Complex Method for Determining the Technical Condition of Electronic Devices Based on a Cognitive Model, Petri Nets and Artificial Neural Network

Abstract

The paper proposes a complex method for determining the technical condition of electronic devices (TC of ED) based on building a heterogeneous cognitive model (HCM), a simulation network model and formation of an artificial neural network (ANN). The well-proven and established in modeling discrete processes Petri net acts here as a simulation network model. The novelty of the given method is the combination of the HCM and the Petri net to obtain additional information about TC of ED and to build on their basis the ANN for making diagnostic decisions under measuring and expert information. ANN is used to solve the classification problem that allows one to identify the state of the ED which are characterized by certain parameter values and range it to one of the several pairwise non-intersecting specified classes. The paper presents a table of flowchart conversion into HCM, Petri net, as well as algorithms for converting HCM and Petri net into ANN with some assumptions. This allows one to avoid the select problem of the ANN structure, which is carried out on the basis of the operational personnel experience and scores of attempts to conduct training. The method proposed is illustrated by the example of determining the TC of microcontrollers in control systems.