August 15, 2007

Focus on O&M (August 2007)

PREVENTIVE MAINTENANCE
Stop valves from leaking money

Testing valves for leakage costs little but can save a lot. Energy drops across many valves exceed 1,250 Btu/lb, and leakage rates can be higher than 1,000 lb/hr. In a plant burning fuel priced at $5 to $10 per million Btu, a single valve with a modest leak (Figure 1) can account for $50,000 to $100,000 in lost fuel costs annually. That's at least an order of magnitude greater than the $3,000 to $6,000 cost to replace it.

 

 

 

The losses multiply quickly in a typical power plant, which may have hundreds or thousands of valves. By wasting precious fuel, leaks can raise a generating unit's heat rate by as much as 1% to 3%. Another penalty is the cost of producing the makeup water that literally goes down the drain.

First measure, then control

Most power plants consider a certain level of energy losses and leakage acceptable. Traditionally, detecting leakage losses has been challenging and time-consuming, and leakage rates have been nearly impossible to pin down. A relatively new technique, ultrasonic emissions testing (UET), can detect and quantify energy losses by relating the characteristics of sounds produced by leaking valves to their leakage volume or rate. As steam, saturated steam, or water moves from a high-pressure line through a leaking valve into a lower-pressure line, it produces turbulence. The nature and the level of turbulence correlate with a measurable and unique ultrasonic signal from the valve seat or area with a leak surface.

Deriving the leakage rate from the amplitude of the signal is analogous to calculating the flow rate through an orifice or constriction using the Reynolds Number. The same variables on which orifice flow calculations depend influence the relationship between the sound signature and the leakage flow rate. These variables include the pressure differential across a valve or orifice, its geometry, and the properties of the fluid being controlled.

In practice, the level of the acoustic signal is measured against background noise to isolate the sound energy associated with the leakage. Measured levels and known values of variables are entered into knowledge-based software that compares the inputs with numbers in an extensive database. The database, developed over the past decade, contains all relevant parameters, including signal levels and laboratory-measured leakage rates for many valves operating under different conditions. The software correlates the test data to the lab data and calculates the leakage rate of the valve under scrutiny with proven accuracy.

Comparing measurement methods

Another technique for detecting and quantifying valve leakage is thermography. Like "night vision" cameras and goggles, thermography uses infrared imaging to measure the heat energy emitted by an object. It can detect even small surface temperature variations inside a pipe.

Testing for leakage requires making several measurements upstream and downstream of a valve and across its body. If the conditions in a pipe are known, the surface temperature gradient can be used as a model to predict the heat flux or the temperature gradient inside it. Heat-transfer programs correlate the level of heat flux to the energy loss or leakage of the valve.

Although thermographic testing is far more common than UET, there are situations where thermographic results must be carefully interpreted:

• An elevated downstream temperature could be produced by residual heat from a recent cycling of the valve. In this case, thermographic data could be misleading.
• UET identifies the leakage at the source and measures it against background noise. With thermography, it is far more difficult to isolate valve leakage from other potential heat sources.
• A leaking valve will always emit an ultrasonic signature. But if the leak is in a saturated environment, it may not produce a temperature rise downstream.
• Thermographic testing may require removing large amounts of insulation from multiple locations along a pipe. UET only requires drilling a few, easily pluggable 1/2-inch holes in the pipe's aluminum jacket to accommodate a 3/8-inch-diameter probe to make contact with the valve and the adjacent pipe (Figure 2).
• UET gives immediate results, but thermographic measurements must be mathematically or computationally modeled to determine leakage rates.

 

 

---