Neural network analysis of impact properties of flux cored arc weld metals for structural steels

ABSTRACT

Today more and more systems use standardised hardware and software which are customised for particular applications using data. These data-driven systems offer flexibility and speed of implementation, but are dependent heavily on the correctness of their data to ensure safe operation. While data is crucial to the correct operation of these data-driven systems, there is much evidence that it is often overlooked. In many cases the data is developed quite separately from the remainder of such the systems, and may not benefit from the same level of hazard analysis, verification and validation.

This paper considers the characteristics of data in data-driven systems, and suggests that in such systems it is appropriate to consider data as a separate component with its development lifecycle. The paper then considers the architectural design of data-driven systems, and outlines the data supply process before considering the verification and validation of such systems.

KEYWORDS: Data-Driven Systems, Verification and Validation, Safety Management, Data Development.

Dr Yih-Chang Chen is an Assistant Professor of the Department of Information Management, Chang Jung University, Tainan, Taiwan. (cheny@mail.cjcu.edu.tw ) Dr N. Story is a Lecturer of Electronic Engineering in the University of Warwick, Coventry, United Kingdom.