Close 

A Strategic Approach to Modernization: Minimizing Risk through Refurbishment

When Duke Energy Geracao Paranapanema acquired eight hydro plants in Brazil, the company discovered three facilities needed immediate repair. However, instead of only fixing the urgent problems, Duke found ways to improve overall operational efficiency and energy production.

By Cesar Teodoro, Antonio F. Canina, and Stanley J. Kocon

During partial privatization of generation assets in the Brazilian state of São Paulo in the late 1990s, U.S. utility Duke Energy acquired the company Companhia de Geração de Energia Elétrica Paranapanema. This company, a spinoff of the Energy Company of São Paulo (CESP), owned eight hydroelectric plants on the Paranapanema River with a total capacity of 2,307 mw of generating capacity. Table 1 on page 14 lists the projects.

Click here to enlarge image

Immediately following the acquisition, a team of Duke Energy specialists visited each of the power plants. The purpose of the visits was to identify and quantify risks associated with existing equipment issues and to identify possible improvements that could increase operational efficiency and energy production of the assets.

The condition assessment highlighted several equipment problems that Duke Energy Brazil needed to address immediately to avoid possible loss of energy and capacity. Among the problems was a crack in the turbine runner in Unit 3 at the Capivara plant; turbine blade cracking in Units 1 and 2 at Jurumirim; and a need for generator rewinds in Units 2 and 3 and for a turbine rehabilitation in Unit 1 at Salto Grande.

Duke Energy Brazil recognized it needed to fix these specific problems to avoid compromising its short-term energy production capabilities. However, the company also wanted to take advantage of the outage periods to improve the overall operational efficiency and energy production of all three plants.

For each plant, the utility prioritized the need for modernization of all facility components, including turbines, generators, hydraulic structures, controls, and balance-of-plant equipment. Priorities were established based on the risk of equipment failure as well as the potential for improvement in the unit’s performance. Using the priorities as a guide, Duke Energy set a refurbishment scope and implementation plan for each project.

Owing to the need for multiple outages to be managed in a relatively short period of time and, in some cases, concurrently, Duke Energy selected a single supplier, Voith Siemens Hydro, to provide equipment and services for refurbishment of all three plants.

Overall approach to prioritizing work

For each project, a team of Duke Energy and Voith Siemens representatives performed detailed “risk-of-failure” analyses, using experience and engineering judgment. The compiled results were used to decide work priorities at each plant. The results also were used when deciding where to replace or repair various components.

Then, Voith Siemens and Duke Energy worked together to prepare a modernization schedule for each plant, taking into consideration river management and outage planning; lead time required for equipment supply; failure risk analysis; and the required plant availability and reliability to meet the terms of Duke’s energy supply sales contracts.

The following sections provide details about work performed at the three plants. Total cost of the work, which took place from January 2001 to August 2005, was US$35 million.

Capivara

The 640-mw Capivara power plant entered commercial operation in 1978. The four vertical Francis units are rated at 160 mw each.

As soon as the units started operation, then-owner CESP discovered cracks on the welded joints of the stay vanes. The cracks, 200 to 300 millimeters (mm) in length, appeared on a third of the vanes. CESP called in the original equipment manufacturer to repair the cracks. In addition, strain gages were installed to detect stress on the stay vanes. Eventually, the manufacturer modified the vanes’ trailing edges, adopting different geometries according to the vane position. Although these efforts resulted in a significant decrease in cracking, the issue was never completely resolved.

In addition to the stay vane cracks, the Francis runner in Unit 3 had developed a significant crack that propagated until one complete blade ruptured midway between crown and band. The original equipment manufacturer performed a local repair. The repair introduced a significant stress that eventually induced a radial crack in the crown, beginning in the outer diameter and propagating 500 millimeters toward the center. This crack occurred over the entire thickness of the crown. Consequently, CESP personnel continuously monitored and inspected the runner every three months to observe the crack propagation. Although the crack had certain stability, nobody could predict its behavior.

When the Duke Energy/Voith Siemens team conducted the risk analysis at Capivara, it considered two options for Unit 3: repair the existing runner (either on site or at Voith Siemens’ factory) or replace the runner. The team concluded that, while repairing the runner on site would be the lowest cost option, the uncertainty of success was too great. Additionally, repairing the runner in the factory would require an outage time equal to or greater than if a new runner was supplied, but would not provide any additional efficiency or power as compared to the original design. When full costs and benefits were considered, the option of runner replacement offered the highest benefit-to-cost with lowest risk.

However, before making its final decision, Duke Energy completed several steps to confirm and quantify the benefits of the replacement runner.

First, the relative efficiency and power characteristics of the existing turbine were established through index testing.

Second, Duke Energy authorized Voith Siemens to conduct homologous model testing to confirm efficiency of the new turbine runner, before committing to the actual manufacture of the prototype runner. In this way, Duke could verify expected efficiency of the new runner, even though the total project definition and approval was not yet complete.


The use of model tests to determine turbine runner efficiency and power improvements at Duke Energy’s 640-MW Capivara plant represents the first time such tests have been performed in Brazil as part of a modernization feasibility study.
Click here to enlarge image

Third, Duke authorized Voith Siemens to conduct homologous model testing of the existing runner to establish precise baseline efficiency. Although the expected power increase could be verified easily with a field test, the relative efficiency increase could only be determined through model testing.

Capivara represents the first time stand-alone model tests have been performed in Brazil as part of a modernization feasibility study to corroborate the efficiency and power improvements that could be achieved with a new runner.

When the test results demonstrated the benefits associated with the new runner could surpass the respective cost, Duke decided to go forward with runner replacement.

Once the turbine rehabilitation strategy was defined, Duke identified additional work to perform at the same time to take full advantage of the unit outage:

  • Replace the mechanical speed governor;
  • Replace the excitation system;
  • Replace the controls (both local and remote) for the main equipment and systems of the plant (turbine, generator, mechanical and electrical auxiliary system, substation, etc.);
  • Replace instrumentation, including temperature and pressure sensors, circuit breakers, and meters;
  • Rehabilitate the auxiliary electric and mechanical system; and
  • Replace the electro-mechanical protection.

Jurumirim

The 98-mw Jurumirim power plant entered commercial operation in 1962. The two vertical Kaplan units are rated at 49 mw each.

The turbines historically presented typical cavitation noise in some parts of their operational range. The runner blades and discharge ring exhibited both cavitation damage and cracks. The discharge ring also leaked water.


This damaged runner blade was among various problems Duke Energy Brazil inherited when it acquired the 98-mw Jurumirim hydro plant. By conducting a benefit-to-cost analysis to determine which of several possible repair/rehabilitation solutions was best, Duke not only fixed this particular problem, but increased overall project reliability and lowered operation and maintenance costs.
Click here to enlarge image

When the Duke Energy/Voith Siemens team conducted the risk analysis at Jurumirim, it determined the only way to resolve the runner blade cracking problem was to provide new blades with a new design. And, the only way to solve the discharge ring cracking/ leaking issue was to replace it with a new design. The team performed an extensive benefit-to-cost analysis to determine which of five possible solutions was best for the project.

The first option, to replace the runner blades with the same hydraulic and mechanical designs, would resolve the material issues with the original blades, but would not necessarily resolve issues related to peak stresses in the transition areas. The team concluded this solution likely would not resolve the cracking problem.

The second option, to replace the blades with the same hydraulic design but with a new mechanical design, would resolve the material and peak stress issues in the transition areas, but would not provide any additional efficiency or power as compared to the original design.

The third possible solution was to replace the complete runner, including the hub, with new hydraulic and mechanical designs with the same diameter. This option would resolve the material and peak stress issues in the transition areas and would provide higher efficiency and power as compared to the original design. In addition, this solution would allow the outage time to be reduced considerably since the new runner would be available at least six months earlier than the other options. However, even with the reduced outage time, the additional cost associated with replacing the runner hub outweighed the benefits.

A fourth option, to replace the complete runner, including the hub, with new hydraulic and new mechanical designs with a larger diameter, offered the same benefits of Option 3, with the added benefit of higher efficiency and significantly higher power. However, because the value for peaking power was not defined by the market, the benefits associated with the higher-powered runner would be negligible.

The fifth option, to replace the blades with new hydraulic and mechanical designs but use the existing hub, would resolve the material and peak stress issues in the transition areas and provide higher efficiency and power as compared to the original design. Plus, it did not carry with it the expense of a new hub. This solution resulted in the highest benefit-to-cost with lowest risk. Thus, Duke Energy selected this option for implementation.


During an outage to solve generator-related problems at the 74-mw Salto Grande hydro station, Duke Energy Brazil also modernized other plant components to increase overall efficiency.
Click here to enlarge image

Once the turbine rehabilitation strategy was defined, Duke identified additional work to perform at the same time to take full advantage of the unit outage. This additional work included:

  • Installation of a new reinforced discharge ring with a stainless steel plate in the water passage designed to avoid cracks;
  • Generator rehabilitation, including new stator winding, new stator core, and new pole coil insulation;
  • Replacement of the mechanical speed governor;
  • Replacement of the excitation system;
  • Replacement of the controls (both local and remote) for the main equipment and systems of the plant (turbine, generator, mechanical and electrical auxiliary system, substation, etc.);
  • Replacement of the electro-mechanical protection;
  • Replacement of the instrumentation, including temperature and pressure sensors, circuit breakers, and meters;
  • Rehabilitation of the auxiliary electric and mechanical system; and
  • Refurbishment of the water intake gates.

Salto Grande

The Salto Grande power plant, constructed in the 1950s, entered commercial operation in 1960. Salto Grande was among the first power plants to be constructed on the Paranapanema River. The project’s four vertical Kaplan units are rated at 18.5 mw each.

Duke’s condition assessment at the Salto Grande project identified specific generator-related issues, such as a high vibration index in the generator stator package, a high corona index, and a history of short circuits in the windings. Most of these problems stemmed from the fact that the existing Class B winding insulation has long outlived its useful life.

Before Duke acquired the Salto Grande plant, CESP had refurbished the turbine-generator in Unit 4 and had already scheduled a generator refurbishment for Unit 1. The generator work involved installing new class F stator windings, a new stator core, and new pole coil insulation. To take advantage of the already scheduled outage of Unit 1, Duke added the following components to the refurbishment:

  • Rehabilitation of the turbine;
  • Replacement of the speed governor;
  • Replacement of the excitation system.
  • Replacement of the controls (both local and remote) for the main equipment and systems of the plant (turbine, generator, mechanical and electrical auxiliary system, substation, etc.);
  • Replacement of the electro-mechanical protection;
  • Replacement of the instrumentation, including temperature and pressure sensors, circuit breakers, and meters; and
  • Rehabilitation of the auxiliary electric and mechanical system.

Results of the refurbishment

Refurbishment of all three plants was accomplished in 504,000 man-hours, without a lost work-time accident. The result – improved reliability and operational safety, as well as a reduction in risk of income loss due to operational unavailability.

In addition, Duke Energy has reduced operation and maintenance costs at all three plants and increased its ability to operate at peak demand. In addition, at the Capivara plant, Duke gained 1.7 percent in efficiency, which translates to a 3-mw increase in guaranteed energy volume.

The best way to measure the results of the refurbishment is to evaluate measurable quantities before and after completion of the work. In 2000, Duke Energy Geração Paranapanema’s machine failure index was 1.39 (i.e., the number of emergency outages per unit per year of operation). In other words, during 2000, among the 29 machines at the eight plants, there were 37 emergency unit outages.

For 2005 (the most recent year for which data is available), the machine failure index was 0.51, representing only 14 emergency unit outages during the entire year for the 29 machines). The average machine failure index of all other electric utilities in Brazil for 2005 was 2.30. s

Mr. Cesar Teodoro can be reached at Duke Energy International, Av. Nacoes Unidas, 12901, Sao Paulo, 04578-910, Brazil; (55) 14-33429011; E-mail: cteodoro@duke-energy.com. Mr. Canina can be reached at Voith Siemens Hydro Power Generation Ltd., Rua Friedrich von Voith, 825 Predio, Sao Paulo 02995-000 Brazil; (55) 11-39444264; E-mail: antonio. canina@vs-hydro.com. Mr. Kocon may be reached at Voith Siemens Hydro Power Generation, 760 East Berlin Road, York, PA 17404 United States; (1) 717-792-7082; E-mail: stanley. kocon@vs-hydro.com.

Reference

Teodoro, Cesar, Antonio F. Canina, and Stanley J. Kocon, “Hydro Modernization in Brazil: 3 Power Plants, 6 Generating Units, 5 Years Initial Concept to Commissioning – A Fast Track Hydro Modernization Case Study,” HydroVision 2006 Conference Technical Papers CD, HCI Publications, Kansas City, Missouri, USA.

Cesar Teodoro is director of operations for Duke Energy International — Geração Paranapanema. Antonio Canina is head of sales for Latin America for Voith Siemens Hydro Power Generation Ltd. Stanley Kocon is director of sales and marketing for Voith Siemens Hydro Power Generation Inc. At the time the work described in this article was completed, Mr. Teodoro was responsible for power generation operation and maintenance for Duke Energy; Messrs. Canina and Kocon were responsible for Voith Siemens’ service and modernization market in Brazil.


To access this Article, go to:
http://www.hydroworld.com/content/hydro/en/articles/print/volume-15/issue-1/articles/cover-story/a-strategic-approach-to-modernization-minimizing-risk-through-refurbishment.html