Statistical analysis and definition of blockages-prediction formulae for the wastewater network of Oslo by evolutionary computing

Abstract

Oslo Vann og Avløpsetaten (Oslo VAV)-the water/wastewater utility in the Norwegian capital city of Oslo-is assessing future strategies for selection of most reliable materials for wastewater networks, taking into account not only material technical performance but also material performance, regarding operational condition of the system.The research project undertaken by SINTEF Group, the largest research organisation in Scandinavia, NTNU (Norges Teknisk-Naturvitenskapelige Universitet) and Oslo VAV adopts several approaches to understand reasons for failures that may impact flow capacity, by analysing historical data for blockages in Oslo.The aim of the study was to understand whether there is a relationship between the performance of the pipeline and a number of specific attributes such as age, material, diameter, to name a few. This paper presents the characteristics of the data set available and discusses the results obtained by performing two different approaches: a traditional statistical analysis by segregating the pipes into classes, each of which with the same explanatory variables, and a Evolutionary Polynomial Regression model (EPR), developed by Technical University of Bari and University of Exeter, to identify possible influence of pipe's attributes on the total amount of predicted blockages in a period of time.Starting from a detailed analysis of the available data for the blockage events, the most important variables are identified and a classification scheme is adopted.From the statistical analysis, it can be stated that age, size and function do seem to have a marked influence on the proneness of a pipeline to blockages, but, for the reduced sample available, it is difficult to say which variable it is more influencing. If we look at total number of blockages the oldest class seems to be the most prone to blockages, but looking at blockage rates (number of blockages per km per year), then it is the youngest class showing the highest blockage rate. EPR allowed identifying the relation between attitude to block and pipe's attributes in order to understand what affects the possibility to have a blockage in the pipe. EPR provides formulae to compute the accumulated number of blockages for a pipe class at the end of a given period of time. Those formulae do not represent simply regression models but highlight those variables which affect the physical phenomenon in question.

Historical analysis of blockages in wastewater pipelines in Oslo and diagnosis of causative pipeline characteristics

“…The analyses presented in this paper and in Ugarelli et al (2009b) concur that sewage pipelines are more prone to blockages than combined flow and stormwater pipelines. This is owing to the fact that the sewage carries with it a host of solid impurities, is denser and more viscous and thus relatively sluggish in comparison to storm-water and mixed flow, with respect to fluidity.…”

“…Comparison with previous studies In Ugarelli et al (2009b) the same dataset analysed in this paper, was analysed with a traditional statistical analysis and with the EPR tool.…”

“…In Ugarelli et al (2009b), formulae derived by means of the Evolutionary Polynomial Regression tool (EPR) (Giustolisi 2006) are provided to compute the cumulative blockages for a pipe at the end of a given period. Those formulae do not represent simple regression models but lay stress on those variables which affect the physical phenomenon in question.…”

A Two‐Stage Data‐Driven Spatiotemporal Analysis to Predict Failure Risk of Urban Sewer Systems Leveraging Machine Learning Algorithms

“…As for the area of data-driven approaches, several studies have used this type of models to analyze the different factors responsible for sewer system failures. For example, Ana et al (2009), Ugarelli, Kristensen, Røstum, Saegrov, and Di Federico (2009), and Younis and Knight (2010) used regression analysis techniques for the selection of important system features that have a strong impact on sewer deterioration. In particular, Ana et al (2009) proposed a backward stepwise regression approach to systematically drop insignificant variables in an iterative manner to improve the predicting capabilities of their model.…”

Sewer asset management – state of the art and research needs

“…Research focused on identifying pipe attributes that are the key variables for the blockage phenomena (e.g. Ugarelli et al 2009), estimating the time intervals between blockages to identify potential trends in the failure rates (e.g. Jin and Mukherjee 2010) and in preventive maintenance scheduling optimization to minimize costs (e.g.…”