Evaluation of an Advanced Diagnostic Technique for the Determination of Diesel Engine Condition and Tuning Based on Laboratory Measurements 2010-01-0154

In the present work is presented a detailed evaluation of an advanced diagnostic technique, developed by the authors, for the determination of diesel engine condition and tuning. For this purpose, an extended experimental investigation has been conducted on a prototype test engine installed in the author's laboratory. During the measurements various operating parameters (i.e. torque, fuel consumption, injection pressure, cylinder pressure, peripheral temperatures etc.) have been recorded at various operating conditions (i.e. engine speed and loads). Initially the engine operated at its normal conditions (i.e. reference state). Then, two “virtual” faults (i.e. reduction of injector opening pressure and increase of cylinder mass leakage) were introduced, that affected engine operation. The diagnosis approach, evaluated in the present work, is based on a thermodynamic simulation model and on the estimation of model parameters, referred to as model constants, which are considered as an index for the condition of specific engine components (i.e. injector, piston-ring assembly etc.). Initially, the diagnostic technique was applied at normal engine operation and thus its ability to estimate engine condition and tuning was investigated. The evaluation procedure was based on the comparison of measured and estimated operating data such as SOI, fuel consumption, power output, exhaust valve timing etc. Further on, the technique was applied at “faulty” engine operation. The objective was to evaluate the ability of the technique to detect the actual cause for each virtual fault that was introduced. From the results, it is shown that the developed technique captures adequately both engine tuning and condition and moreover it provides the actual cause for an engine irregularity. This is most important for field applications (marine, power generation etc.), where the amount of available engine data is limited and usually not adequate for direct estimation of engine tuning and condition.