Estimation in Centrifugal Compressors via Particle Filtering

A turbocharger is a fundamental device to improve power of internal combustion engines. In this paper, we focus on plants constituted by two turbochargers, each of which composed by the centrifugal compressor and the turbine, working together. In order to ensure the reliable and efficient operation of the turbochargers, new maintenance strategies are being developed, often based on the continuous measurement and monitoring of key variables. To this purpose, a model, which can be constructed from first principles, is a fundamental tool. However, some main variables are not accessible, posing an estimation problem. Herein, an innovative approach based on particle filtering is proposed. Its validity is assessed through experimental results reported in the paper.