Dynamometer tests of the new sensor and its FEA thermal results correlated to within 4%.

Dynamometer tests of the new sensor and its FEA thermal results correlated to within 4%.


A relatively simple 2D thermal model let Delphi Corp., Troy, Mich. (www.delphi.com), find a more durable automotive oxygen sensor. The new design senses wider fuel-air ratios than existing sensors. This, in turn, can lead to higher fuel efficiency and lower exhaust emissions. Delphi thermal models were run in FEA software from Algor Inc., Pittsburgh (www.algor.com).

Existing sensors only signal a rich (excess fuel) or lean (excess oxygen) ratio. "The new sensor has to detect ratios in-between and work from 40 to over 1,000°C, making thermal optimization a trick," says C. Scott Nelson, Delphi senior project engineer.

Durability, low cost, and a short length were also design goals. Lengthening the sensor is an easy way to lower temperatures, but longer parts vibrate more.

Design strategies were to restrict vertical heat flow while promoting radial flow and conduction through components. Nelson experimented with different thermal-conductivity values without a particular material in mind and then looked for materials with similar properties. "The software let me reduce the temperature at two critical locations by 20% which kept peak temperatures below the material's maximum threshold," he says.

— Paul Dvorak

Delphi's C. Scott Nelson used Algor FEA software to lower the working temperature of a wide-range oxygen sensor. The one-half-symmetric model shows a few of the sensor materials. The simple model let Nelson do many more iterations than would be feasible through manual analysis.
Delphi's C. Scott Nelson used Algor FEA software to lower the working temperature of a wide-range oxygen sensor. The one-half-symmetric model shows a few of the sensor materials. The simple model let Nelson do many more iterations than would be feasible through manual analysis.