third phase of the København Metro. This will see the network extended south from Lergravsparken to the city's airport. Nine bidders submitted tenders for the two negotiated design-and-build contracts which cover the civil works and the electrical and mechanical systems. The civil works invitation includes an upper price limit; any bids above this will be disregarded.

Prequalified bidders submitting tenders to promoter Østamagerbaneselskabet for the civil works are Skanska Danmark, Hoffmann and Arkil Novejfa, NCC-SRS, MT-Højgaard and a joint venture of Pihl with Aarsleff. The E&M systems bidders are Ansaldo Trasporti, MT Højgaard, the Aarsleff-Wicotec-Semco joint venture and Intabb.

The tendering procedures for Phase 3 make use of valuable experience gained with the first two phases. Following the inauguration of Phase 1 on October 19 (RG 12.02 p730), a seminar was held in the Danish capital on November 21 - 22 to review the experience of all the parties involved in its construction.

Legal framework

In June 1992 the Danish parliament passed the Ørestad Act for the creation of a new town south of the city centre. This is being promoted by Ørestadsselskabet (Ørestad Development Corp), which is owned by the Municipality of København (55%) and the Danish state (45%). The land was given to ØSK free of charge.

The provision of high quality public transport links with central København was integral to the scheme. Financing for Ørestad would come from land sales, land taxes and public transport revenues. But the act did not stipulate the form that the specified transport routes should take. ØSK was to adopt the most appropriate system, not simply the cheapest. Financing for each phase would be reviewed once the previous stage was complete.

In 1994-95 ØSK commissioned studies of different transport options: a modern tramway, with on-street running in the city centre but segregated at-grade right-of-way elsewhere; a light rail line running underground in the city centre; and a fully segregated driverless light metro.


There are many ways to tender a large, complex project such as a metro, and ØSK spent the second half of 1994 determining the best method of implementation. There were reviews of contracting concepts, the number of contracts required and tendering procedures. Critical issues were compatibility between designs, management of interfaces between companies and contracts, timing, financial size, division of risk, and keeping the number of contracts to a minimum. Combining the tunnelling works that formed a major part of phases 1 and 2 offered cost savings, and the act was amended to permit this.

A decision had to be made as to whether to adopt a design-and-build model, or to issue separate design and construction contracts. Then there was the choice between a single contract, or separate civil and systems contracts. A design-and-build contract would give the contractors a high degree of autonomy, reducing ØSK's ability to influence the final outcome. It was feared that some bids might be based on sub-standard designs, or radical concepts that would make comparison difficult.

ØSK wanted to minimise its exposure to contractors who might be tempted to submit low bids and try to recover losses through claims. Separate design and construction contracts would provide for close supervision and direct participation, at the expense of being more time-consuming.

To ensure bids were comparable, aesthetic, functional, maintenance and technical standards had to be detailed in the tender documents. ØSK would be exposed to interface management risks, but the risk of a failure by the contractors would be reduced. As a result of these considerations, a design-and-build model was chosen for Phase 1.

Single contract

ØSK considered a single construction contract the most attractive, as interfaces would be minimised, and supervision simplified. However, only very large joint ventures or consortia would be able to bid. And because of the size of the contract there was a risk that a failure on the part of the contractor could have disastrous consequences for the whole project.

Railway systems contractors are accustomed to negotiation throughout the contract period, giving them the freedom to provide answers to a client's needs. In contrast, such negotiations are less common among civil engineering contractors, who generally work from detailed designs provided by their clients.

Whilst separate design and construct contracts would be feasible for the civil works, this would not be realistic for the railway systems. The civil works would account for over 70% of the project value, and there was concern that the railway systems aspect could suffer as a result. So ØSK decided to award two contracts.


In October 1994 ØSK opted for a light metro, which was seen as offering the lowest overall cost for the number of passengers it could carry, even though the initial costs were much higher than for light rail.

A metro's ability to attract custom was a deciding factor. Automation is suited to handling irregular passenger flows, as trains can be diverted to suit a changing situation. Over 600 routes are built into the control system. And the presence of a steward, rather than a driver, on the trains meant ticket gates would be unnecessary.

In March 1995 contractors were invited to prequalify, in accordance with tendering rules laid down in EU Supplies Directive 93/38. Public works tenders in Denmark are subject to special rules, but it was established that these did not apply to design-and-build projects.

A negotiated procedure with prior advertising was adopted, as other EU tendering models prohibited negotiation with bidders, and this was felt to be essential. But because the operations were to be sub-let to a third party, it was unclear if the tendering approach would be permitted. The directive applied to 'an operating system', and the metro did not yet exist. The situation was referred to the European Commission, which decided that this approach would be allowed.

Bidders were asked to prequalify for the civil works (Contract A) and for the transport systems, operation and maintenance (Contract B). They were also invited to bid for a combined contract, including O&M for Phases 2 and 3 (Contract C). Operations and maintenance for the first five years were included, as it was believed that the contractor would be best placed to assess how the systems were functioning.

Lump-sum contracts with provision for inflation were requested, in line with other tenders at the time. Approaches such as partnering were not considered, but according to Poul Pr??sius of COWI, if the contracts were being awarded today a different approach might have been taken.

In July 1995 six firms prequalified for Contract A, five for B and four for C. Tender documentation included details of ØSK's aspirations. These included 100sec headways, high levels of service and information, and good interchange with existing transport modes. Architecture and design was to be simple, unpretentious and in harmony with the environment.

The documents detailed the scope of works, conditions of the contracts, technical specifications, and the basis for pricing and payment.

For legal reasons Denmark's AST93 general and special conditions were written in Danish. All other documentation was issued in English, which was chosen as the main project language to reflect the international nature of the tendering. ØSK Managing Director Anne-Grete Foss commented that the language and cultural differences made the project seem 'like the Tower of Babel.' She said that the Danes 'love plans and diagrams, but the Italians prefer making things work, even if no-one knows how they work.'

Bidders were able to adopt the illustrative designs, or to develop other proposals, as long as they complied with the tender documentation. In addition they were able to submit alternatives. At meetings in September and October 1995 the bidders, could question the implementation plans, and suggest amendments.

Tenders closed in February 1996. A four-stage evaluation was carried out by geographically separate consultants. As Contract C was a combined bid for A and B, this evaluation was handled by the project managers.

Factors considered included technical quality and design, construction, manufacturing and installation methods, procedures and approvals, organisation, price, employer's risk and payment plans. Financial evaluation covered tender sums, upgrading cost and potential consequential liabilities. Contract B criteria included life-cycle costs, commissioning plans and the use of proven technology. The O&M aspects included training and education, technical quality and reliability, quality assurance and environmental impact.

The first phase of review identified major areas of non-compliance and extreme pricing. Detailed discussions continued with adjusted bids in the second phase. A third phase was intended to produce final bids, but some quantitative changes were made to the requirements before the final bids were submitted in July 1996.

Separate awards

The most economically advantageous bid for each contract was selected, and it was found that the single contract C was less attractive than separating contracts A and B.

In October 1996 ØSK awarded the DKr2·9bn Contract A to the multi-national København Metro Construction Group (Comet), and the DKr1·7bn Contract B to Ansaldo. This included construction of the operations and maintenance depot, which was the only civils work in Contract B.

At 15 months the whole process, from issue of tender documents to signing, lasted somewhat longer than expected, and debriefing with the unsuccessful bidders found that most felt the tendering costs were excessive.

A further 70 smaller contracts totalling DKr800m covered station and shaft areas, utility relocation and other works. To ensure aesthetic and technical compatibility, the design-and-build approach was dropped in favour of design by ØSK's own consultants. This allowed the time-consuming negotiated procedure to be abandoned in favour of mathematical models to assess the best bid.

A feature of the contracts was that works could not be handed over until approved by the Danish rail safety authority Jernbanetilsynet. This meant the contractor carried the risk after completion. Comet Project Director Peter Jefferies told the seminar delegates that only the customers are reasonably equipped to carry the safety assessment risk. Contractors are unable to understand requirements at the tender stage, and he suggested that 'it will all end in tears' unless contractors reject attempts to shift such risks onto them. He felt that 'a witch doctor with chicken bones would have more chance of estimating risk' than a civil works contractor.

Jefferies felt clients should not have too high hopes of architectural excellence from a competitively-tendered design-and-build bid. While contractors can meet basic design requirements, aesthetic requirements are difficult to satisfy. The risks devolved to the contractor are minimal, but where 'subjective client satisfaction' is essential, he feels the client should retain control of the detailed architectural design.

The project was carried out under close media scrutiny and amid great public interest, which could be disconcerting for managers not used to what Jefferies described as the 'advanced, even exaggerated, sense of democracy' in Denmark, which 'makes decision-making very slow.' Jefferies said he would have preferred to negotiate behind closed doors.

Tonni Christiansen, Deputy Director of Rambøll said strict interface management was a key issue. It was essential to determine the appropriate levels of specification. The owner should focus on the performance of the final system, not the technology required to produce it, and should avoid creating unnecessary constraints which could favour one bidder.

He agreed that the clients should retain responsibility for interfacing with any parties 'interfering' with design requirements. Much effort had been put into creating clear and detailed documentation of interfaces between the two main contracts, but many discussions were still needed. 'Although a certain issue may not be one's responsibility, it can still turn out to be one's headache,' he warned.


Project chronology

Ørestad ActJune 1992
Consultancy contracts 1994
Contracts signed Spring 1994
Invitations to prequalify March 1995
Invitations to tender July 1995
Discussions with interested parties September-October 1995
Bids received February 1996
Best and final offers July 1996
Contracts signed October 3 1996
Opening of Phase 1 October 19 2002
Tenders received for Phase 3 December 3 2002

Pre-contract consultants

Project management COWI

Civil works COWI

Rolling stock Carl Bro

Railway technical Rambøll, NIRAS installations Halcrow

Traffic Carl Bro

Architecture & town planning KHR AS

Environmental issues Carl Bro/Environmental Resources Management joint venture

Over-specification adds cost

One problem identified during construction of the Metro was the over-specification of the toughness required from steel components. Normal Danish minimum design temperatures are -10íC, but the specification required the stations to use material suitable for -40íC.

Such specialised steel is expensive and difficult to procure. Comet Engineering Manager Edward Garvey explained how a contract change request was made, but not before the above-ground stations had been built to 'more or less withstand the next Ice Age.'

A 230-page book in English, with numerous black and white illustrations, contains all the papers presented at the Metro Inauguration Seminar on November 21 and 22. DKr250 from COWI A/S, Parallelvej 2, DK-2800 Kongens Lyngby, København,

  • CAPTION: A total of 35 three-car trainsets is being delivered for Phases 1 to 3. Services on Phase 1 will require 10 sets to maintain 3 min headways. According to Karsten Fick of Carl Bro, the design parameters required 'room for all, not seats for all', as the mean journey is expected to be 4 km or 6 min
  • CAPTION: Except for stairs providing access to the trackbed from the platform ends, the stations are structurally independent of the viaducts
  • CAPTION: Station entrance totems display departure information. Stations have limited facilities, as they 'are not urban waiting rooms, just a short stop on the way to a destination', says Operations Director Carsten Fich