Severe runner blade cracking problems initiated a turbine refurbishment programme at Rocky Reach hydro power project in the US

Public Utility District No.1 of Chelan County in Washington state owns and operates the second largest non-federal hydroelectric generating system in the US. Two of the District’s three hydroelectric projects, Rocky Reach and Rock Island, are part of an 11-dam

system on the Columbia river. These, along with the third hydroelectric project, Lake Chelan, have a combined total generating capacity of over 2000MW and produce about 11,000GWh of power annually. The power is then sold to Chelan County residents, Alcoa’s Wenatchee aluminum smelter, and four principal power purchasers (Puget Sound Energy, Avista Corporation, PacifiCorp, and Portland

General Electric).

The project

Rocky Reach is a run-of-river scheme that consists of a system of dams and abutments, a gated spillway, fish passage facilities, power distribution systems, and a power house with 11 generating units. An extensive rehabilitation programme focusing on the turbine generator units and their appurtenances has recently been undertaken.

Units 1 through 7 are six-bladed vertical Kaplan turbines, rated at 140,000hp at 28m head and 90rpm, and were built in the early 1960s by Allis Chalmers. Each turbine has a semi-spiral casing, 20 stay vanes, and 20 adjustable carbon-steel cast wicket gates, semi-spherical discharge ring, and an elbow draft tube. The runner has six adjustable blades, with the blade servomotor in the intermediate shaft between the turbine and generator shafts. Turbines of the remaining units, 8 through 11, were originally vertical propeller turbines, rated at 177,000hp and 85.7rpm, and were built in the early 1970s by Dominion Engineering Works.

Each turbine has a semi-spiral casing and elbow draft tube. The distributors have 20 stay vanes and 20 adjustable carbon-steel fabricated wicket gates. The original turbine runners had five fixed blades made of carbon steel.

In the early 1990s, Units 1 through 7 began experiencing severe runner blade cracking, starting at the interface between the blades and the trunnions and progressing diagonally into the blade leaves. The situation was aggravated further because no stress-relieving grooves were provided in the original design. Attempts to repair cracks in carbon steel Kaplan blades in situ without proper stress relief have generally failed. Therefore this remedy was not even considered for Rocky Reach turbines 1 through 7.

A 1992 feasibility study confirmed the technical and economical feasibility of replacing the entire runner assembly with new state-of-the-art runners. A basic philosophy and approach for the turbine rehabilitation project was developed.

The approach used competitive bidding, turbine model development by two manufacturers, and competitive model testing of both models in an independent hydraulic laboratory. The winning manufacturer was then awarded the rest of the contract for design, manufacturing, delivery, and installation of the replacement turbines.

Voith Riva Hydro of Milan, Italy and Sulzer Escher Wyss of Zürich, Switzerland were the two best-qualified bidders. They each built and tested a homologous model of the existing turbine on a 1:20 scale, and then designed and developed a replacement runner. The existing water passages were identical and the specifications required the models be tested under the same hydraulic conditions to be fair to both bidders.

Under a separate procurement process, ASTRÖ of Graz, Austria was selected as the independent laboratory where the models of both bidders were retested.

In June 1994, after completing the competitive model testing, the District awarded the contract for the rehabilitation of units 1 through 7 to Voith Riva Hydro. During the model development, new ideas related to the safe passage of juvenile fish through the turbines were identified. To fully explore these ideas, Voith Riva Hydro conducted additional turbine model development.

Fish-friendly modifications, primarily based on the elimination of wedge-shaped gaps between runner blades and adjacent components, were developed, implemented on the model and tested. Elimination of the upstream blade-hub gap was later applied to all 11 Rocky Reach turbines and other hydroelectric projects along the Columbia river, such as the Wanapum and Bonneville.

The manufacturing phase started in June 1994, immediately following completion of the model testing. Replacement parts were manufactured by Voith Riva Hydro in Milan and by various subcontractors throughout Europe.

Scope of rehabilitation

The rehabilitation of units 1 through 7 consists of the replacement of the turbine runners with state-of-the-art Kaplan runners, having six, stainless steel blades, while maintaining the same diameter and setting as the existing runners. The increased efficiency of the replacement runners, and the elimination of the outages due to runner blade structural failure, more than offset the project cost. The discharge ring inner surface is being overlaid with stainless steel and restored to the original shape.

The thrust bearing is being modified by adding a hydraulic balancing system for the thrust pads, so that the bearing capacity will be adequate under the most adverse operating conditions. The generator stator sole plates are also being replaced to provide controlled thermal movement. In addition the governor hydraulic system is being refurbished while retaining the original 350psi (2.4MPa) working pressure, the main pumps were replaced, the volume of the oil sump and accumulator tanks increased, and the regulator and unit controls replaced with modern digital technology.

All parts that are to be reused in the project are inspected and restored to original condition or replaced, as necessary.

A feasibility study performed on rehabilitation of the four propeller turbines indicated positive benefits for replacement of the existing propeller runners with state-of-the-art-Kaplan runners.

The new runners are larger in diameter, have five stainless steel blades, and are set approximately 0.9m lower than the existing runners. Rehabilitation resulted in improved turbine performance and resolution of the vibration and structural problems. Modification to the stay vanes, in order to improve the flow pattern at the inflow edges, is also yielding added efficiency benefits. The distributors are being refurbished and new self-lubricating bushings installed.

The diameter of the new semi-spherical stainless steel discharge rings was increased to accommodate the larger diameter and lower setting of the new runners and to maintain the original hydraulic capacity. The thrust bearings are being replaced to match new loads.

In addition, the conversion from a propeller to a Kaplan-type turbine necessitated the replacement of the entire governor system, the addition of oil heads, and modification of the turbine and generator shafts to accommodate the oil control pipes. The operating oil pressure of the new governor system was increased to 785psi (5.4 MPa). Each turbine is controlled by a new digital governor and a programmable logic controller (PLC).

Site delivery of the replacement parts for units 1 through 7 was completed by the end of 1999. While the rehabilitation of units 1 through 7 was moving forward into the installation phase, bid documents for the rehabilitation of units 8 through 11 were being prepared. During the preparation of those documents, the need to enhance the overall operating flexibility of the power house was identified, in addition to solving the vibration, cavitation and blade cracking problems of the existing units. Rehabilitation of units 8 through 11 was originally scheduled to follow rehabilitation of units 1-7.

The need to integrate the rehabilitation of units 8 through 11 into the overall programme became more pressing in early 1995 due to increased restrictions on when units would have to remain in service to aid downstream fish passage. Since the flexibility to sequence units for rehabilitation, regardless of whether they belonged to the unit 1 through 7 or to the 8 through 11 group was an important factor to the District, the runner replacement and turbine rehabilitation for units 8 through 11 was added to the existing contract for the unit 1 through 7 runner replacement.

Another benefit of this arrangement was that the design features, namely the fish friendly enhancements, developed by Voith Riva Hydro for units 1 through 7 would also be available for Units 8 through 11. The scope of work was finalised, contracts negotiated, and performance guarantees established. The rehabilitation of units 8 through 11 was accomplished in two phases.

Under the first phase, Voith Riva Hydro performed the design of replacement turbine runners and appurtenant parts, and developed a homologous turbine model, including the fish-friendly features. Finally, the turbine model performance was verified by the independent hydraulic laboratory (ASTRÖ). After satisfactory completion of the first phase, the District issued a notice to proceed for the second phase, which consisted of the manufacturing and installation. Manufacturing of the Unit 8 through 11 components started in June 1996.

As with units 1 through 7, the runner blades, integral with the trunnions, were cast of 13-4 Cr-Ni stainless steel at the Creusot Loire Industrie foundry in France. The blades were machined at the Voith Riva Hydro plant in Milan, which in the meantime acquired the five-axis milling technology. The runner hubs were cast by River Don foundry in Sheffield, UK. Components were shipped back to the Voith Riva Hydro plant in Milan, where the individual systems, such as the runner, were assembled to the maximum extent possible to facilitate the field assembly and installation. The governor pumping units were manufactured by Rexroth in Pennsylvania, US and, after shop inspection, were shipped directly to the jobsite. The site delivery of the replacement components for units 8 through 11 was completed in early 2000.

Downstream migratory fish passage had become an issue as well, and it had been determined that the majority of fish pass through units 1, 2, and 3 on their way down the river during their migration period of April to August. It became obvious that units 1, 2 and 3 could not be shut down for rehabilitation during this migratory period. At that time, the development of the fish-friendly turbine features was still in progress and it was beneficial to rehabilitate units 1, 2, and 3 as the last units of the unit 1 through 7 group, so that they could fully benefit from the results of the development. The schedule was modified to accommodate fish considerations to the final sequence of rehabilitating units 6, 7, 5, 4, 3, 9, 2, 10, 11, and 1.

As a result of a December 1999 incident involving unit 3, in which a rotor rubbed the stator causing considerable damage, the rehabilitation program was suspended for most of 2000. The programme resumed in November 2000 with the rehabilitation of unit 10, which was completed in July 2001. The rehabilitation of the remaining two units should be completed by 31 March 2003.

Development and testing

There were several aspects of the model development approach used for units 1 through 7 that greatly benefited the District. Firstly, the competitive model testing forced the manufacturers to continue improving the turbine performance even if they had met their guarantees. This was because one competitor was not aware of the achievements of the other and the second phase of the contract was to be awarded to the competitor with the lower evaluated contract price, the equipment price reduced by the efficiency benefit. This project was large enough in terms of cost, benefits, and number of like units that the competitive model testing was justified. For smaller projects this is not always the case.

Another important aspect was that the acceptance of the turbine performance was done on the model level. That eliminated costly and uncertain field efficiency testing as a part of the contract. Having both models of the same size and tested under the same hydraulic conditions in the independent laboratory made the results directly comparable. Building and testing the model of the existing turbines was the next important step. That provided a benchmark for accurate determination of the performance benefits obtained from the runner replacement. Stepping up the incremental efficiency from the model to the prototype level using the hydraulic affinity laws was deemed much more accurate than attempting to perform field efficiency tests before and after the turbine rehabilitation. The hydraulic design for units 8 through 11, significantly different than that for Units 1 through 7, had to address not only a conversion from a fixed blade propeller turbine to a double regulating Kaplan, but also an increase in runner diameter by about 5% and a setting lowered by nearly 1m. For this reason it was pointless to build and test a model of the existing runner.

Similarly as for units 1 through 7, an opening tendency of the runner blade torque was required for all conditions to allow the mechanical speed limiters to maintain the maximum overspeed below the generator design speed. The official model tests for units 8 through 11 were also performed in ASTRÖ’s independent hydraulic laboratory in Austria, where the results of the model tests previously carried out in the manufacturer’s laboratory were confirmed. Once the contractual tests were completed, the leading edges of the selected stay vanes were extended to further increase the turbine efficiency.

Following the contractual model test development programs, which produced ‘standard’ Kaplan turbine designs for Units 1 through 7 and units 8 through 11, the District pursued investigation of fish friendly turbine features on the model turbine level. The tests were carried out in Riva’s laboratory in Milan and in ASTRÖ laboratory in Graz. These focused on the minimising of the runner hub-blade gaps upstream and downstream of the blade trunnion. Also fully spherical discharge ring was developed for Units 8 through 11, closing the gap between the runner blade periphery and the discharge ring along the entire length of the blade. Attempts were made to model the movement of small fish that would pass through the turbine, on a 1:20 scale hydraulic turbine model. This was done by means of inert, neutrally buoyant particles made of a material having density equal to 0.98 the density of water. The model was modified to allow observation of the ‘fish’ using special endoscopes and a high speed TV recording systems.
Related Articles
Under inspection
Spotlight on … uprating and refurbishment

Bibliography

Hron, McKee, Bramati, Rossi, “Rocky Reach Kaplan Turbine Replacement,” presented to Waterpower ‘97 conference in Atlanta, GA, August 1997.
McKee, Rossi, “Rocky Reach Kaplan Turbines: Development of Fish Friendly Runners,” presented to Hydropower conference in Barcelona, Spain, June 1995, and to 4 a Conferenza Internazionale Hidroenergia ‘95, Milano Italy, September 1995.
Hron, McKee, Strickler, “Large Kaplan Turbine Runner Replacement, ” presented to Waterpower ‘95 conference in San Francisco, CA, July 1995.
McKee, Rossi, “Rocky Reach Kaplan Turbines: Development of Fish Friendly Runners,” presented to a Conference on Uprating & Refurbishing of Hydro Power plants, Nice, France, October 1995.
Lang, Christman, “Fish Bypass System Impact upon Turbine Runner Performance at Rocky Reach Dam,” presented to IAHR Symposium, Valencia, Spain, September 1996.