Estimating in-cylinder pre-combustion mixture temperatures using acoustic resonances

The Air-Fuel ratio of automotive engines during the warm-up period is difficult to control and contributes a substantial portion of the emissions on the EPA test cycle. High bandwidth estimation of the in-cylinder charge temperature provides opportunities for improvement in spark ignition (SI) engine control algorithms. Pressure sensor based algorithms for estimating air-fuel ratio (AFR) have been shown to be improved by bulk temperature information. This paper explores the suitability of using acoustic resonances to estimate charge temperature in the presence of an unknown AFR. A technique that allows for the estimation of the average bulk charge temperature during a 1-2 msec interval based on acoustic resonance data gathered from a pressure transducer is described. A parameter estimation algorithm suitable for extracting the required frequency information from short data sets is identified. The variation in temperature estimates as a function of AFR has been explored using a computer simulation that accounts for the change in the ratio of specific heats with changing mixture strength. The results of the simulation show a relatively small effect of AFR on temperature when the AFR is under closed loop control. The performance of the acoustic resonance based temperature estimates has been evaluated by comparing them to predictions based on a polytropic compression. All evaluation of the sources of errors in the method indicates the potential for approximately 1% accuracy in the temperature estimates.