NRN PhD publishes in renowned journal CMAME
18 Jul 2018
Writing for renowned academic journals is a highly competitive business. With this in mind next month will see NRN AEM funded PhD student, Eilir Pryse, have his work on stochastic systems be published in sought after journal Computer Methods in Applied Mechanics and Engineering (CMAME).
With Eilir's present research, working on NRN project 125 at Swansea University, he aims to address the most pressing issue in the design and analysis of complex stochastic structural dynamic systems, that is, the 'curse of dimensionality'. Eilir and the team want to exploit the huge potential that could be gained by exploring reduced order modelling in the context of explicit consideration of system uncertainties.
Having his work published in a journal which holds the development of computational methods for engineering systems at its core, Eilir’s innovative research has been given the platform to showcase such innovative research – research which is seen to be at the forefront of significant development of computational research in applied mechanics and engineering.
“Work published in a leading journal like this means that it will be read by most of the active researchers in the field. This way, the NRN funded research will have the widest academic impact”
– Professor Sondipon Adhikari, PI, NRN125
So what does this paper have that allows it to stand out from the rest? It 'proposes a set of hybrid projection methods to analyse structural dynamic systems which contain stochastic properties' says Eilir
"The mathematical models used to model physical systems are often idealizations of physical processes. Many engineering problems are concerned with materials that are intrinsically random, therefore, merely using the average value of the material properties would not establish a system’s behaviour with a desired confidence or reliability. Fortunately, the entire subject of uncertainty can itself be addressed in a scientific and a mathematically precise way by utilising stochastic computational models.
Although the process of analysing stochastic computational models is computationally expensive, one of the proposed methods avoids many of the computational overheads. Although the proposed method is computationally inexpensive, it can produce accurate and reliable results."
Being published in such an acclaimed journal in the field of applied mechanics and engineering has allowed this PhD to gain great experience in producing a publication encompassing original scientific thinking which has ultimately been acknowledged by esteemed peers as Eilir explains -
“I am extremely grateful for the support I have received from both the paper’s co-authors (Prof Sondipon Adhikari and Dr Abhishek Kundu) as well as the NRN. From my experience, there are several advantages of writing journal papers whilst studying for a PhD. The process of writing and receiving responses from reviewers has been very rewarding, and I hope that the experience will be advantageous for the future. “
Image Caption: When applying two of the proposed methods to a scholastically parameterised cantilever beam, the following log error plots can be depicted. The depicted methods are similar, however the method used in the RHS plot utilises a error minimising approach. It is apparent that applying the simple error minimising approach can dramatically lower the log of the approximate error.