Conference on Applied Inverse Problems, July 20-24, 2009, Vienna, Austria
Poster Presentation
Marcin Janicki: Modification of Function Specification Algorithm for Real Time Estimation of Source Temperature in Electronic Circuits

Tue, 21 July, 2009, 17:15-18:15, Foyer

Temperature is an important factor affecting both electronic circuit operation and its reliability. Therefore, certain applications require continuous temperature monitoring during their operation so that to protect circuits from thermal destruction. Obviously, the best solution in such cases is to place temperature sensors directly where heat is generated. However, quite often this is not allowed due to certain design constraints. Then, circuit temperature has to be monitored using remote sensors located away from heat sources and the heat source temperature has to be estimated using some inverse method. Owing to its simplicity and ease of practical digital filter realisation, the best suited from the inverse algorithms for this particular purpose is the Beck's function specification method.
However, this algorithm has also some drawbacks. Namely, due to the fact that the exact time instants when changes in the power dissipation occur are not known in the considered application, the function specification method cannot be applied in its classical form and the estimation has to be carried out employing steady state solutions of the heat equation. This however, together with the time averaging of several consecutive samples, leads to significant estimation errors just after each change in the power dissipation pattern. Thus, a modification to the standard function specification algorithm is proposed, where measured temperature values from sensors are delayed according to the distance between sources and particular sensors. The benefits from the proposed modification are illustrated based on the simulated and measured data for the real test integrated circuit containing matrices of heat sources and temperature sensors. For this structure the heat equation was solved using the 3D Green's function method.
This research has been supported by the Polish Ministry of Science and Higher Education grant No. N515 008 31/0331.

< Back | ^ Top


This page was made with 100% valid HTML & CSS - Send comments to Webmaster
Today's date and time is 10/23/21 - 04:36 CEST and this file (/events/conferences/aip2009/poster/talk.php) was last modified on 12/20/12 - 17:06 CEST