WDSA Task Committee on distribution system model database

A brief update on the status of the task committee. I received word from Lindell Ormsbee that the ASCE has approved the task committee, and we will plan to discuss this more at EWRI in Portland in the first week of June, 2014.

TASK COMMITTE ON RESEARCH DATABASE FOR WATER DISTRIBUTION SYTEMS
PROPOSAL TO THE WATER DISTRIBUTION SYSTEMS ANALSYS COMMITTEE

Since the late 1960’s, the water distribution system research community has utilized several water distribution networks as baselines for use in comparing different analysis and optimization algorithms. One of the most widely studied water distribution system is the New York Tunnel system, first introduced to the research community in 1969 by Schaake and Lai (1969). Alperovits and Shamir (1977) subsequently introduced a hypothetical two loop system that has been used to evaluate different optimization algorithms (Kessler and Shamir 1989, Geem 2006). Other commonly studied systems include Anytown, USA (Walski, et al. 1987) which was used as the basis for the original “Battle of the Network [Optimization] Models” and the Hanoi, Vietnam network introduced by Fujiwara and Kang (1990). The “Battle of the Water Sensor Networks” utilized the theoretical systems “Network 1” and “Network 2” (Ostfeld, et al. 2008). The hypothetical network of C-Town was used as the case study for the “Battle of the Water Calibration Networks” (Ostfeld, et al. 2011). EPANET’s example networks “Net2” and “Net3” have frequently been used in water quality parameter estimation studies and sensor placement studies (Berry, et al. 2006, Watson, et al. 2010, Pasha and Lansey 2009, and Hart, et al. 2011). Researchers at Texas A&M have produced models of “Micropolis” and “Mesopolis,” two hypothetical water distribution systems (Brumbelow, et al. 2007). Möderl, et al. (2011), developed a system that generates virtual water distribution system models. Additionally, researchers at the University of Exeter have created “Exnet,” a large hypothetical system, for the purpose of testing multi-objective optimization algorithms. Most recently, Jolly et. al., (2013) developed a database of 12 models based on moderate sized systems in Kentucky.

In most cases, these test networks have represented either hypothetical networks or highly skeletonized versions of larger actual systems. Unfortunately, comparison of different algorithms on the basis of a single system may lead to erroneous conclusions, since the possibility exists that any proposed algorithm may be able to take advantage of the unique state space of the particular problem while not being as robust for other systems (Maier, et al. 2003, Khu and Keedwell 2005). Secondly, while comparing the results from a new optimization routine with a previous solution may show a more optimal solution, it might also deviate from a practical or feasible solution (Walski 2001). This kind of comparison also fails to show that the algorithm in question will provide the optimal solution for any configuration rather than just a single, highly skeletonized, example water distribution system. With increasing processing speed of modern computers, researchers are able to analyze much larger and more complex systems than ever before.

The purpose of this proposal is to suggest the establishment of a Task Committee within the WDSA Committee to develop an online database of water distribution systems for use by the water distribution systems analysis committee as well as the larger water distribution research community for the purposes of evaluating and comparing algorithms for use in support of the design and operation of water distribution systems. Expected products will include:

  1. A technical report describing each of the systems as well as instructions for accessing the online database,
  2. a CD which will contain the associated models for each of the systems,
  3. the online database,
  4. technical sessions at the WDSA meeting and EWRI Water Congress that discuss the database, the systems, and example applications of their use in evaluating various algorithms.

Sorry for bringing up an old topic/event but this task committee sounded interesting. What’s the current status of the proposal; are there any plans for an online database?

Hi there,

To answer @patricklabruzzo question, there seems to be an online repository, which seems to be relevant:

http://www.uky.edu/WDST/database.html

However you need to download the files one-by-one. @samhatchett, @eladsal, I suppose we could use them in an OWA repository on Github, right?