ABSTRACT
Bridges and Culverts form critical links within transportation networks. They are expensive long term infrastructure investments and are therefore essential to a country’s economic functioning and society’s everyday lives (Tang, 2007). They are prone to deterioration due to unfavourable chemical, mechanical and physical consequences (Scharven, Hartmann & Dewulf, 2011). Nonetheless, they are expected to remain safe and functional for the duration of their design life. Thus bridge management systems have been fundamental to maintaining and preserving these (Nordengen, and Nell, 2005). However, the data they contain is solely used for structural prioritization. Hence, the purpose of the study was to utilize the Struman Bridge Management System database to investigate the relationships that exist between the inventory and inspection data of the RC bridges and RC culverts within the Western Cape Province. The study involved conducting data mining activities to simplify the data and extract useful information. These activities encompassed investigating structural defects, their predominance and spatial distributions (in terms of district municipalities) as well as their relationships with inventory data such as structure type, age and span length/width. STATA logit models were used to investigate whether the relationships were statistically significant and to determine odds ratios of several ‘events’. In addition to this, the average CIs were also investigated to assess the condition of the structures relative to the inventory data. The results indicated that RC bridges and RC culverts had similar defects and consequently similar forms of deterioration. Several deterioration mechanisms and causes of defects were identified. These included chloride ingress, traffic loading, undermining and several others. There were no specific spatial distributions identified along the coast or inland as the prevalence of the predominant defects varied for the different district municipalities. Also, location (in terms of district municipalities) gave more meaningful results when considered with the mean age and the standard deviation of the age of the RC structures in that location. There were no statistically significant relationships identified between the predominant defects and the various structure types. However, findings from data mining suggested that the most dominant structure types had higher percentages of RC structures with defects and were in the worst condition. Moreover, there was a general increase in RC structures with predominant defects with increasing age categories as well as increasing span length/width. Also, the average structural condition showed a general decrease with age and span length/width. Nonetheless, the average CIs of the RC structures suggested that their average overall condition was good. Lastly, the study makes recommendations pertaining to the collection of BMS data, the analyses BMS software should be used to conduct, the parameters CI indicators should take into consideration and further research that would be a continuation of the study, more complicated than previously considered and has important implications for upland water quality.
