Advanced Research Projects Energy (ARPA-E) Act Proposed in Senate and House
IC Corporation Working with Enova on First Hybrid School Bus

Neural-Network Engine Controller for Higher Efficiency and Lower Emissions

Use of the neural network (NN) controller (right) reduces heat release in lean operation (shown) and with high EGR. Click to enlarge.

Researchers at the University of Missouri-Rolla are developing a neural-network-based engine controller in an attempt to increase engine efficiency while reducing NOx emissions.

Artificial neural networks are adaptive software systems which learn based on the successful connections they make between nodes. The goal of the work is to create a controller that essentially learns on-the-fly how to operate an engine more cleanly and efficiently.

Lean operation on a spark ignition (SI) Engine can reduce emissions (HC, CO and NOx) by as much as 30% and also can improve fuel efficiency by as much as 5%–10%. Engines operating with high EGR (exhaust gas circulation levels) can further reduce emissions by as much as 50% to 60%.

One problem with operating the engine either extremely lean or with high EGR is cyclic dispersion—cycle-to-cycle variability in engine output—of heat release.

A good example of people experiencing cyclic dispersion is when they’re sitting in their car at a stop light and they feel their car shaking. The more EGR you can add, the lower your NOx emissions. The question is how far can we push it and still keep cyclic dispersion in a reasonable range.

—Jim Drallmeier

The goal of the neural-network (NN) controller is to allow operation under both lean and high EGR regimes while minimizing cyclic dispersion heat release.

The neural-network observer part of the controller will assess the total air and fuel in a given cylinder in a given time. It then sends that estimate to another neural network, which generates the fuel commands and tells the engine how much fuel to change each cycle.

—Jagannathan Sarangapani

This controller observes what an engine cycle is doing, makes measurements in that period of time, reduces that data, and decides how you need to push the engine in the next cycle. It does all that before the next cycle starts. We’re talking about a matter of milliseconds.

—Jim Drallmeier

Although increasing EGR can reduce nitrogen oxide emissions, it can cause significant cyclic dispersion—cycle-to-cycle variability in engine output—in heat release.

A smart controller that can reduce cyclic dispersion could open new avenues for engine efficiency. The research team is also exploring the feasibility of the neural network controllers for hybrid vehicles and fuel-cell/gas vehicles as well.

The National Science Foundation and the Environmental Protection Agency are jointly funding the three-year, $515,000 project. The researchers are collaborating with Caterpillar Inc., and Oak Ridge National Laboratories.



The comments to this entry are closed.