Underground Mysteries Solved

Hydraulic modeling helps a Georgia city make water system improvements and plan for future growth in the aftermath of a tornado
Underground Mysteries  Solved
Jennifer Suttles, project manager at Woodard & Curran, uses hydraulic modeling software to examine Americus’ water infrastructure. (Photos and graphics courtesy of Hugh Ryan)

Interested in Infrastructure?

Get Infrastructure articles, news and videos right in your inbox! Sign up now.

Infrastructure + Get Alerts

When a tornado nearly leveled sections of Americus, Ga., the city got down to the business of rebuilding. One major reconstruction project was the replacement of the destroyed Sumter Regional Hospital, the area’s major health care facility.

To serve the new facility and provide for more growth and development, the city needed to ensure adequate pressure throughout its water distribution system. Some of the infrastructure is over 100 years old. City officials knew some areas of town had much different pressure than would be expected based on maps of the system.

Addressing these issues required a thorough understanding of the whole system, and the best way to develop that was through a systemwide hydraulic model. An initial data collection phase provided information for the model, which accounted for known water system attributes such as pipe sizes and materials, valve locations, fire hydrants, and elevated water tank levels and sizes.

The model identified areas of low pressure and inadequate fire flow, but that was far from the only benefit. It also helped the city find infrastructure it didn’t know it had, and identified closed valves in the system, solving mysteries about low pressure. Finally, the model highlighted areas where opening or closing valves would improve pressure, and where new or upgraded pipes would be needed to accommodate future demand.

 

Incomplete information

Americus, known as the “Shining City on a Hill,” is located in Sumter County in southwest Georgia. It is home to the headquarters of Habitat for Humanity International, the century-old Windsor Hotel and the Rosalynn Carter Institute for Caregiving. It is also a popular tourist destination, being near the birthplace of former president Jimmy Carter and the site of the Civil War’s infamous Andersonville prison.

Out of sight from the wandering tourist, beneath the city’s roadways and historic buildings, lie old water pipes, some dating back to the early 20th century. As challenging as it is to run an aging water system, it grows much more difficult when information on the infrastructure is outdated, inaccurate or missing. The lack of information is accentuated when water mains break and it is nearly impossible to fully isolate the system for repairs.

 

Building the model

The Americus water distribution system consists of 134 miles of 2- to 20-inch pipe spread over nearly 11 square miles. To serve its 17,000 people, the city has two water treatment plants fed by groundwater and eight elevated storage tanks. The city now supplies an average water demand of 2.85 mgd, but future demand has been estimated at 3.34 mgd by 2030, a 17 percent increase.

To develop the hydraulic model, Americus hired consultants Woodard & Curran to examine infrastructure maps and information, including Geographical Information System (GIS), record drawings and water maps.

Engineers entered the data into the Bentley WaterGEMS hydraulic modeling software to create an initial infrastructure schematic for the entire city. They extracted residential, commercial, industrial and institutional water consumption data from the city’s utility billing system, and then accurately located the data within the model to represent varying daily and seasonal demands on the waterlines at any given location.

As a better picture of the system emerged, officials chose nine strategic locations for fire hydrant tests that calibrate the model and replicate how the system worked in reality. During a fire flow test, flows from an open hydrant are recorded, along with initial (static) and final (residual) pressures on neighboring hydrants, to document the system’s reactions to high flows.

These tests aid in refining pipe roughness values, a factor that contributes to pressure. Well-calibrated roughness values result in a model that accurately represents field conditions. In addition to fire flow testing, engineers installed five pressure data loggers throughout the system to record pressure over a 24-hour period.

The pressure loggers captured the behavior of the system at specific points before, during and after the fire flow tests. The data gave the team a better idea how the system reacted to the fire flow tests and provided information on how the system performed over time, from daytime high demands to nighttime low demands.

Once preliminary data was entered into the hydraulic modeling software, along with elevated storage tank data from the city’s SCADA system, the initial typical values assumed for pipe roughness (originally based on material) were refined based on the results of the fire hydrant tests.

In many cases, because of the age of the system, the roughness coefficients were lower than typical for the pipe material. (The lower the coefficient, the rougher the pipes, and the more potential pressure loss.) Newer pipes, such as large transmission lines installed more recently, had higher roughness coefficients.

 

Discovering the system

During calibration, the team discovered several anomalies. First, the modeled elevated storage tanks did not accurately represent the filling/emptying cycling seen in reality. In addition, field measurements showed extremely low pressures in some sections of the distribution system where the model indicated pressures should be much higher.

Finally, parts of the system thought to be closed off were still under pressure, and vice versa. Behaviors that did not represent how the system should act based on known attributes of the infrastructure indicated the presence of pipes and valves that were not in the city’s water map or GIS.

To find the “missing” infrastructure, team members changed the model to try to explain why the system was behaving differently than expected. In some cases, the modification required inserting a partially or even fully closed valve in areas where real pressures were lower than modeled.

Insertion of these valves made the model more accurately represent the physical system, and when technicians performed field checks of those areas, they indeed found valves. In one case, simply opening a closed valve provided adequate pressures to an area that had experienced zero pressure during fire flow testing.

To resolve another anomaly, the team added a pipe to the model where the city thought an old abandoned line was located, based on a 1911 map. The model showed that the pipe was, in fact, still in service, and that explained why the city had difficulty isolating parts of the system.

 

Benefits of knowing

One immediate benefit of the modeling was a better understanding of how to isolate specific areas of the system without affecting other customers throughout the city. In addition, identifying potentially closed valves allowed the city to locate and open them to increase pressure and fire flows to homes and businesses experiencing lower pressures.

In areas where inadequately sized infrastructure was responsible for chronic low pressures and flows, the model supported recommendations for new pipes or pipe upsizing to improve water availability.

For example, the model showed that additional piping would be needed to ensure that developments like the new hospital would receive adequate daily pressures and fire flows during emergencies. The new line would also provide redundancy in the system in case the existing line went out of service.

Team members scrutinized the recommendations for impacts to the city’s historical areas and worked to minimize them wherever possible.

Another benefit of the model was the determination that one of the city’s eight elevated storage tanks could be taken out of service, saving on maintenance costs. The tank was temporarily out of service for maintenance during the fire hydrant testing, and the model showed that the system could meet both existing and projected demands with the remaining seven tanks.

 

A learning process

Since the final hydraulic model report was issued, the city has continued to collect data and update the model. Tank elevations have been surveyed where estimates had been used previously. The city hired a company to exercise valves, giving a much better idea of their working condition and whether they were open or closed. The city continues to incorporate this information into the model to maintain as accurate a system as possible for emergencies and future growth.

The hydraulic model, with collected field data, helped the city conclusively identify unknown or forgotten infrastructure, from missing pipes to closed valves. It helped the city plan for growth and provided greater certainty in day-to-day operations, leading to better service to customers without unnecessary capital expense.

The city showed foresight in developing the hydraulic model and making a small investment to improve it continuously for the future. F

 

About the Authors

Bernard Kendrick (bernard.kendrick@cityofamericus.net) is public works director in Americus, Ga. Jennifer Suttles (jsuttles@woodardcurran.com) is a project manager and water resources expert with Woodard & Curran. Lorraine Campagne (lcampagne @woodardcurran.com) is a water resources engineer with Woodard & Curran.



Discussion

Comments on this site are submitted by users and are not endorsed by nor do they reflect the views or opinions of COLE Publishing, Inc. Comments are moderated before being posted.