Data Needs for Sanitary Sewer Peak Flow Projections

Why is my engineer asking me for so much data? Best practices for flow and rainfall data collection for sanitary collection systems. 

This blog post was adapter to an article in the April 2020 issue of Water Environment and Technology. Here is a link to download a PDF of the article. The data collection practices outlined here are useful for developing antecedent moisture models (AMMs) of collection system. There are many useful resources at the AMM learning library.

Introduction

Owners and operators of sanitary collection systems are often faced with the challenge of determining how much peak flow their systems generate. This might be needed for master planning purposes, or to plan for capital upgrades, or to correct a capacity or overflow issue. Peak flows in these systems are often driven by rainfall, due to inflow and infiltration (I&I) dynamics. This means it is important to understand the relationship between rainfall and flow to answer the peak flow question. All systems are different and exhibit different levels of I&I depending on the system age, condition, material, construction method, and many other factors. Under these circumstances, it is critical to collect good flow and rainfall data to understand the dynamics that are driving flows, so that a good estimate can be made of design peak flows. In this article, I am going to outline some best practices for flow and rainfall data collection, and explain why they are important for developing design peak flows from your sanitary collection system.

Why Collect Good Flow and Rainfall Data?

If you own or operate a sanitary collection system, collecting good data is key for good system operations. But at some point you are probably going to be faced with the question of how much peak flow your system generates. It is prudent to be aware of the data that is needed to answer this question, and to implement a standard practice of collecting this data now so that it is available when you need it. The costs of corrective programs for I&I or system overflows are directly related to the magnitude of the projected design peak flow rates. Poor available data leads to poor projections of design peak flows. This can lead to unnecessary oversizing of upgrades and needless extra expenses. Data collection is very cheap compared to the cost and risk of designing a major upgrade without good data on which to base the design.

Why Models Are Used?

A hydrologic model is often used to determine design peak flows for sanitary collection systems. A hydrologic model is a numerical routine that describes the relationship between rainfall (the system input) and the flow (the system output). These are often referred to as rainfall-runoff models. There are many different hydrologic models available that are suitable for modeling sanitary collection systems. Why do engineers use such models to estimate design peak flows? The purpose of a model is to simulate an unobserved condition, such as the flow from a design rainfall event. Because these events are rare and occur sparsely in nature, it is often necessary to make projections of design flows from observations of flows from smaller storm events. A good model is calibrated and validated to several storm events with various durations, intensities and wetness conditions. Therefore, it is necessary to have good flow and rainfall data to use in developing such a model. The model will only be as good as the data that was available to develop it.

When is a Model Not Necessary?

If the purpose of a model were to simulate unobserved conditions, then a model would not be needed when the desired conditions have been observed and measured. For example, some systems have selected a design storm event based on an actual large event that has occurred and caused problems in the system. That is a perfectly legitimate method to select a design condition. Another example when a model is not necessary is when there is a lot of observed data available. A great example of this is a USGS stream gauge with a long period of record like 50-100 years. With so much data available, it is not necessary to use a hydrologic model to determine design floods such as the 10-year or 25-year flood. The level of these floods can be determined from a statistical probability analysis of the observed data. This is in fact the methodology that FEMA uses to determine floodplains. In the two above examples, it may still be useful to build a model to understand unobserved conditions in the system. For the system that selects a large observed rain event as the design condition, it may still be useful to develop a hydrologic model of the system to understand the distribution of flows across the system at unobserved locations, or to understand the shape of the hydrograph for storage design. In the USGS stream gauge example, a hydraulic model may be useful to understand the water depth at locations along the stream where there is no stream gauge.

Best Practices for Data Collection

The dynamics of I&I in sanitary collection system can occur very quickly, with flows rising after a rainfall within minutes or hours. The large events of interest occur rarely, with useful large events occurring only a few times per year. With these dynamics, the basic framework of a good data collection program incudes the following elements:

1. Continuous measurement – Daily measurements will not cut it for a flow response that can come and go in a matter of minutes or hours. At a minimum, hourly data should be collected. A five-minute sample frequency is even better!
2. Long-term monitoring – Because large storm events rarely occur, and because conditions can change drastically from season to season and from year-to-year, permanent long-term monitoring is necessary. This does not mean that you have to collect data from dozens or hundreds of locations on a permanent basis. But you will increase the accuracy of your peak flow projections tremendously if you have at least one location where you collect long-term data. Often for small and medium communities, this can be at the influent of the wastewater treatment plant. For larger systems, this may mean making measurement on each major branch of the system.
3. Uncensored data – Data that is affected in some way, or censored, by components of the system will not be as useful. Storage facilities, system overflows, on-off pump stations, hydraulic restrictions, and diversions will all censor the data, making the peak flows recorded less useful. If you can, collect flow data at locations that are unaffected by these facilities. For example, place meters upstream of an on/off pump station, or upstream of a diversion to a storage tank or an overflow. If this is not possible, measure the flow of both the interceptor and the diversion. Measuring the flow downstream of an overflow or a diversion will produce data with limited value, as a significant portion of the peak flows for large events will not be measured (censored).
4. Tabulate peak flows on your monthly operating report (MOR) – If you operate a wastewater treatment plant, you probably have an influent flow meter, usually a flume, to measure influent flow rate. Regulations require that the daily average flow be recorded on your MORs. It is a good practice to also record the daily peak flow on your MOR. That is the highest flow that was recorded during the course of the day. For small and medium systems, these data can be invaluable for checking design peak flows from a model against.
5. Collect good rainfall data – Do not overlook the importance of good rainfall data. It is half of the data needed to develop a good flow model. In addition to the best practices above, be sure to collect rainfall data from within or nearby the system of interest. A rain gauge 10 miles away at the airport is not going to cut it, due to spatial variation of large thunderstorm rain events.
6. Data storage and backup – Be sure that the data is accessible and backed up. Collecting all of this great information is not going to do any good if it can’t be accessed in an old legacy system, or if it is lost because it was not backed up. While these data are often collected by SCADA systems, SCADA is not the best repository for long-term storage, because it was designed for short-term control functions. If you do plan to use SCADA as your long-term data repository, be sure to let your integrator or IT staff know that this is an expected use case for the system. Often long-term data storage and accessibility needs are not considered for SCADA systems, and this can lead to data loss when system upgrades are made or a new system is installed. By the way, this is a great reason to use H2Ometrics for long-term data storage!

Closing Comments

Having been involved in sanitary collection system modeling projects for numerous communities around the country, I would highly recommend that owners and operators of these systems implement these best practices for system monitoring. This will ensure that when the time comes to understand design peak flow rates, you have good data on which to base the peak flow projections. This will provide more confidence in the peak flow projections, which allows for better optimized upgrades, lower cost, and reduced risk of failure. For small and medium systems, this could be as simple as placing a continuous, permanent data logger on the rain gauge and influent flow meter at your wastewater treatment plant, or logging peak flows on your MORs. For larger systems, consider placing a collection system flow meter and rain gauge on each major branch of your collection system. A good rule of thumb is one flow meter and rain gauge for each 10,000 – 25,000 people served, depending on the density and number of major system braches. For a community of 100,000 people, this would result in 4-10 flow meters or rain gauges. For a system this large, with an operating budget of tens of millions of dollars and long-term capital improvements that could be on the order of a hundred million dollars, spending a few thousand dollars a month on metering to have great data for developing good design peak flows is a very good investment. This will pay dividends in the long run.

H2Ometrics and Data Collection

H2Ometrics is a cloud-based water and sewer data analytics platform. We collect data from sensors like flow meters, rain gauges and depth monitors, store and host it on our cloud server, and then provide browser-based analytics tools to get value out of the data. H2Ometrics was designed specifically to perform many of the long-term data storage and analytics functions described in this post. To learn more, please see the overview of the platform, our customer case studies, and demo videos that show the tools in action.