"The big thing now is getting water-distribution hydraulics modeling more into the operational side," says Walski. "The next wave is, how do we get these great tools for analysis into the hands of the operations people so they can leverage this capability?

"Operating a water system is 95% routine and 5% panic," adds Walski, "but when you have a major pipeline break and suddenly everyone's asking if we can close some valves and isolate the problem, having an analysis tool incorporated [and] in operation so you can play with those kinds of decisions is a huge benefit. And also, there are huge potential energy savings. When I walk around a utility these days, I see potential everywhere, like little nuggets of gold lying around," Walski says.

Out of Sight

Most water customers are only vaguely aware of the systems that, in the best cases, bring clear water reliably from the tap. They may see hints of that supply in the big, silent water tanks on high ground, the blank-walled buildings that crop up in odd places with the hum of pumps coming from within, in the treatment plants on the edge of town and, now and then, the mud-caked pipes exposed when crews rip up streets to perform mysterious chores before bandaging the holes with patches of pavement that gradually sag like scars of old wounds.

Laura Jacobsen manages the planning division for Nevada's Las Vegas Valley Water District, where water is husbanded like gold by a system whose modeling and monitoring capabilities are said to be among the best in the industry. Hydraulic modeling has been much less user-friendly until recently, Jacobsen says. "There were no graphics. You had to enter data manually, and there was no way to visualize results. You were interpreting tables of numbers," she says.

"With a district like ours, with 4,700 miles of pipe, this is an enormously complicated problem," adds Jacobsen. Her district system delivers 437 million gallons per day to 356,000 accounts serving a population of 1.34 million people. There are 41 reservoirs and 52 pumping stations with 256 individual pumping units. "That's a lot of decision variables," she says. "Look at the street grid in your city—many, many blocks. And there is probably a pipeline under every one of them.

"It can be a challenge. You need reliable information from your SCADA system, but the databases are in different formats," Jacobsen adds. Programmers at her utility created their own translation programs to bring SCADA data in automatically. Her utility's software vendor, Innovyze, has come out with a product to automate data import, as has Bentley, although Jacobsen says she has not yet tried it.

GIS started out as a mapping system and a way to track assets, says Brown, but modelers quickly realized the GIS data they needed takes special handling. The interconnections of the pipe are critical for hydraulic analysis, for example. "On a map, you might have pipes crossing at an intersection, but are they connected or do they bypass one another?" says Brown. "What is the diameter, the material and the installation year? The information is not directly used in the model but is indirectly used to establish roughness coefficients. And you need elevation data to calculate pressure, which is typically captured from another source."

Capturing everything is the gold standard. "We have constructed an all-pipes model using [a] 1:1 relationship with our GIS system," says Jacobsen. "All the ID tags for pipes in our models match the numbers in our GIS system. That lets us automate the model construction and update process to a great extent. It also lets us maintain information about the system outside the model, such as maintenance data, which can be referenced by tag number.

"We have integrated with our SCADA system, so we can pull boundary conditions for any time and mimic what's happening in the system in real time. Then, we do a daily operating plan to minimize water age and minimize energy costs, and then we pull the information [back] into the SCADA system, which runs the plan," Jacobsen says.