Towards assessing sewer performance and serviceability using knowledge based systems

To sustain the functionality of an ageing and steadily deteriorating sewerage network more emphasis must be placed on its effective maintenance. To determine the point at which proactive intervention in the system is necessary the relationship between sewer condition and performance needs to be understood and robustly defined. The key to this analysis is the ability to define the rate at which assets deteriorate and how this may be influenced by external factors. This paper describes a methodology that aims to produce a performance assessment model by applying techniques of expert and knowledge based systems, which have been used for failure analysis in other branches of engineering. Existing data from sources such as repeat CCTV surveys, event histories and hydraulic models can be used to identify changes in the internal condition of sewers over time. Pipes are categorised into either "no change" or "rapid change" groups and related to i) their physical attributes/susceptibility to surcharging/local environment and ii) to their past performance, to determine the influence deteriorating condition has had on performance. This will allow a model database to be collated of pipes in many different circumstances, with known performance histories. From this a series of cases can be built up forming a suitable archive from which to predict suitable management strategies of any pipe of known characteristics, using techniques such as case based reasoning (CBR) and classification by similarity (CBS). The paper also evaluates the current criteria for targeting maintenance activity in the UK and considers the extent to which existing " rules" are capable of identifying performance deficiencies in a sewerage network (in contrast to condition deficiencies). Suitable database structures and appropriate case retrieval functions/algorithms are discussed as a first step in the development of a tool for prioritising the asset management of sewers with a high probability of rapid deterioration and impaired future performance