Diagnosis methodology for the turbocharger groups installed on a 1 MW internal combustion engine

The issue of internal combustion engine (ICE) diagnosis attracts great interest because modern engines need continual control of the operational status, in order to obtain high efficiency in energy conversion and accurate control of the polluting emissions. In particular, in reference to an alternative ICE of 1 MW, the present study relates the development, through the design of neural simulators, of the turbocharger maps to reproduce the operational states characterized by new&clean conditions and allowing the evaluation of particular "health state" indices of such a module. In detail, after an experimental campaign, turbocharger fundamental characteristics referred to new&clean conditions, such as the compressor isoentropic efficiency and the mass flow elaborated by the turbine, were evaluated at different operation conditions of the alternative ICE Subsequently, the neural simulators were developed through the training and test of different neural architectures.