ROLE OF RANDOM FACTORS IN NONLINEAR REGRESSION: A CASE STUDY FOR ESTIMATION OF THERMOPHYSICAL PARAMETERS

Abstract eng:
In order to design an optimal experimental setup the designers have to take into account uncertainties connected to the investigated system. The input random factors associated with for example values of loading, specimen dimensions or measurement errors influence behaviour of the system, which thus becomes also uncertain. From this point of view, the experiment design is a very important because it effects amount of information which can be obtained from the experiment. More specifically, accuracy of the identified parameters from indirect experimental measurements depends on experimental settings. In this contribution we demonstrate a role of random factors in a nonlinear model calibration on an illustrative example of one dimensional heat conduction. The thermophysical parameters such as thermal capacity and thermal conductivity are identified on a basis of noisy measurements from experiments with different setup. The experiments vary in a number of sensors and number of observed time steps. The presented statistical analysis shows dependence of the parameter estimation on the choice of measured quantities involving different uncertainties.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket