Like most auto engines, HCCI engines inject a premixed charge of fuel and air into the cylinder. But combustion is not set off by a spark plug. Instead, the mixture is compressed until it autoignites, much like in a diesel at a much lower temperature. This virtually eliminates NOx emissions and lowers throttling losses, which could lead to a 30% boost in fuel economy. HCCI engines would also use a higher ratio of air to fuel, high compression ratios (on the order of 12:1), and burn a variety of fuels, including methane.

Researchers at Honda hope to use HCCI technology on a V6-powered Accord sedan to boost mileage to 50 mpg, double the 25 m pg it gets today. They ultimately want to add electric power to the Accord to get an HCI gas/electric hybrid that gets 70 mpg.

But there are several technological problems to solve before HCCI engines can be mass produced for consumers. The biggest challenge is controlling the combustion at low and high speeds so that HCCI engines run smoothly.

The trick, according to scientists, is to ensure the air/fuel mixture ignites exactly when the piston reaches its zenith. But HCCI engines inherently lack traditional methods of timing ignition. In diesels, for example, ignition starts when fuel is injected into the hot, compressed air.

And in standard engines, sparking the plug initiates ignition. HCCI engines also lack the means to control the heat-release rate. In diesels, this takes place by adjusting the rate and duration of fuel injection, while in standard engines, engineers manage it via control of spark plug intensity and spark duration.

Overall, the challenge is to establish closed-loop control of the fuel and air systems to keep combustion optimized despite changes in speeds and loads. If the control is off, for example, and the there is a misfire, the gas mixture injectedin to the cylinder for the next cycle-could be too cold for autoignition and the engine could stop.

HCCI engines also tend to have low power densities, so a 250-cu-in. HCCI engine will generate less horsepower than a traditional internal combustion engine with the same displacement. This means friction losses in HCCI engines are a larger fraction of total power output than in traditional car engines.

At Lawrence Livermore Laboratory, researchers specialize in mathematical modeling of HCCI combustion, a carryover from the days when they tested and built nuclear warheads. They are studying two options for controlling HCCI engines. One is to recirculate exhaust gases into the fuel/air mix to quickly raise its temperature. The other is to add dimethyl ether to the fuel/air mixture to improve combustion.

Several researchers suggest bootstrapping the technology rather than striving for a widespread automotive application right off the bat. They want to use current state-of-the-art HCCI techniques to build stationary engines, such as a 5-MW generator. They would run at constant speeds under constant loads, sidestepping most of the control issues. What they learn from designing, building and operating the generators can then be used in later applications.