Peter Chalk Centre

University of Exeter

Stocker Road

Exeter

EX4 4QD

Tel: +44 (0)1392 263637

E-mail: CCWI2019@exeter.ac.uk 

17th International Computing & Control for the Water Industry Conference

1st - 4th September 2019
University of Exeter, UK
  • Facebook Social Icon
  • Twitter Social Icon
  • Instagram

9C Asset management

Please note the time shown on this page is automatically adjusted by the server according to the time zone set in your computer.

Collaborative

Ruth Allen

Chair:

to

Development of Long-term Renewal Plan Optimization Model for Suspension Risk Mitigation of Water Distribution Network

Kibum Kim

Presenter:

Authors:

Kibum Kim, Jeewon Seo, Jinseok Hyung, Taehyeon Kim, Taeho Choi and Jayong Koo

This research developed a renewal plan optimization model for enabling water distribution network renewal at minimal cost within water supplier’s budget while minimizing water suspension risk. The renewal plan optimization model is developed to use multi-objective dynamic programming in consideration of life cycle cost of water distribution network and water suspension risk in water distribution network. A developed model, which pursues the simultaneous minimization of life cycle cost and water suspension risk, was applied to Y city. Compared to the conventional methodology, the optimal renewal plan could further reduce annual average water suspension risk by 28.38 % at a cost reduced by 12.41 %. Thus the feasibility of the developed model could be verified by demonstrating that further mitigation of water suspension risk was possible at lower cost despite of the small annual deviation of budget due to restraint on allowable annual budget. In addition, an analysis showed that a renewal plan with the application of the optimal water suspension risk mitigation method could reduce the average annual water suspension risk to 10.67 m3/yr at a life cycle cost of 62,247 million KRW. This means that the renewal plan enables the management of water suspension risk at a reduction level of 41.94 % compared to the conventional methodology, even at a cost reduced by 8.11 %, thereby reducing the estimated length of water suspension time by 2.62 times from the current 2.28 days (about 55 hours) to 0.87 days (about 21 hours). This research developed a renewal plan optimization model that considers the stability of water supply, about which the preceding researches had little concern. As the developed model has been proved to be capable of giving better achievement at lower cost compared to the conventional methodology, its utilization is expected to enable more effective maintenance of water supply facilities.

to

Presentation moved to session 4A

Presenter:

Authors:

to

Predicting future failure rates of pipe networks using a combination of physical degradation modelling and statistical inference

Presenter:

Authors:

Jojanneke van Vossen, Karel van Laarhoven and Roel Diemel

Most drinking water companies quantify the performance of the pipe network in terms of number of failures. Being able to predict the number of future failure rates is vital for estimating the amount of investments required for replacements. Many existing methods consist of statistical extrapolation of actual pipe failure data. Common challenges with the data are that these are collected from relatively short and incomplete time windows with respect to the lifespan of pipes. This makes it difficult to choose a suitable statistical model for extrapolating failure rates based on the data alone. A solution is to use physical models on pipe degradation. However, these in general require many local input data, making it difficult to use them on a network scale. We investigate a method in which physical inferences about degradation are used to choose a model for statistical fit of failure data and use that model for prediction of future failure rates. We present results of the use of the Comsima tool for calculating stresses in pipes in combination with failure data from a pilot area to assess the possibilities and limitations of this approach.

to

Water Tank Crack Detection and Segmentation using Deep Convolution Neural Network and 3D Reality Mesh Model

Zheng Yi Wu

Presenter:

Authors:

Zheng Yi Wu and Rony Kalfarisi

Water tanks are often used to store water at the elevated location for maintaining the required pressure and adequate water service. The tanks need to be inspected to ensure its integrity. However, it is not easy to inspect the elevated tanks. In addition, it is conventionally visual inspection but subjective and time consuming. In this paper, images are collected by Unmanned Aerial Vehicle (UAV), a deep convolution neural network is trained to automatically detect and segment the cracks. The images are used to construct 3D reality model by photogrammetry technology. The cracks are annotated on 3D mesh for intuitive presentation and statistical assessment.

to

The impact of planned Level of Service on investment selection using the integrated water resources and investment planning model

Damian Staszek

Presenter:

Authors:

Damian Staszek, Savic Dragan and Guangtao Fu

Traditionally water companies use separate models for water resource planning and for water resource investment planning. They estimate their deployable output (DO), defined as the output of a commissioned source or group of sources or bulk supply, in a water resources model. The DO estimation is performed for an assumed level of service, under chosen drought scenarios. The output of water sources can be constrained by the environment, licences, pumping plants and/or well/aquifer properties, waters transfer or water quality issues - and these constraints have to be taken into account in the DO calculation. The DO estimation is then used as an input to an investment model. There is a lack of two-way interaction between the water resources model and the investment model in the traditional approach. Any investment selected by the investment model does not affect the initially-estimated deployable output. We propose an integrated model for water resources and investment planning, where any investment selected affects the forecasted deployable output. A case study of Bristol Water is presented in this paper, and results from Bristol Water’s draft WRMP 2019 are compared with results achieved from the integrated optimisation model. We investigate the effect of various assumed LoS on investment selections. Methods and Materials

to