Inverse Thermal Problems in Electronics

Recent technological advances have led to miniaturization of electronic devices and increase of their operating frequency. Both these factors have augmented dissipated power density and consequently have brought new challenges in the thermal management of electronic systems. Since electronic circuit failures are caused mainly by elevated temperatures, thermal analyses have become indispensable in electronic circuit design. The main objective of the project is to elaborate effective methods for solving various inverse problems occurring during the thermal analyses of electronic systems. Owing to the developed solutions, it will be possible to design thermally optimized electronic systems incorporating built-in components performing real time temperature monitoring for control or protection purposes. The first issue concerned by the project activities will be the optimization of electronic system thermal models, both the three-dimensional distributed ones and the compact boundary independent ones. The distributed three-dimensional models are extremely important for all the thermal analyses of electronic systems, because they are practically the only ones, which render possible fairly accurate simulations of even the most complex cases. The optimization in the case of distributed models will consist in the identification of various model parameters such as temperature dependent material thermophysical properties or structure geometry. The heat equation solutions will be obtained using numerical methods, and when possible the analytical one based on Green`s functions. The compact boundary independent models consisting of several elements are created by numerous simulations with different boundary sets and subsequent reduction so as to minimize the error with respect to measurements and the distributed model simulations. The compact models will be derived in two ways, the above described standard one, and the new one based on transient temperature measurements. Owing to the proposed approach it will be possible to reduce the entire compact model generation process assigning some physical meaning to its components and to derive at the same time the greens functions necessary for the analytical solutions. The second project goal is to establish, having created the optimised thermal models, efficient procedures for the optimal positioning of heat sources in electronic circuit layout and the determination of the best, from the temperature estimation point of view, location of temperature sensors for a given heat source configuration. Finally, the project will deal with the problem of real time estimation of power dissipated in the heat sources from remote sensor measurements. The developed algorithms will be based on the previously elaborated thermal models and the optimal sensor configurations. All the created algorithms will be developed specifically for their direct practical application in the design of electronic systems and real time monitoring of circuit temperature. The main goal of the research on the monitoring system will be to identify not only the best spatial configuration for a given problem but also the optimal sensor readout time and the value of the inverse algorithm regularizing parameters as well as its stopping criteria.