Deterioration forecasting of concrete bridge elements in Victoria, Australia using a Markov Chain

Road authorities invest significantly on the planning of maintenance, repair, and rehabilitation of concrete bridge structures. Deterioration caused by service conditions, unanticipated events and deferred maintenance of old bridges are diagnosed using a selected condition monitoring method. Converting these conditions monitoring data to a meaningful practical decision supporting tool is a problem faced by all road authorities worldwide. Whilst there are numerous methods proposed in literature to forecast deterioration using discrete condition data, there is no consensus on the best method or the method appropriate for a given set of data. This paper presents a research where the Markov Process has been used to predict deterioration curves for three different elements of concrete bridges in Victoria, Australia.Data from 5000 bridges and culverts in Victoria has been used in deriving the forecasting curves presented in this paper. Initial analysis of the data indicated that a deterministic method will offer a very poor correlation. Markov chain is a special case of Markov process, which has been used by many researchers to predict the future condition of a deteriorated asset. The data has been collected in four consecutive inspections of structures following a standard bridge inspection manual. In addition to the condition given using a scale of four levels, the data also include the annual average daily traffic, percentage of heavy vehicles and the age of the bridges. Deterioration curves derived for three types of elements are presented in this paper. The Markov chain appears to offer a suitable approach for forecasting deterioration using the set of data analyzed.

[1]  David B. Dunson,et al.  Bayesian Data Analysis , 2010 .

[2]  Jamshid Mohammadi,et al.  Analysis of Bridge Condition Rating Data Using Neural Networks , 1997 .

[3]  Yi Jiang,et al.  BRIDGE PERFORMANCE PREDICTION MODEL USING THE MARKOV CHAIN , 1988 .

[4]  Hojjat Adeli,et al.  Neural Networks in Civil Engineering: 1989–2000 , 2001 .

[5]  Ying-Hua Huang,et al.  Artificial Neural Network Model of Bridge Deterioration , 2010 .

[6]  J. L. Bogdanoff.  A New Cumulative Damage Model: Part 1 , 1978 .

[7]  Stephen P. Brooks,et al.  Markov chain Monte Carlo method and its application , 1998 .

[8]  Samuel H Carpenter,et al.  PAVEMENT PERFORMANCE PREDICTION MODEL USING THE MARKOV PROCESS , 1987 .

[9]  John O Sobanjo.  A Neural Network Approach to Modeling Bridge Deterioration , 1997 .

[10]  David Veshosky,et al.  Comparative Analysis of Bridge Superstructure Deterioration , 1994 .

[11]  E Czepiel.  BRIDGE MANAGEMENT SYSTEMS LITERATURE REVIEW AND SEARCH , 1995 .

[12]  S Setunge,et al.  Deterioration prediction of concrete bridges with artificial neural network (ANN) derived from discrete condition data , 2013 .

[13]  Wayne J. Davis,et al.  Optimal maintenance decisions for pavement management , 1987 .

[14]  Gongkang Fu,et al.  Methodology of Homogeneous and Non-Homogeneous Markov Chains for Modelling Bridge Element Deterioration , 2008 .

[15]  Kamal Golabi,et al.  A Statewide Pavement Management System , 1982 .

[16]  Yi Jiang,et al.  The development of performance prediction and optimization models for bridge management systems , 1990 .

[17]  Teresa M Adams,et al.  FAULT-TREE MODEL OF BRIDGE ELEMENT DETERIORATION DUE TO INTERACTION , 1997 .