Innovative techniques for the Condition Monitoring of an Internal Combustion Engine using Turbocharger speed

A numerical and experimental analysis of the turbocharger speed in a four-stroke turbo-diesel engine is presented in this work. The first part of the research is focused on the possibility to monitor the performance of the engine injectors that are characterized by a time deterioration, which is the effect of deposits of particulates in the nozzles. With this aim three different methodologies were developed, exploiting the numerical model, based on time and frequency content of the turbocharger speed. The first methodology is based on the evaluation of the area between the turbocharger speed and the line connecting the extreme values of the influencing windows related to the cylinder object of the analysis. The second methodology is based on the acceleration of the turbocharger, the peak value of the acceleration in each contribution window is evaluated and compared with the expected one to quantify the amount of fuel injected in each cylinder. Finally, the methodology based on the frequency content of the turbocharger speed is presented, for this case two main parameters are considered: the module of the fourth order, that allows to measure the fuel quantity injected in all cylinders, and the first order phase to detect if there is a problem in the injection in one cylinder and to detect which cylinder is involved. A specific experimental campaign was carried out to verify the reliability of these three methodologies, this experimental campaign was conducted with the engine provided by Yanmar Japan and exploited for the calibration of the numerical model. For the purpose of detecting injection issues, two different set of injectors were tested, one, set new, with a homogeneous performance among cylinders and one, used and deteriorated, with a different performance for each cylinder. From the analysis of the results related of the methodologies developed a very good reliability of each method was detected. In addition, the possibility of detect misfiring events was positively tested. Finally, a technique to distinguish among different typical faults of diesel engines was presented. This strategy relies on different engine parameters such as the temperature and the pressure of the exhaust manifold. Exploiting the information coming from these sensors and the turbocharger speed it is possible to detect the fault that is responsible for a change in the engine condition among selected possibilities: EGR variation, Injection variation, in terms of timing and injected fuel mass, and turbine backpressure. The fault detection strategy was developed and tested both numerically and experimentally confirming its ability to monitor the condition of the engine.