| The research presented in this thesis was aimed at investigating a sensor validation model capable of detecting, isolating, and compensating multiple simultaneous failures of sensors that participate in controlling stationary internal combustion engines. The proposed sensor validation model was embedded in a software application designed to operate in real-time for preventing the introduction of erroneous information to the control system, thus avoiding unnecessary or incorrect control actions and guaranteeing personnel and equipment safety. In addition, the developed sensor validation system was integrated with an emission estimation system and an emission control system for ensuring reliable emission quantification even when sensors operate inadequately. The resulting three-module software system was integrated with an engine monitoring system REMVue (REM Technology Inc., Calgary, Alberta, Canada) for (1) providing real-time pollutant emission estimations, and (2) supporting failure diagnosis and compensation. |
