Save Article Instructions
Close 

Prediction Trumps Prevention

By Steve Blankinship, Associate Editor

If prevention is worth a pound of cure, then knowing just when prevention is needed - and when it isn’t - is worth even more. Predictive maintenance practices have helped one of the largest utilities in the United States to elevate the availability at two of it largest coal-fired plants.

With more than 41,000 MW of generation and a enough transmission/distribution lines to circle the earth, Atlanta-based Southern Company serves regulated and competitive markets across the Southeast. Two of Southern’s best performers - Plant Bowen in Georgia and Plant Miller in Alabama - consistently rank among the best performing coal plants in the nation. According to Southern’s chief production officer Jerry Stewart, both plants owe their success to a combination of predictive maintenance, employee safety (an 86 percent reduction in reportable injuries over the last 12 years) and a belief that providing the lowest-cost power is in the best interest of both the company and its customers.

The 3,200 MW Plant Bowen is about 60 miles northwest of Atlanta. Construction on the first of four supercritical units started in 1967 with the fourth unit entering service in 1975. (Southern has 13 supercritical units in all.) Consuming about 8 million tons of Central Appalachian coal annually, Bowen’s four boilers are from Combustion Engineering. Two turbine/generators are Westinghouse and two are GE. Units 1 and 2 are rated at 700 MW while 3 and 4 are rated at 884 MW. Cooling towers provide 100 percent of cooling requirements. As a plant built before new source regulations, Bowen is getting scrubbers. All units had electrostatic precipitators when they were built. Selective catalytic reduction has recently been added to all four units to meet ozone regulations.

With 400 employees today (down from 600 in the late 1970s), Bowen has raised its annual capacity factor from less than 70 percent in the early 1980s to 81 percent today, a figure that would have been in line with nuclear power performance a decade ago. Even with the demand growth, Bowen has achieved forced outage rates below 2 percent. High availability has been achieved through enhanced maintenance practices and training.


Plant Bowen’s two stacks rise 1,000 feet above Bartow County near Cartersville, Georgia.
Click here to enlarge image

About four years ago, Southern adopted the Plant Reliability and Optimization Program developed by EPRI, a structured maintenance program based on planning maintenance using predictive parameters rather than performing preventive maintenance at set time intervals.

Stewart said a large part of the approach is based on the knowledge that when maintenance is performed, things can potentially cause a future problem. “For example, we analyze oil before we change it rather than just doing it after a specified time.” If the analysis shows no problem, it isn’t changed. The same applies to predicting equipment failure. “We take regular vibration readings,” Stewart said, along with recording temperatures and monitoring trends. “Our maintenance planning has become more effective.”

Located about 20 miles from Birmingham, Plant Miller is Alabama’s largest power plant, producing about a fourth of the state’s electric power. Miller has four 660 MW subcritical units comprised of Babcock & Wilcox boilers and GE turbine generators. The units burned Alabama sub bituminous when first commissioned, but switched to Powder River Basin coal in the late 1990s. During Southern’s peak season (May - September) Miller is the fleet’s lowest-cost performer, after hydro and nuclear. It has an operational staff of about 320.

“Miller has consistently set unbelievable performance numbers,” says Stewart. The plant’s forced outage rate for all of 2006 was 0.64, and has been below 1 percent for seven consecutive years. During peak season, it’s especially critical that Miller run well; during the most recent peak season, Miller performed at 0.28 with Units 2 and 3 performing at 0.01 and 0.02 respectively.

Even small fractions of percentiles of availability can quickly translate into millions of dollars - lost or saved. “The more expensive power is what we have to buy from somebody else,” Stewart said. “The best reliability we can have during that May through September timeframe, the better off the customer is.”

To Stewart, reliability is achieved during scheduled maintenance outages when the plant isn’t running. The Miller and Bowen units come off line every two years and stay off for 30 to 37 days as in-depth routine maintenance is reformed. Those 37 days are key to achieving 23 months of reliability, he said.

Bowen and Miller reflect an industry trend to stretch maintenance intervals. “We don’t do scheduled spring and fall maintenance programs every year anymore,” he says. “This spring we’ll take Miller 3 and 4 off for 37 days each. We took Units 1 and 2 off last year.” Units 1 and 2 share a cooling tower, as do 3 and 4. So 1 and 2’s outages overlap by one week so that equipment used for both units can also be looked at. The process repeats every two years.

Stewart says Southern is trying to move to a 23-month run cycle across its fleet. He concedes that some U.S. coal-fired units probably won’t ever go that long - lignite units, for example - because of fuel quality issues. He also sees a few U. S. utilities trying to go to three years between outages. It’s too early to know how such intervals will affect reliability. Plants will likely have to complete some cycles to learn what equipment is wearing out beyond two years.

Finding the right balance remains a big challenge for coal plant operators. “That’s what we’re always trying to figure out,” Stewart says. “We’re not tied to anything - 12 months, 18 months or 24 months. We’re trying to figure out what is that optimum between outages that will allow us to provide the most megawatts, but still balance that with forced outage rates.”


To access this Article, go to:
http://www.power-eng.com/content/pe/en/articles/print/volume-111/issue-2/departments/managing-the-plant/prediction-trumps-prevention.html