PREDICTION OF PERFORMANCE MAPS FOR HOMOGENEOUS-CHARGE COMPRESSION-IGNITION ENGINES

Abstract

Performance maps are useful for evaluating potential engine designs because they encompass the full range of engine speeds and loads encountered in a normal driving cycle; however, these maps are time-consuming and expensive to construct experimentally. An efficient method is presented for constructing performance maps for homogeneous-charge compression-ignition (HCCI) engines by determining the optimum operating conditions (equivalence ratio, valve timing, etc.) and associated performance characteristics (fuel consumption, peak cylinder pressure, etc.). The numerical procedure used to construct the performance maps combined engine cycle simulations to model the gas exchange process and perfectly stirred reactor (PSR) simulations to model the compression, combustion, and expansion processes. A fast numerical solver was employed to exploit sparsity in the PSR model equations and allowed the use of very detailed kinetic mechanisms for primary reference fuel and n-heptane. A major limitation of HCCI engines is that they are limited to low load operation because of engine “knocking” at higher loads. A fundamental criterion is presented for predicting the HCCI knock limit and the corresponding upper bound on engine load. This modeling strategy is first used to simulate a baseline engine that has been tested experimentally. Subsequently, several other cases are described which investigate the influence of compression ratio, fuel, and supercharging on HCCI performance. It is observed that acceptably high loads (11 bar brake mean effective pressure) can be achieved without knocking using an HCCI engine with a relatively low compression ratio, low-octane fuel, and moderate boost pressure. For the cases investigated, knocking occurred for equivalence ratios greater than about 0.6, and the fundamental origin of this limit is discussed.

Keywords:

• HCCI combustion
• engine knock
• operating range
• performance maps

We are grateful to Prof. John Heywood and Prof. Ahmed Ghoniem for helpful discussions. Thanks also go to Prof. Paul Barton and Dr. John Tolsma for assistance with the DAEPACK software. Financial support for this work was provided by Ford Motor Co. and the U.S. Department of Energy under agreement DE-GC04-01AL67611.

Notes

RON = research octane number, cad = crank angle degrees, EVO = exhaust valve opening, EVC = exhaust valve closing, IVO = intake valve opening, IVC = intake valve closing.

^a Estimate based on coolant temperature.

PRF 92 = 92 octane primary reference fuel (92 vol% 2,2,4 trimethylpentane, 8 vol% n-heptane).

¹Note that the partial decoupling of the cycle simulations and PSR model discussed previously is still valid in this case even though ignition does not occur near top center—the residual fraction is very low (2–4%) over much of the map, so there is little communication between cycles and details of the combustion process have little effect on the charge conditions at IVC.

η_f,b = brake fuel conversion efficiency, r _c = compression ratio, PRF = primary reference fuel.