Modeling state of waste water system in Dakar, Senegal

Abstract: This report contains a method to mathematically determine the condition of pipes in a sewer system in Dakar, Senegal. First a survival analysis model is established based on probability density functions. The functions are then fitted to a data set consisting of the age of the sewer pipes and their inspected status. The status is ranked on a scale from 0 to 5 where 0 is a practically new pipe and 5 being a pipe with severe deterioration. The data fitting is done using the maximum likelihood function. Based on the established functions a prediction of the deterioration of the system is done, modeling how the proportion of pipes in different states develops. It is found that a model describing the deterioration can be established. It is also found that a few hundred inspected pipes is sufficient to model large quantities of similar pipes with good accuracy as long as consistent classification of the pipes is applied.