Written By:
A routine water model calibration becomes atypical during changing pandemic water demand patterns, revealing the importance of thoughtful data interpretation.
At a time when everyone is working to limit physical contact and maintain safe distances, water distribution system modeling is the perfect project: sensor installation can be performed by a single person, data is collected automatically, and modeling can be completed remotely. WithersRavenel was working on a routine model calibration with Beaufort County and Ferguson at the start of the COVID-19 pandemic in the U.S., providing fascinating insight into how business shutdowns and stay-at-home orders affected both the water industry and residential usage patterns.
Located in the Coastal Plain region of eastern North Carolina, Beaufort County provides water to 13,044 customers spread over a 958-square-mile area. Of those customers, 12,979 are residential, 62 are commercial, and 3 are industrial. This distribution of customers is not unusual, and WithersRavenel expected to see familiar patterns: residential customers use more water in the evenings and on weekends, while commercial and industrial customerswho use more water overallhave their peak times during Monday-Friday working hours.
Beaufort County At A Glance
- 883 Miles of Water Main
- 12 Elevated Storage Tanks
- 2 Water Treatment Plants
- 11 Booster Pump Stations
- 2 Water Treatment Plant Pump Stations
- 2 Valve Stations
- 6 Wells
- 843 Hydrants
- 7,820 Valves
WithersRavenel just completed validating the water model and allocating loads when Ferguson delivered the data from the Automated Meter Reading System. We converted that data to a histogram, which broke users into categories based on overall usage. Surprisingly, more than 1,800 of the residential customers used less than 250 gallons per day on average. A typical family of four uses closer to 320-400 gallons per day on average, and while it is not impossible for smaller or more water-conscious households to use less than that amount, it is rare to see so many of them.
After rechecking the numbers, we concluded the math itself was correct, which pointed to faulty assumptions in the analysis. We dug deeper into the data with Ferguson and discovered that many of these homes are actually second homes and not primary residences. As a result, many of them are used only on weekends. This increased usage by part-time residents should have caused a noticeable rise in overall usage and specifically a spike weekend usage, but it was offset by industrial shutdowns. The net effect was little change in overall volumes, but a distinctive shift in demand patterns.
Taking the time to review the hourly data with Ferguson caused us to revisit and revise our assumptions about Beaufort’s system. Going forward, this insight will enable our team to offer more accurate model explanations, such as recognizing sudden water pressure drops as a consequence of part-time resident usage and not necessarily the result of a leak as previously thought. It will also help the County select times for system flushing operations that do not disrupt system pressures.
Quality models and the inherent data analytics that accompany the modeling systems are critically important tools for managing our state’s water systems. Water modeling not only improves water system performance and reliability, but also enables utilities to become more efficient in its day-to-day operations and can provide for substantial cost savings when used effectively. At WithersRavenel our team of water, financial, and data-mining strategists can help you more efficiently and effectively manage your water utility operations.