Title: Stochastic modelling of failure interaction: Markov model versus discrete event simulation

Authors: Michael Sony, V. Mariappan, Vivek Kamat

Addresses: Department of Electricity, 220 kV S/S Amona Navelim, Goa – 403107, India. ' Roever Engineering College, Anna University, Perambalur, Tamilnadu, 621212, India. ' Goa College of Engineering, Farmagudi Ponda Goa – 403401 India

Abstract: Markov models have traditionally been used to model failure interaction of load sharing parallel system to evaluate the availability of systems. However, complexity of the Markov model increases as the number of components in parallel increases, also solving the Markov model requires high level mathematics and hence practicing engineers find it difficult to model, solve or interpret the results using this. Discrete event simulation (DES) is a more versatile modelling technique that can also be used to model failure interaction. A next event based simulation algorithm is designed and coded in MATLAB 7.0. The Markov and DES model is applied to a live case in Goa Electricity Department and the respective outputs are compared. A full factorial design of experiments is conducted on both the techniques and also statistical model adequacy techniques are conducted to ascertain the veracity of the model. The results indicate that DES model is as accurate as Markov model.

Keywords: next event approach; system simulation; Markov modelling; failure interaction; discrete event simulation; availability; power systems reliability; availability simulation; full factorial design; model limitations; load sharing parallel systems; stochastic modelling; electricity systems.

DOI: 10.1504/IJAOM.2011.040657

International Journal of Advanced Operations Management, 2011 Vol.3 No.1, pp.1 - 18

Published online: 15 Jun 2011 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article