Transportation accounts for 65 percent of U.S. oil consumption. More efficient vehicle engine designs would reduce U.S. dependence on foreign oil and help mitigate global climate change caused by carbon dioxide emissions. And cleaner engines would contribute to reducing toxic pollutants in the atmosphere, such as nitrogen oxide, hydrocarbons, and carbon monoxide. One promising technology, the homogeneous charge compression ignition engine, has potential for high efficiency and low emissions.

In the HCCI engine, fuel is premixed with air, as in a spark-ignited engine, but with a high proportion of air to fuel. When the piston reaches its highest point, this lean fuel-air mixture spontaneously combusts due to the heat of compression, as in a diesel engine. A feature of the HCCI engine is that it burns cooler than spark-ignited or diesel engines. Lower combustion temperature considerably reduces the emissions of nitrogen oxide. In addition, premixed combustion in HCCI engines reduces particulate matter emissions to very low levels. An HCCI engine can operate using almost any fuel, as long as the fuel can be vaporized and mixed with air before ignition.

Schematic of dual manifold intake system used in a test engine for cylinder-by-cylinder combustion control and optimization. Control valve motors, not shown, go on top of the mixing tees.

HCCI technology could be scaled to virtually every size and class of transportation engine, from small motorcycles to large ships. Stationary HCCI engines could replace the spark-ignited and diesel engines currently used for distributed power generation, especially in combined heat and power applications, where recovered exhaust heat is used for hot water, space heating, or other tasks.

Although the advantages of the HCCI engine are clear, significant challenges must be overcome before the engine is commercially viable. Once engineers resolve these issues, intermediate-size HCCI engines could achieve approximately 40 percent peak efficiency versus 35 percent for spark-ignited engines. Diesel engines can achieve high efficiency similar to HCCI engines, but diesels are major sources of NO_x and particulate matter emissions.

The greatest challenge is controlling the HCCI engine's ability to operate under a wide range of speeds and loads. The HCCI engine does not have a direct combustion trigger, such as the spark plug or fuel injector used in conventional spark-ignited or diesel engines. Instead, combustion is achieved by controlling the temperature, pressure, and composition of the fuel and air. The required control system is fundamentally more challenging than for conventional engines because ignition is sensitive to small changes in temperature. When a load is suddenly added, as when a vehicle goes from idle to cruising speed, the control system must adjust the temperature, pressure, and composition rapidly enough to maintain stable combustion.

Researchers at the Lawrence Livermore National Laboratory in California have developed new methodologies for addressing these challenges. LLNL researchers Daniel Flowers, Salvador Aceves, Francisco Espinosa-Loza, Nick Killingsworth, and Tim Ross teamed up with two University of California professors, Robert Dibble of Berkeley and Miroslav Krstic of San Diego, to demonstrate controlled HCCI combustion in a stationary engine for distributed generation.

This effort has been funded by the California Energy Commission's Public Interest Energy Research Program and the Department of Energy's Office of Energy Efficiency and Renewable Energy. Caterpillar contributed the loan of a Caterpillar 3406 spark-ignited natural gas stationary engine with six cylinders and 15-liter displacement.

Use of extremum seeking for minimization of HCCI fuel consumption. Extremum seeking delays combustion timing from 3 degrees to 9 degrees after top dead center, reducing fuel consumption by more than 10 percent.

Converting the Caterpillar engine to HCCI mode required the modification of many engine components. The stock pistons were replaced with pistons typically used in the diesel version of the Caterpillar 3406, with higher compression ratio (16:1) and shallow bowl geometry. This modification is necessary for improving the thermal efficiency of the engine and to promote fuel ignition. Spark plugs were replaced by pressure sensors (one for each cylinder).

The team replaced the stock carburetor with one tuned for liquid petroleum gas. The LPG carburetor, when used with natural gas, delivers the correct fuel-to-air ratio (lean to about one-third of evenly stoichiometric) for HCCI combustion due to the higher molecular weight of LPG.

Most of the effort in modifying the engine was dedicated to the critical issue of combustion control. LLNL researchers implemented a thermal control system that consists of a dual manifold intake. Part of the intake fuel and air are heated by exhaust gases before flowing into an insulated hot manifold. A cold manifold has air and fuel flowing at ambient temperature. Combustion control is achieved by blending hot and cold gases to obtain appropriate ignition timing, much like adjusting hot water and cold water to get the right temperature for a morning shower.

Cylinder-by-cylinder control is essential for satisfactory operation. Because HCCI is very sensitive to temperature changes, even small differences of a few degrees between cylinders can cause unsatisfactory performance. Temperature differences invariably exist in multicylinder engines because of slight differences in compression ratio, or because coolant gets hotter as it flows sequentially from cylinder to cylinder.

The variability of temperature has a strong effect on combustion timing. The engine coolant path from cylinder 6 to cylinder 1 determines the relative temperature and time of ignition of all cylinders. Cylinder 6 is cooled by incoming water from the radiator and is therefore the coldest and last to ignite. Cylinder 1 is at the end of the cooling passage and it is hottest, igniting earliest.

Hot cylinders that ignite early produce a sudden and violent combustion that may damage the engine. Late-burning cylinders generate high emissions of hydrocarbons and carbon monoxide due to incomplete combustion.

The dual manifold control system implemented in the Caterpillar engine addresses this problem by detecting cylinder-to-cylinder differences in combustion timing and independently adjusting the intake temperature of each cylinder. Mixing valves in each of the six cylinders compensate for the natural variability in cylinder-to-cylinder temperature and produce consistent combustion.

Automatic control of the mixing valves is achieved using servo motors. The combustion timing is calculated based on the measured pressure signal from each cylinder, and the mixing valve position is then scheduled by the control algorithm to yield the desired combustion timing.

Engine operation is typically optimized with a mapping procedure. At each operating point, multiple timings are tried. The optimal combustion timing is determined from the mapping data. This procedure is time- and labor-intensive.

The LLNL team applied an innovative method to speed up the process: extremum seeking, a non-model-based, real-time optimization method that iteratively modifies the input of a system so that a desired performance metric reaches a local optimal value. Extremum seeking differs from conventional optimization methods. It performs a non-model-based parameter search, which is independent of whether the system is linear or has significant nonlinearities. The extremum seeking method offers no guarantees for finding a global optimum, but has the ability to simultaneously optimize multiple objectives (like fuel consumption and emissions). Unlike other non-model-based optimization methods, extremum seeking lends itself to convergence analysis, which means that its convergence rate properties can be quantified and tuned.

For the engine application, extremum seeking determines the combustion timing that minimizes the HCCI engine's fuel consumption. Extremum seeking delays the combustion timing from 3 degrees to approximately 9 crank angle degrees after top dead center while simultaneously reducing fuel consumption by more than 10 percent.

Aside from the Caterpillar engine experiments, the LLNL team has had a long involvement with HCCI engines that goes back almost 10 years (see "Otto or Not, Here it Comes," June 2000). Ongoing research at LLNL focuses mainly on developing high-fidelity analysis tools that can assist engine manufacturers in delivering on the high efficiency and low emissions of this promising engine technology. Dependence on foreign oil, local air pollution, and climate change are difficult problems that require innovative solutions. HCCI engines are a practical option that can contribute to a solution.

Daniel Flowers is an engineer and Nick Killingsworth a postdoctoral in the Energy and Environment Directorate at Lawrence Livermore National Laboratory. Robert Dibble is a professor of mechanical engineering at the University of California, Berkeley.

home | features | breaking news | marketplace | departments | about ME back issues | ASME | site search

© 2007 by The American Society of Mechanical Engineers