‘Do More With Less’ with the Help of Hydraulic Models
Bridging the Gap Between Hydraulic Models
Today, utilities must operate declining infrastructure with shrinking budgets while facing increasing regulatory pressures, energy cost increases, climate change, and many other challenges. Every day they are challenged to “do more with less.” One of the best ways to do so is to maximize return on investment for assets that most utilities already have — namely, their system hydraulic model.
If we could fully understand the performance of a utility system from observed data and past experiences, there would be no need for hydraulic modeling. Hydraulic models interpolate and extrapolate what is known to answer questions for utility decision-makers that the available data cannot address. Countless hydraulic models have been developed over time to support master planning through system capacity assessment. It’s surprising how rarely these models have been used to help with the everyday management of the system.
In the last 20 years, pipeline hydraulic modeling platforms have been extended to integrate a wide range of enterprise systems. Consequently, today’s modeling platforms can, in addition to providing capacity assessment, support other aspects for utility operations – from capital planning to operation optimization. Since most of the utilities today already have a hydraulic model, they can often reap additional benefits from these models with limited additional investments. To obtain these new benefits, one must combine a traditional understanding of the utility’s system with inside knowledge of what modeling tools can (and sometimes cannot) do.
Over 1,000 SCADA to hydraulic model connections were established to calibrate Denver Water’s model under real-time conditions, resulting in a robust decision support tool for water quality analysis.
Potable Water Systems
Hydraulic models of water distribution systems have been in everyday use for over 20 years, primarily to support planning decisions. However, since the earliest days, pressure-pipe models also have been used for other analyses – water quality being one of the early advanced model uses. The Denver Water project is an excellent example of forward thinking, showing how these models can be leveraged over time.
Denver Water provides water to approximately 1.4 million people through over 3,050 miles of pipe. Distribution system operations are complex and variable, making water quality of paramount interest. HDR helped pioneer the use of chloramines as a disinfectant and implemented a robust monitoring and sampling program to ensure high quality. However, during lower-demand and high-temperature periods, there is still a potential for nitrification at the edges of the system.
As commissioned by Denver Water, our team completed an ambitious hydraulic and water quality model calibration project involving over 1,000 permanent SCADA points; dozens of temporary flow, pressure, and valve position measurements; and a system-wide tracer study (Figure 1). The tracer, a National Sanitation Foundation (NSF)-approved calcium chlorine solution, was monitored by over 50 permanent and temporary conductivity probes. Using SCADA data to drive real-time simulation of actual historical conditions in the system, the model was successfully calibrated and used to provide solutions for water quality problems.
By combining model results and over five years of sampling data, a robust water quality management plan was developed including operational, maintenance, and capital improvement recommendations. Upon completion of the project, Myron Nealey, Denver Water Senior Hydraulic Engineer, said: “Combining hydraulic modeling with an innovative method of conducting a tracer study throughout Denver Water’s distribution system helped in analyzing water quality and hydraulic issues. This study gave a variety of practical operational solutions to undertake for areas of Denver Water’s and its distributors’ distribution systems.”
Unlike water quality analyses, risk assessment and system redundancy evaluation is a relatively new application for hydraulic models. It is well documented and relatively straightforward to prioritize capital investment to support growth and expansion, but the prioritization of pipe replacement and redundancy is more difficult. And while pipe condition assessment can forecast the likelihood of failure, the other half of the equation, consequence of failure, is typically not easily determined.
In the case of a confidential client, HDR utilized hydraulic modeling tools to automate the criticality evaluation for every individual pipe and determine the consequence of pipe failure. Over 70 percent of this utility’s demand is comprised of wholesale customers located along the periphery of the system. This aspect of their system configuration concerned the utility if there was a pipe failure.
We used the existing, in-house model to determine pipe redundancy needs by analyzing the consequence of pipe failure. We developed an automated modeling protocol that evaluates the impacts of individual pipe failure on the system performance. Automatically analyzing the results of thousands of model simulations aided in determining Pipe Criticality Ratings and helped optimize the capital investments for their system.
Wastewater Collection Systems
While advanced uses of pressure pipe models have focused primarily on new ways to utilize existing solver results, gravity system models have made major advances expanding the capabilities of solvers. The work done for San Antonio Water System is a great example of how existing models can be extended with new solvers.
SAWS has a long tradition of using hydraulic models to analyze sanitary sewer system performance, identify capacity deficiencies, and determine needed improvements. A key component of any collection model is the rainfall derived inflow and infiltration methodology. There are many methods for modeling RDII, but sometimes traditional approaches can fail.
SAWS’ models were extensively calibrated and validated over time, suggesting there might be additional RDII processes contributing wet weather flow in several basins. Subsequent investigation of the pipe and manhole in the creek floodway indicated that the RDII problem is likely caused by both traditional RDII processes and the direct flow transfer between the creek and the system through defective pipes and manholes. Further testing has shown that manhole lids could have significant inflow rates with only an inch of surcharging. A manhole lid near a creek can become a significant inflow source once the creek level rises.
To evaluate the cross connections between the pipe and river system, we integrated the creek hydraulic model and sanitary sewer hydraulic model (Figure 2). This integrated model allowed the examination of interactions between the creek and sanitary sewer, resulted in better calibration of both models and provided a complete understanding of the existing sanitary collection system and the impacts of the proposed improvements. The integrated model was not only able to better replicate the system behavior during large and/or consecutive rainfall events, but also helped SAWS determine the appropriate corrective measures to eliminate sanitary sewer overflows. A model with traditional RDII approach would result in solutions that would not only allow for more creek water to enter the pipe system, but also aggravate the SSO problems downstream.
While SAWS needed to integrate river and pipe models to fully understand what is going on in the system, the New York City Department of Environmental Protection needed to extend existing collection system models to accurately predict the annual count, volume and duration of combined sewer overflows. To do this, the HDR team conducted extensive DEP model validation based upon intensive flow monitoring at the key CSO regulators in the system. Monitoring was conducted for one year at multiple locations in each regulator to capture the complexity of wet-weather flows in these tidally impacted structures. The flow monitoring included depth and velocity at influent channels and overflow weirs and tide-gate opening angle to record overflow occurrence.
Data from the five-minute Gauge Adjusted Rainfall Radar was applied to the model and long-term simulation results were compared to the observed hydraulic data. The performance of the collection system model was evaluated in several ways. The first assessment evaluated the ability of the model to represent the total service area flow to each of the two pumping stations over the entire 12-month period. A second assessment compared the overall accounting of which rainfall events were fully captured by the combined sewer system, and which caused an overflow. Figure 3 presents a bar chart comparison of the model to observe events for the key four regulators. With the exception of some snowfall or small rainfall events, the model matched the exact CSO occurrence over the 12-month study period. The analysis also compared total CSO discharge volume over the study period, CSO discharge volumes on a storm-by-storm basis and the time-series wet-weather flow coming into and overflowing each regulator during each storm. The result of this detailed model validation was a robust and demonstrably accurate collection system model, suitable to evaluate existing and proposed system performance. In some cases, the model was found to provide higher reliability than flow measurements. As a result, the model can be applied with confidence even without the need for additional flow metering in the validated area.
Stormwater Collection Systems
Similar to wastewater collection systems, stormwater models have seen major advances through improved and integrated solvers. Integrated one-dimensional (1D) and two-dimensional (2D) solvers utilize the computer power of a multiprocessor, including hundreds of processors in a graphics processing unit, to simulate gravity flow both in piped networks and on the overland surface flow. Our modeling experts integrated the results of a combined 1D/2D model with a computational fluid dynamics model of a key weir structure to mitigate the flooding problems (Figure 4) and bring the airfield storm drainage system in compliance at Andrews Force Base, the home of Air Force One. This is one of several airfield inundation projects conducted using 1D/2D model integration.
To ensure the result accuracy meets needed design criteria, a three-dimensional CFD model of the key weir structure was used to determine the head loss across the structure over a range of operating scenarios. Then the CFD results were used to improve the 1D/2D model. The existing 1D model was extended into a combined 1D/2D model to evaluate needed infrastructure improvements and bring the airfield into compliance.
A similar two-step approach is being used to model the City of Cedar Rapids’ stormwater infrastructure and identify system deficiencies, evaluate alternatives to improve system performance and estimate the capital needed for the improvements. The first step focused on a citywide, macro-scale model incorporating the major pipes, open channels, and detention facilities (Figure 5). The results gave the City a broad overview of the performance of major conveyance components and allowed prioritization of succeeding detailed analysis. The model also provides a starting point for more detailed basin-scale models. The basin-scale models include a much more extensive pipe network for targeted smaller basins, and simulate 2D overland flow as well as pipe and channel 1D flow.
The phased nature of the modeling efforts allowed the City to continue large-scale CIP planning while implementing the key projects in prioritized areas. Scott Olson is a Cedar Rapids City Council member who was involved in creating the plan as chairman of the council’s infrastructure committee. He noted in a Cedar Rapids Gazette article: “The needs are substantially greater than the funding we currently have available. So you can see it is more important we have a prioritization system so our limited resources are put to good use, and those projects have an impact.” The two-step modeling approach is the key to guiding this prioritization of projects in Cedar Rapids.
The efficient use of modeling tools makes sense on smaller-scale projects, too. On the South Awbrey Butte Drainage project in Bend, Oregon, we were able to integrate 1D and 2D models to identify bottlenecks in the stormwater system and define potential solutions.
The study area contained many modeling challenges with steep-sloped streets being the principal flow conduits, many areas with no piped conveyance, and drywells as the primary means of stormwater disposal. An integrated modeling approach allowed for more detailed surface water routing and more detailed identification of localized flooding.
Model animations showed floodwater moving along streets and through properties that matched the observed flooding. The integrated 1D/2D model shows the benefits to private property owners and city assets by modeling any corresponding flood depth with infrastructure improvements at a relatively fine resolution. The City of Bend was very interested to see how the model output could help them optimize and compare alternatives.
Modeling expertise goes well beyond the everyday software use and includes intimate knowledge of the hydraulic solvers, software platform, back-end data management, and automation and customizing tools. Our experts maintain proficiency and we hold current licenses for virtually all key commercially available modeling software platforms. HDR employees, including the primary author of Bentley Systems’ data management patent and former Wallingford Software and MWH Soft (now Innovyze) staff members, provide additional return on modeling investment by bridging the gap between the existing models and data and long-term system needs.