On-Line Compressor Cascade Washing for Gas Turbine Performance Investigation

Abstract

On-line compressor washing for industrial gas turbine application is a promising method of mitigating the effects of compressor fouling degradation; however there are still few studies from actual engine experience that are inconclusive. In some cases the authors attribute this uncertainty as a result of other existing forms of degradation. The experimental approach applied here is one of the first of its kind, employing on-line washing on a compressor cascade and then relating the characteristics to a three-dimensional axial flow compressor. The overall performance of a 226MW engine model for the different cases of a clean, fouled and washed engine is obtained based on the changing compressor behavior. Investigating the effects of fouling on the clean engine exposed to blade roughness of 102μm caused 8.7% reduction in power at design point. This is equivalent, typically to 12 months degradation in fouling conditions. Decreases in mass flow, compressor efficiency, pressure ratio and unattainable design point speed are also observed. An optimistic recovery of 50% of the lost power is obtained after washing which lasts up to 10mins. Similarly, a recovery of all the key parameters is achieved. The study provides an insight into compressor cascade blade washing, which facilitates a reliable estimation of compressor overall efficiency penalties based on well established assumptions. Adopting Howell’s theory as well as constant polytropic efficiency, a general understanding of turbomachinery would judge that 50% of lost power recovered is likely to be the high end of what is achievable for the existing high pressure wash. This investigation highlights the obvious benefits of power recovery with on-line washing and the potential to maintain optimum engine performance with frequent washes. Clearly, the greatest benefits accrue when the washing process is initiated immediately following overhaul.

Prediction Reliability of a Statistical Methodology for Gas Turbine Prognostics

“…Larsson et al [36] found out that, for a mechanical drive site, the deviation from the nominal reference value of compressor efficiency and mass flow due to fouling could be up to 3% and 2% respectively over a time frame of 13 days, but, after compressor wash, the degradations are partially restored by about 2%. In particular, with regard to performance recovery, Igie et al [22] state that an optimistic recovery of 50% of the lost power and of all the key parameters can be obtained after compressor washing [22]. Therefore, scenarios #1 through #4 reproduce all these different situations.…”

“…Therefore, scenarios #1 through #4 reproduce all these different situations. Scenarios #5 and #6 are also considered in this paper to account for a considerable decrease of 10% of the parameter Q(t), since fouling can cause an 8.7% reduction in power over a one-year period [22]. Recovery is instead assumed the same as in scenarios #1 through #4.…”

“…Kurz and Brun [21] review and discuss fouling mechanisms in axial compressors, based both on technical literature and on the authors own tests. Ugie et al [22] investigate the effects of fouling for a typical oneyear degradation. Fabbri et al [23] present a methodology which can provide an indication of the best time to intervene on a component of which the performance degrades over time; this methodology needs a collection of data to define the ideal condition, the degradation curve and the average cost of output exergy flow.…”

Application of a Statistical Methodology for Gas Turbine Degradation Prognostics to Alstom Field Data

“…The most common measures to reduce fouling are compressor washing and inlet filter exchange. Among the many papers produced on fouling effects, the following specifically deal with the time evolution of compressor fouling and its effects over time on gas turbine performance [19][20][21][22][23][24][25][26].…”

Método para optimizar la planificación de los lavados “off-line” en compresores de turbinas de gas

“…Se han realizado amplios estudios sobre lavados on-line (Igie et al (2011), Syverud (2007 o Schneider (2010)). En ellos se explica que el lavado on-line se realiza normalmente con la turbina funcionando a plena carga y por lo tanto, con los VIGVs abiertos del todo.…”