Shani Wallis, TunnelTalk

Fixed-price, all-risk contracting was taken to a new level for procurement of the 120MW Esti hydropower project in Panama. TunnelTalk visited the project in 2003 and filed this site visit report.

In the early 2000s in Panama, the overseas division of Sweden's heavy construction firm Skanska, was civil contractor and sponsor of an all-risk, fixed-price contract to build the 120MW US$213 million Esti hydropower facility.

Up-stream face of earth-filled concrete-faced Barrigon Dam

Owner of the project, private power generating company AES Panama, a subsidiary of international parent company AES of Arlington, Virginia, in the United States, acquired plans for the plant in 1998/99 when it bought a 30-year concession to own and operate three hydro plants from the government of Panama.

The purchase was concluded as part of the Government's privatisation of the IRHE state-owned electricity generating industry, to meet the International Monetary Fund's criteria for reducing the country's national debt. Within the sale agreement, the Government insisted that, along with the three operating plants, AES buys and develops the much-needed capacity of the proposed Esti project.

EPC arrangement
At that time, Esti was being developed by IRHE and under a US$199 million contract for the civil works awarded to ICA of Mexico. IRHE and the Government however could not finance the project. With privatization, AES accepted responsibility to fund and construct Esti and also accepted a very tight deadline to have the first of two units on-line by the original ICA contract end date of 15 November 2003. Under AES ownership, the ICA proposal was still not financially viable. The contract was terminated and the project redesigned and rebid in 2000, but starting from scratch was a major risk. Under the privatisation agreement, any delay to operation of the plant's first unit by 15 November 2003 would incur a penalty of more than US$150,000/day. To mitigate these risks, AES adopted the design-build type of procurement and known in the power industry as an EPC, Engineer-Procure-Construct, contract.

Adit 1 at the upper half of the 4.8km long headrace tunnel

"We adopted the EPC concept essentially for time concerns," explained AES Panama chief engineer, Ambrosio Ramos. "With a power on-line deadline of November 2003, and no procurement or construction contracts awarded at the beginning of 2000, time was of the greatest concern. Under the EPC concept, the same group is responsible for the detailed design of the civil, hydraulic and mechanical works, as well as design and procurement of all the associated power generating equipment, and all the construction activities. Only under such a method could we attempt to meet the deadline and avoid the penalties. Other advantages of the concept include the fact that procurement is let on a fixed-price, all-risk basis.

The owner knows exactly how much the project is going to cost at the start of the construction. There can be no hidden extras and no prospect of cost overruns. As such, the owner and the contracting group can arrange their funding and financing needs with greater certainty."

According to Rod Jorgensen, Project Manager at Esti for AES Panama, this is believed to be the first application of the EPC contract for procurement of a hydropower plant involving extensive earth works, dam building and tunnelling. "I have managed EPC contracts in the past, and while it is a major advantage to have all aspects of project procurement under one umbrella of responsibility, the geological and geotechnical aspects of a hydro scheme have introduced very specific issues."

Having adopted the concept against little or no option, only two groups priced the project. Of these the US$213 million fixed-price offer from the Skanska-led group was accepted over a Spanish-based bid. Within the EPC group SwedPower is the designer, Alstom is supplying the generators, GE Hydro is supplying the turbines and other mechanical components, and Skanska itself is managing all the civil works.

In accepting the EPC contract, Skanska took on considerable risk. As well as all geological and geotechnical risk, AES passed on the government's heavy penalty for any delay beyond the 15 November 2003 due date. Set at US$180,000/day, the penalty includes liquidated damages as well as any lost revenue and the cost to AES of buying power on the expensive spot market to meet customer obligations. There would also be the cost to the EPC group for operating on site beyond the due date. These penalties motivated all companies associated with the project to get construction moving as soon as possible and ensure that all aspects of procurement remained strictly on schedule. In fact, while the EPC contract was signed officially in February 2001, Skanska had agreed to start work earlier, in September 2000 in efforts to save time and help meet the power on-line deadline.

Fortunately for the group and despite major design, geological and scheduling problems, Skanska progressed construction efficiently. The last blast in the 4.8km long, 65m² (8m x 8m) headrace tunnel took place in early February 2003, and when TunnelTalk was on site the week after, concrete lining of the tunnel as required had started. Work was also progressing on fitting the first of the two 60MW Francis turbines in the above ground powerhouse and the construction team was confident of meeting the 15 November deadline and claiming a small bonus for early completion.

Fig 1. Plan of the Esti project

Construction dilemmas
The most costly of the various aspects that have taxed Skanska's EPC fixed-price contract was realisation early in the construction phase that the alignment of the headrace tunnel was not going to be suitable for the topographical and geological conditions. Given the area's almost flat, slightly downward sloping topography, passing under a valley at the Rio Esti area, the fall in elevation from the intermediate Barrigon reservoir to the power station was only 122m.

This made a conventional low-pressure almost horizontal headrace, leading to a vertical high-pressure shaft into the penstocks and the turbines impossible.

Instead, the design called for a fully pressurised headrace with a 60m deep x 6m to 8m diameter surge shaft about half-way along. The scheme therefore comprises a 25m high dam on the Chiriqui River that diverts 90m³/sec of flow from its reservoir along 6.2km of concrete lined canal to a head water reservoir created by a 50m high concrete-faced earth-filled dam at Barrigon - that also dams the tailrace discharge from an existing upstream powerhouse - and the 4.8km long high-pressure headrace tunnel that feeds a maximum 118m³/sec to the powerhouse at a velocity of 1.9m/sec and under a maximum 23 bar pressure (Fig 1).

Given the flat topography and the limited overburden, the potential for hydrofracturing in the relatively weak highly weathered volcanic tuff of the area was recognised but it was not until after the urgent need to get excavation started that detailed tests were carried out to assess the exact density of the material. Tunnelling had started from the powerhouse end using extensive geological investigation data gathered from bore holes for the earlier more expensive ICA project, but the project design and the headrace tunnel alignment had been altered significantly, principally by AES and its project engineers MWH, to arrive at a less expensive scheme.

During the summer of 2002 an extra 30 or 40 investigation bore holes were taken by Skanska/SwedPower and it was also possible to carry out in-situ hydrofracturing tests from within the section of tunnel already excavated. These unfortunately confirmed that both stresses and the material density at 2.4Mg/m³ was too low, and the risk of hydrofracturing too high, for the tunnel alignment. The alignment had to go deeper to increase the overburden and to significantly reduce the hydrofracturing risk explained Lars Hässler, Chief Geotechnical Engineer for the Skanska team.

This decision was not taken lightly, since in addition to lost time, the cost consequences were many: some 180m of already excavated 8m x 8m primary supported drill and blast tunnel would have to be abandoned; a 70m deep shaft to connect the deeper alignment to the section of excavated tunnel leading to the turbines in the powerhouse would have to be constructed; 50m of this shaft would have to be steel lined (Fig 2), and the total 300m of steel lining for the powerhouse penstock connection would have to be increased from 4.8m to 5.3m diameter. In addition, future dewatering of the system would require extensive pumping for a large percentage of the unusual bottom up, rather than top down, pressure shaft design.

Fig 2. A high pressure shaft in reverse. The unusual design avoids potential hydrofracturing in the shallow cover of weathered volcanic tuff above the original high-pressure headrace alignment in red

This was not the end of the project's geotechnical troubles. In addition to further geological problems in the headrace tunnel, unexpected problems surfaced on both dams. The most significant of these was caused by a late change of dam type causing relocation of the Barrigon Dam. This in turn gave a construction start without sufficient investigations, and eventually required a much deeper than expected foundation plinth and a false abutment. In the tunnel, the greatest delay was caused by a 170m length of very poor Class 5 rock that took four months to excavate.

Tunnel excavation
The 4.8km long headrace is excavated entirely by drill+blast, using a fleet of 2 and 3-boom Atlas Copco and Tamrock jumbos, and progressing initially from the powerhouse adit and subsequently in each direction from two intermediate adits. The lengths of the blast rounds and the immediate support requirements were applied according to five rock classes; Class 1, being the best and requiring little, if any, support, and Class 5 being the worst. As it happened, most rock encountered was Class 2 with one significant length in very poor Class 5 conditions. At a rough estimate, overbreak on the drill+blast work, according to tunnel superintendent Bert-Ove Sjostedt, was 10-11% or about 70m² actual on a 63m² theoretical excavation line.

"In these weathered volcanic tuff deposits this was not exceptionally much" said Sjostedt, "but slightly more rock bolts and immediate support shotcrete was used than expected. That in turn was offset by longer lengths than anticipated in better rock Class 2. In the reach of poor Class 5 rock, overbreak was high, and a great deal of pre-grouting was required to stabilise the material. The rounds were shortened and spiling was also needed to control potential ravelling of the face."

All primary support in the headrace used wet-mix, steel-fibre reinforced shotcrete dosed with 45-50kg/m³ of Dramix fibre imported from Belgium. For Classes 2, 3 and 4, the primary shotcrete thickness is up to 150mm while in Class 5 it increases to 200mm. There are no steel arches in the primary support design and suggestions of embedding rebar arches in the primary shotcrete in Class 5 conditions was also rejected.

A 170m section of the headrace took four months to get through

When TunnelTalk was on site, just a week after the final breakthrough, all discussions concerned the types and quantities of final lining required in the tunnel. It was never envisaged, by any of the associated parties, that the entire tunnel would be lined and in-situ concrete was not specified for any rock class or condition in SwedPower's preliminary design. Steel lining along the tunnel length, other than at the penstock, had not been envisaged either. The approved SwedPower design determined that reaches in good Class 1 rock would remain unlined with the remainder lined with fibre or mesh reinforced shotcrete as required.

In reality, however, so little of the best rock was encountered that "it has been agreed that very little of the headrace can remain totally unlined," said Mases. "Nearly all requires some form of lining." In addition to fibre or mesh reinforced shotcrete, some applied as the primary support, heavily rebar reinforced in-situ concrete to a final i.d. of 7.6m had to be added to the scope of works. When TunnelTalk was on site, it was anticipated that a section of steel lining in the 170m length of very poor Class 5 rock might also be required. This, as well as the exact lengths of other forms of lining in the tunnel, was the subject of much study and concern. At that time (early 2003) it was also the core of what could be construed as conflicts of interest. Skanska as the leader of the EPC contract considered it its responsibility to make these fundamental design decisions, bearing in mind its fixed-price contract and its contractual responsibilities. On the other hand, AES as owner and operator of the plant for a 30 year private ownership concession period, wanted the final say, set against the fact that the warranty period on the design-build EPC contract is only three years and that, given the scheme's redesign and the cost of pumping, the system is unlikely to be dewatered for inspection within the three-year warranty period.

A subsequent report from Panama confirmed that some 6% or 260m of the 4.8km long, 8m x 8m excavated tunnel was lined with wet mix shotcrete reinforced with four layers of wire mesh, with another 240m lined with 400mm-800mm of heavy steel bar reinforced in-situ concrete. The project's 230,000m³ of shotcrete and concrete was batched on site with the batching plant working 24h/day, 7 days/week and producing a peak output of 3,000m³/week. Sika's Plastement retarder was used to give the wet mix concrete an extra 40 minutes life and Sika's alkali-free accelerator Sigunit 520 AF was dosed at 7.5kg/m³ at the nozzle into the wet mix shotcrete. In the tunnel the 24,000m³ of primary and secondary shotcrete was mainly applied using two Normet mobile robot shotcreting units with on-board wet mix pumps and accurate liquid accelerator dosing units. The in-situ concrete was cast behind 9m long forms built on site.

Formwork erected for in-situ concrete lining in the worst rock class conditions

Work on tunnel lining started concurrent with excavation in February 2003 and by early August all was complete and the system ready for watering up. The first of the system's two 60MW Francis turbines and its associated generating equipment was installed and ready for wet testing in August.

EPC post-mortem
"Time was the greatest problem on the project," said Mases for the Skanska construction team. "Had we had just one more year - six months to undertake more comprehensive geotechnical and geological investigations and another half year to concentrate on the detailed design - this would have made a tremendous difference.

We would have had tighter control on our own construction time and cost elements and the owner could have spent less on supervision. The concept needs a strong designer to advise and direct not only the EPC group's civil contractor but also the owner's team and its experts. A strong design team operating directly on site can save significant amounts of money by making quick on-the-spot cost and time saving design decisions."

For AES, Jorgensen said that the EPC form of procurement provides a "whole different set of problems for the owner. Contractors working to a fixed lump sum could be motivated to cut corners," he said. "This places a greater emphasis on inspection and we have 2-3 times the inspection force than is usual for design-bid-build projects. Here on site AES has 39 staff. This includes MWH personnel who are responsible also as our design consultant/quality controller for reviewing SwedPower's design work."

In describing the concept further, Jorgensen explained that in addition to the US$180,000/day late penalty, there is also the possibility of earning bonuses. "In addition to early completion, there is the potential for a bonus for improving generating capacity. A smoother tunnel lining for example will reduce head losses and could increase the specified minimum 118m³/sec flow capacity through the turbines." This line of thinking placed further emphasis on the tunnel lining issues, which were already the crux of many intense discussions. While agreement was eventually reached on the lining issues, Jorgensen said that AES would seek to extend the civil warranty period on any future EPC hydro projects contract.

Esti's EPC contract is based on FIDIC terms and conditions and while there is no disputes review board (DRB), disputes can go to arbitration if they remain unresolved after 90 days.

"Within the project's US$213 million fixed-price lump civil works account for about US$150 million," said Mases. "Against that we set a contingency, but this, and more, has been consumed given the situations encountered. A contingency of at least 15% of the bid price and priced into the contract bid would be recommended for future EPC contracts. Contingency funds need also to cover procurement of generating equipment and any potential penalties."

"Although the EPC method involves 100% risk responsibility for the construction team, I do see the concept becoming more common and for two reasons," continued Mases. "First the client knows the exact price of the project from the outset and this, in managing budget and finances, is more important than obtaining the lowest price. Secondly, it is a better way for the contractor to operate. As well as suiting the scope of the job to his particular way of working, the work is set against a fixed set of owner requirements. There can be no owner-initiated change orders and the element of self-certification by the contractor should require less client supervision, which makes life easier for the owner as well. Yes, we like the method and would go into it again - with the suggested improvements and refinements."

Back in the head office in Panama City, Ramos for AES said he too would recommend the concept for future hydropower projects. "It is fast, it is a fixed-price for the owner's fixed project requirements, and the owner is not involved in the co-ordination of the different project suppliers and contractors."

Footnote:
EPC contracting has become a regular procurement concept for hydro project developments in Latin American and South American countries and a contract for major repairs of the Esti project headrace tunnel was awarded in 2011 to SELI of Italy.

References

Progress at Chile's Chacayes hydro scheme - TunnelTalk, March 2010
EPBM push in Panama - TunnelTalk, Feb 2011
Panama awards urgent Esti headrace rehabilitation - TunnelTalk, Dec 2011

Search

TunnelTalk Articles TunnelCast Videos Archive Feedback Current Diary Dates World News Briefs Career Moves Accolades & Tributes Stay Connected Procurement Job Opportunities Equipment Sales DIRECTORY Marketplace About Us Advertising Contact Us Legal Notices Home Office     Latest Alert Alert Signup Alert