Support vector machine applied to the problem of thermal diagnostics of design defects in radio electronic devices

Abstract

The paper describes a new approach to diagnostics of the technical condition of printed circuit board (PCB) assemblies of electronic devices based on mathematical modelling of thermal processes and on the support vector machine as a tool for classifying design defects of electronic equipment. This work continues the previous study of thermal diagnostics of malfunctions in PCB assemblies. The developed method combines mathematical modelling and physical tests of PCB assemblies under study with subsequent analysis of the obtained characteristics. The tools for method implementation include special-purpose software suites for PCB and hardware design such as Altium Designer, SolidWorks, NI Multisim as well as mathematical modelling software packages and the Python programming language.