Search Results

Several race car competitions seek to limit engine power through a rule that requires all of the engine combustion air passes through a hole of prescribed diameter. As the approach and departure wall shapes to this hole, usually termed orifice or restrictor are not prescribed, there is opportunity for innovation in these shapes to obtain maximum flow and therefore power. This paper reports measurements made for a range of restrictor types including venturis with conical inlets and outlets of various angles and the application of slotted throats of the ‘Dall tube’ type. Although normal venturis have been optimized as subsonic flow measuring devices with minimum pressure losses, at the limit the flow in the throat is sonic and the down stream shocks associated with flow transition from sub-sonic to sonic are best handled with sudden angular changes and the boundary layer minimized by the corner slots between the convergent and divergent cones.

In recent times new tools have emerged to aid the optimization of engine design. The particle swarm optimizer, used here is one of these tools. However, applying it to the optimization of the S.I. engine for high efficiency and low NOx emission has shown the preference of ultra lean burn strategy combined with high compression ratios. For combined power, efficiency and emissions benefits, there are two restricting factors, limiting the applicability of this strategy, knocking and cyclic variability. In the ultra lean region, knocking is not an important issue but the variability is a major concern. This paper demonstrates the application of a variability model to limit the search domain for the optimization program. The results show that variability constrains the possible gains in fuel consumption and emission reduction, through optimizing cam phasing, mixture and spark timing. The fuel consumption gain is reduced by about 11% relative.

Increasing engine power and efficiency using a particle swarm optimization technique is investigated by using thermodynamics based quasi-steady engine simulation model. A simplified engine friction model is also incorporated to estimate the brake power output. Further, a simple knock model is used to make sure of knock free engine operation. Model is calibrated and validated to a Ford Falcon AU six-cylinder gasoline engine. Nine different engine-operating parameters are considered as input variables for the optimization; spark timing, equivalence ratio, compression ratio, inlet and exhaust vale opening timing and durations, maximum inlet valve lift and manifold pressure. Significant improvement of the engine power output for a given amount of induced gas is observed with the optimized conditions when compared to the corresponding power output with the reference engines normal operating conditions.

A real time energy management (EMS) optimizing algorithm is introduced that performs similar to offline dynamic programming (DP) for parallel HEVs. The EMS and the DP are compared, especially with the addition of a local hill climbing technique, to the example performance prediction of the fuel consumption of a 1.67 tonne large car using a 50 kW Honda Insight engine (representing 65% power reduction from standard) as reference. Then the performance of the vehicle in HEV mode, with a parallel 30 kW motor/generator is examined. The average improvement of this vehicle over five drive cycles from around the world is about 50% reduction in fuel consumption. Next the engine is replaced with an advanced SI turbocharged engine with assisted ignition which returns the performance to that expected of this class of car i.e. 0-100 km/h acceleration time of 7 s. This results in a 14% average reduction in fuel consumption across the five cycles compared with the base Honda engine.

There are two parts to achieving the optimization reported here. The development of an engine simulation model and an optimization algorithm. The engine performance is evaluated using a quasi-dimensional engine combustion model with sub models to incorporate friction, heat losses and abnormal combustion, that is knocking. After extensive search and development a new Particle Swarm Optimizer (PSO), has been developed. Optimization includes, for the first time, the search of discontinuous design variables. The input variables considered for this investigation are manifold air pressure, air-fuel ratio, spark timing, compression ratio, valve timing events including valve open duration, maximum valve lift and engine speed. This enables the identification of the maximum thermal efficiency at a given power output at any engine operating speed.

The challenge of tough fuel consumption reduction targets and near zero NOx emission standards can be met by optimization of the full range of engine design variables. Here these are explored through an engine simulation model and the application of an optimizing algorithm that can work in discontinuous data space. The combustion model has main features that include flame propagation, the effects of turbulence, chamber shape interaction and NOx formation. Two engine configurations are used to illustrate the application of the model and optimizer. Both allow the adoption of extra lean burn possible with LPG as fuel and EGR through an external route or cam phasing. In the first the compression ratio and cam profiles are fixed, in the second study they are also optimized.