Development of a Novel Technique to Address Compressor fouling in a Turbo-shaft Engine

Compressor washing is commonly used in gas turbine engines to retrieve engine power. Severity of fouling should be known to decide on mounted or uninstalled washing and also to optimize the time and money. The present study aims to develop a system for predicting and scheduling the washing process. One 1 MW turbo shaft engine has been taken as the model for this study. The deviations in performance parameters have been quantified based on test data over a period. Deterioration of engine health parameters namely efficiency and flow capability of compressor, gas generator turbine and power turbine are considered for analysis. Sensitivity analysis and ranking of the measurements were done using a correlative technique suggested by Stamatis. The interdependency and observability of the measurements were checked. The fault signatures of selected measurement set on component degradations were isolated and estimation charts were formed to predict the optimum time intervals for compressor washing. The study forms a base platform to apply techniques like artificial neural networks for the accurate forecasting of optimum cleaning intervals for turbo shaft engines.