Home

Purpose

The purpose of this paper is to describe a machine learning (ML) approach that was implemented to optimize wet weather operations at the Neuse River Resource Recovery Facility (NRRRF). This was a collaborate effort between Hazen and Raleigh Water to investigate if a ML approach could be used to predict influent flow to the NRRRF 72-hours in advance as a function of explanatory variables with an emphasis on being able to accurately forecast the largest events. Benefits The project objectives were achieved, and this presentation will demonstrate that many facilities can follow this approach to predict the shape of their influent flow hydrograph and implement their wet weather strategy accordingly in real-time. This paper will also include recommendations for future projects such as including ongoing model retraining to more accurately and quickly incorporate changes in collection system behavior into your model.

Introduction
Raleigh Water owns and operates the NRRRF, which treats an average daily flow of 48 million gallons per day (mgd), is permitted to treat 75 mgd, and can convey a hydraulic peak hour flow of 225 mgd. In the past few years the NRRRF has experienced extended 24-hour sustained flows around 150 mgd and peak hour flows around 184 mgd during wet weather. The NRRRF is subject to a stringent total nitrogen (TN) load allocation established under the Neuse River Basin Nutrient Management Strategy that requires the Neuse River RRF to achieve an annual average TN concentration of less than 3 mg/L at permitted flow. In addition, NRRRF must also meet a quarterly average effluent total phosphorus limit of 2 mg/L. The biological nutrient removal process's performance is typically inversely correlated to the amount of flow and load entering the NRRRF, hence the ability to minimize the peak flow through the facility would improve effluent quality and process reliability. The NRRRF's 32-million-gallon primary effluent equalization basin was designed to withhold a significant portion of the flow and load entering the NRRRF during high flow events. While the equalization volume is substantial and offers a useful tool for operators, its utility could be optimized by knowing how much flow will enter the NRRRF for a given rainfall event over the duration of the event and how high the peak flow will be. However, even if a facility does not have dedicated equalization basins such a predictive tool can still be beneficial as other wet weather strategies can be planned for and implemented. Prior to this project, NRRRF Staff utilized collection system pump station data to estimate when the peak flow of the storm will reach the NRRRF and its magnitude, which provided about 30-60 minutes of advance warning. Raleigh Water also has a collection system model and collection system flow monitors. The collection system model is useful planning tool, but there is currently not a way to use this tool in a real-time fashion. ML uses algorithms that assign weights to independent variables and seeks to minimize error when predicting a dependent variable. The open source software package Python was used in this evaluation. ML is an alternative to traditional mechanistic models. The project outcome described herein (predicting influent flow 72 hours in advance) could likely be achieved using a well-calibrated collection system model and forecasted rainfall data as well. Some of the advantages of the ML approach are that it utilizes non-proprietary software, can be deployed to provide continuous retraining, and the results can be viewed in customized format that integrates with existing tools such as the NRRRF's secondary clarifier guidance program. This is not to say that ML models are superior or should replace traditional scientific approaches.

Results
A well-calibrated model predicting influent flows to the NRRRF was developed using a ML-based approach. A ML model was trained to over six years of hourly influent flow data to predict influent flow using the following explanatory variables: the past 12 hours of influent flow, streamflow data, and rainfall data. The model utilizes hourly rainfall forecasts and real-time streamflow data in its predictive algorithm. It is worth noting that collection system monitoring data was also considered but was unavailable for the same time period as the rest of the data, but this information ought to be considered if available. Models were developed to predict flow at each time step from 1 to 72 hours into the future. Thirty-eight storms in 6 years of training data were accurately predicted. The predictions are displayed in a web-based Microsoft Power BI dashboard tool that includes a tool to estimate the optimal point to fill the equalization basin to maximize its utility (see Figures 1 and 2). Microsoft Azure was used to develop an automated data pipeline to update model predictions each hour. The project was deployed in a test mode in December 2019 and completed in July 2020. Since then, at least two major storm events including Hurricane Isaias have occurred and been well predicted (see Figures 3 and 4). This tool also integrates with the secondary clarifier guidance program that calculates required clarifier surface area as a function of SVI, influent flow, RAS flow, and mixed liquor concentration. Per the original design the equalization basins were provided to limit the maximum day flow to the secondary system to two times the influent design flow, or 150 mgd. Use of the equalization tank during high flow events is important for maintaining sustained flows to the secondary process below the capacity of the clarifiers. This is also why the 24-hour running average flow is displayed on the model prediction screen.

Conclusion
Machine learning tools have a place in the water industry and are likely to become more commonplace in the next decade.

Introduction
Rainfall-runoff dynamics of surface water, combined sewer, and separate sewer systems can be highly impacted by antecedent moisture conditions, or the relative wetness or dryness of the system. Accurately simulating these dynamics is critical for developing predictive models of systems that are sensitive to antecedent moisture.

Rainfall-runoff systems are prevalent throughout the natural and built environment, and include surface runoff, stormwater systems, combined sewer systems, and separate sanitary sewer systems. Modeling of these systems is performed for applications like flood control, combined sewer overflow control, and separate sanitary sewer overflow control. Billions of dollars in capital improvements are designed each year based on the outcomes and accuracy of models. The purpose of a model is to simulate unobserved conditions from a mathematical description of the system based on past system performance. In that respect, for a model to be useful, it must be capable of making accurate predictions of future events.

Overview
This paper presents applications for modeling rainfall runoff responses with an antecedent moisture model. The model was derived using the principles of system identification from the field of aerospace control systems to find the simplest mathematical model that accurately describes the relationship between system inputs and the flow output. Developing and testing the model was done primarily from observations in the Midwest U.S. where both preceding rainfall and seasonal hydrologic conditions impact antecedent moisture dynamics. For these systems, the model described here is perhaps the most parsimonious that can accurately simulate these dynamics. This provides several advantages to the modeler, including ease of use, fewer parameters to calibrate, ability to quickly identify optimal parameters, and ease to represent in a numerical computer routine. Physical interpretation of the model structure and parameters is possible, providing the modeler with useful insights into the physical processes driving the rainfall-runoff dynamics. The paper contains the following major components: - An overview of antecedent moisture effects on rainfall-runoff systems.

Details of the model, its development, and the equations, including a description of system identification and the parsimony principle. - Commentary on the use, applications and physical interpretation of the model. - Processes for application of the model including calibration, validation, continuous simulation, frequency analysis and design. The model was initially developed between 1995 and 2000. It has been updated and applied to hundreds of systems and presented in several papers over the years. Until recently, the details of the model, including the equations have been held as a trade secret. The equations and model process were recently released into the public domain. The publication of this paper will be accompanied by a series of spreadsheets that show the model computations and a series of tutorial videos that show how the model works.

Model Development Antecedent moisture conditions, or the relative wetness or dryness of a system, can have a tremendous impact on rainfall-runoff dynamics. The magnitude of the runoff (flow) from rainfall can be affected by how wet the drainage area is from prior conditions. Wetter conditions can produce more runoff, and drier conditions can produce less runoff. Wetness conditions can be affected by a multitude of hydrologic conditions that include items such as prior rainfall, depression storage, air temperature, evaporation, evapotranspiration, solar radiation, soil types, and many other factors. Hydrologic systems can exhibit a wide variation in their response to antecedent moisture conditions. The impact could be as small as initial depression storage on an impervious surface that only affects runoff by a small percentage, or it could be as large as varying wetness conditions in a separate sanitary sewer system that change inflow and infiltration volumes by an order of magnate or more. Understanding the relative impact of antecedent moisture on these systems is critical for engineering design and system operations. These effects can be critically important for developing accurate predictive models of systems that are sensitive to these effects. The importance of accurately accounting for antecedent moisture effects has been covered extensively in the literature. Simulating the rainfall-runoff dynamics of such systems requires the use of an accurate continuous model that is developed to simulate many storms that may occur over a wide range of antecedent moisture conditions. Compared to single event simulation, a continuous simulation can more correctly represent antecedent conditions by incorporating processes of both dry weather periods and wet weather periods. The characteristics of the system identification approach and parsimony are well suited for developing models to simulate rainfall runoff responses and antecedent moisture effects. These principles have been applied to develop the model for simulating the rainfall runoff responses and antecedent moisture effects described in this paper. The development, process and equations of the model will be reviewed

Applications
The paper will cover applications of the model for system design, including:
- Spreadsheet companion to the equations.
- Data requirements and observation data content for developing a model.
- Diurnal flow filtering from total flow signal in sewer applications.
- Calibration and validation processes.
- Model performance quantification through a rigorous accuracy of fit process
- Long-term continuous simulation and frequency analysis for system design.
- System benchmarking under identically simulated antecedent moisture conditions using the model.

INTRODUCTION AND PURPOSE The Great Lakes Water Authority (GLWA) is Michigan's largest regional wastewater collection and treatment system encompassing eighty-six wholesale customer collection systems across three counties and 944 square miles in southeast Michigan. The regional collection system serves approximately three million people, representing one third of the state's population. GLWA operates the regional collection system and Water Resource Recovery Facility (WRRF) through a lease from the City of Detroit. The leased facilities include 183 miles of trunk sewer, interceptors, and outfalls; 6 pumping stations, 16 in-system storage devices, 8 retention treatment basins and screening and disinfection facilities, the WRRF and associated metering and control facilities. The City of Detroit Water and Sewerage Department (DWSD) owns and maintains the wastewater collection system and water distribution system within the City of Detroit. GLWA Member collection systems include 77 untreated CSOs, several points of SSO during large weather events, 12 CSO retention and treatment facilities, and 14 sanitary retention basins, and an estimated 3,000 MS4 separate stormwater outfalls. A high degree of structural optimization has already been achieved through local water quality improvement initiatives, NPDES permit compliance, and operating protocols developed over more than 30 years of combined sewer overflow (CSO) control. The Regional Operating Plan (ROP) is intended to further optimize operations by analyzing the detailed system response to storm events, including expanded sharing of real time SCADA data. This presentation describes the development and integration of GLWA's regional collection system model with a GIS-based digital twin platform that integrates near real time data from SCADA, radar rainfall, and overflow reports with the collection system model. The digital twin is a key element of GLWA's ROP, which was developed to optimize wet weather operational performance across six major wastewater utilities that are wholesale customers to GLWA's regional collection system. This close integration of model and data in the digital twin allows for rapid post-event monitoring and near-real-time analysis of collection system performance during wet weather, facilitating continuous optimization of wet weather operations by GLWA and its Members. BUILDING THE REGIONAL WASTEWATER COLLECTION SYSTEM MODEL AND THE DIGITAL TWIN GLWA developed a new Regional Wastewater Collection System model using EPA's Stormwater Management Model (SWMM) as part of its Wastewater Master Plan project from 2017 to 2020. The collection system model consists of over 4,400 individual model subcatchments and 16,000 pipes covering the entire GLWA service area, including detailed operating rules to represent dry and wet weather operations at key facilities. The model was calibrated to metering data to meet the calibration metrics issued by the Chartered Institution of Water and Environmental Management (CIWEM) Code of Practice for the Hydraulic Modelling of Urban Drainage Systems. The digital twin builds upon the collection system model, and integrates the following elements (Figure 1): - The calibrated and validated collection system model - Gage-adjusted radar rainfall data from Vieux and Associates at a 1-kilometer grid cell resolution adjusted to 36 regional rain gages. The radar rainfall data are sent daily at 5-minute increments and applied to all 4,400 model subcatchments - 540 individual measurement points at 5-minute intervals sent daily from GLWA's Ovation SCADA system. These points represent river stage, pump and gate operating records, and flow metering. - Post Event Reports generated after significant storms to document combined sewer overflow events. The digital twin was built in PipeCAST, a web-based GIS platform that integrates monitoring data and modeling results to facilitate decision-making and post-event analysis in near real time. Its foundation is built upon the concept of the digital twin, which is a virtual copy of the collection system that is used to generate real-time simulations from current weather and precipitation conditions (Figure 2). The PipeCAST system automatically receives data from the radar rainfall grid, climate data from NOAA, and SCADA data through a secure application programming interface (API) connection and runs the collection system model for the previous day. The direct connection to SCADA allows the model to simulate dynamic backwater conditions at each CSO outfall and actual pump station operations. The PipeCAST interface provides a GIS-based visualization of metered and modeled flow, depth, and velocity in the collection system, rainfall statistics, and a dashboard that shows key statistics related to overflow frequency and duration. The GIS-based interface can also provide high-level statistics on system performance, such as pipe capacity during wet weather (Figure 3), which can be used to identify potential locations for additional in-system storage. Since the digital twin integrates data and runs the collection system model daily, GLWA can quickly review system performance after storm events and compare model simulations to observed measurements to use as the basis for ongoing model and measurement improvements. APPLYING THE DIGITAL TWIN GLWA is using the PipeCAST digital twin for post-event analyses, where users review wet weather performance relative to the ROP goals. As an example, GLWA used PipeCAST to evaluate system performance for a recent wet weather event in July 2020. This event was 2.05 inches in 10 hours at the Wayne County Metropolitan Airport, and was approximately a 2-year, 1-hour event. The rainfall was most intense north of Detroit, as seen in the color-coded radar rainfall map in Figure 4. GLWA used the digital twin to evaluate how in-system storage and CSO control facilities conveyed flows from the Conant-Mt. Elliott Sewer that serves Oakland County and central Detroit (Figure 5). Flow and system performance were monitored during this event using the digital twin by reviewing incoming flow from Oakland County, in system storage, and performance at the Leib Screening and Disinfection Facility (SDF) (Figure 6). During this event, the incoming flow from meter SE-S-1, (blue line) peaked at approximately 245 cfs, followed by a short-duration peak from the local combined sewer drainage system in the DWSD collection system (meter DT-S-11, orange line). The in-system storage device, controlled by an inflatable dam, was activated, and no discharge occurred from the Leib SDF facility. Based on the review of data using the PipeCAST digital twin, GLWA was able to conclude that the system performed well during this large event, maximizing in-system storage, and minimizing discharges to the Detroit River. CONCLUSIONS GLWA developed a digital twin of its regional collection system to enable post-event analysis to support the ROP. The digital twin was developed using PipeCAST, a GIS-based tool that integrates metering data and modeling data into a platform that allows for rapid review and assessment of system performance during wet weather. The digital twin is being used to complete post-event analyses following significant precipitation events to further optimize wet weather operations, reducing CSO to receiving waters and maximizing secondary treatment at the WRRF.

Overview: This presentation will describe the approach and share the results of an analysis performed to quantify dry weather flows (DWF) and rainfall derived inflow and infiltration (RDII) into a manifold sanitary sewer force main system with 87 contributing pump stations. The presentation will detail a method of analyzing pump station inflows without the need for installation of expensive flow meters upstream of every pump station in a collection system. The approach also allows for evaluation of historical storms that may have been the catalyst for system-wide assessments prior to any flow metering or other hydraulic assessment plans or studies.

Approach and Current Status: This approach utilizes existing SCADA sensors and data loggers installed in pump stations and transforms pump startup/shutoff activity into a contiguous dataset of inflow rates based on wet well geometry and pump operational envelopes as depicted in Figure 1. This flow data can then be used to analyze DWF diurnal patterns and magnitudes as well as timing and magnitude of any wet weather responses. This data can, in turn, be used to assess performance at individual pump stations in support of a system-wide sanitary sewer master plan update. This approach has been used to develop an updated sanitary sewer system master plan and create a short list of system improvements and capital improvement projects. Key projects from this updated plan have already started detailed design and critical projects are expected to be completed by 2022. This method will also be used to assess pre and post construction I/I and pump station performance.

Results and Benefits: SCADA information was successfully transformed from inconsistent on/off recordings into a contiguous flow hydrograph, as depicted in Figure 2. By analyzing all pump station performance throughout the system holistically, a more complete assessment of system-wide hydraulic performance can be achieved without exorbitant costs supporting system studies, which reserves budget for physical system enhancements. As part of a master plan update, this approach has been employed to economically assess performance of pump stations under existing conditions by simulating a variety of actual and synthetic rainfall conditions. These assessments of individual pump station capacity limitations, force main velocity and head loss performance, and water reclamation facility inflows under both existing conditions and future expected population growth and land development were then used to create an updated sanitary sewer master plan. This plan was executed for a fraction of the cost of a plan with a comprehensive flow metering network.

Conclusion: Using SCADA data is an economical way to analyze actual system performance, assess individual pump station or system capacity, update and recalibrate hydraulic models, and develop potential improvements for sanitary sewer collection systems with any number of pump stations. The approach described in this presentation was executed for a fraction of the cost of a plan with a comprehensive flow metering network installed in gravity sewers and was applied successfully to a collection system with 87 individual pump stations.

Purpose The purpose of the presentation is to illustrate that for digital-water solutions to be successful we need to shift the focus from 'digital' and 'smart,' data and tools to the people and value added. It is easier and cheaper than ever to collect data, numerous tools are available for data processing/visualization, vendors offer different digital twin platforms to integrate and deliver this new information. The key component in the future success of digital-water implementations is not in the digital domain, but instead it is in a clear definition of value added / return on investment and close collaboration with the utility staff on the implementation of the solution. The Jefferson County, AL, application of intelligent algorithms to optimize and prioritize capital investments and asset rehabilitation demonstrates a collaborative approach through every stage of the project and delivers simple to understand optimal solutions based on cutting-edge technology.

Benefits This presentation will provide the following benefits to the audience and industry:

-- demonstrate how formal optimization software can be applied using best practice techniques to evaluate alternatives.

-- provide an apples-to-apples comparison of the SSO Remedial Measures Plan (RMP) developed using traditional trial-and-error modelling techniques and the optimized RMP developed using optimization software

-- illustrate how performing sensitivity analyses on the effectiveness of I/I reduction identifies basins that are cost-effective and ensures conservative conveyance capacity upgrades.

-- demonstrate how asset condition and sediment data can be incorporated into an optimization to determine whether it is more cost-effective to clean the sediment or replace and upsize based on a holistic consideration of conveyance, storage, and I/I reduction alternatives.

-- demonstrate how the optimization approach can incorporate multiple design storms in a single analysis to determine the cost-effective strategy that eliminates SSOs in both events.

-- illustrate how Optimizer can be applied to prioritize the schedule of implementation of a capital program to maximize return on investment.

Background Jefferson County owns and operates wastewater collection systems across nine treatment plant basins serving approximately 600,000 people. A total of approximately 200 reported SSOs are currently recorded by Jefferson County. In 2019, Jefferson County engaged WCS and Hazen to undertake an optimization demonstration project to evaluate conveyance, storage and I/I reduction alternatives for the Valley Creek basin which had the highest density of reported SSOs. This project successfully demonstrated life-cycle cost savings on the order of $54 M (30%) when compared to the Baseline RMP (developed using traditional trial-and-error modelling). It also demonstrated that a prioritized capital improvement schedule could achieve 40% SSO volume reduction within the first 10% of capital expenditure and 98% reduction within 70% of the total capital expenditure required to eliminate all SSOs and achieve surcharge objectives. The Jefferson County, System-Wide RMP Optimization project was undertaken and completed in 2020 based on the success of the Valley Creek demonstration project. The system-wide optimization incorporated five treatment plant basins that are all interconnected either by existing or potential future flow diversion structures.

Details The primary objective of this study was to apply intelligent algorithm optimization technology, 'Optimizer', to evaluate and prioritize remedial measure alternatives including conveyance, storage, inter-basin transfer, treatment, and inflow and infiltration (I&I) reduction alternatives to eliminate SSOs in the 2-year design storm (6-hour and 24-hour events). Secondary objectives include performing sensitivity analyses for key assumptions and to identify aspects of the planning strategy that may require further investigation. The 'Optimizer' software integrates improvement alternatives, comprehensive life-cycle costs, design criteria, and the calibrated hydraulic model of the collection system. In a single analysis, the software applies intelligent algorithms and cloud computing to evaluate tens of thousands of possible solution configurations with respect to life-cycle cost and hydraulic performance. Scenarios and sensitivity analyses are easily completed by adjusting relevant assumptions or inputs and rerunning the cloud-based optimization.

-- Improvement alternatives evaluated in the optimization include:

-- Parallel relief sewers and upsized gravity sewers
-- Sediment removal.
-- Pump station upgrades. Force mains.
-- New storage facilities.
-- Inter-basin diversion controls
-- High rate treatment facilities.

I&I reduction. Unit cost rates adopted for the optimization analysis are planning level estimates developed by Hazen based on the County's bid tab records. The unit costs include capital, O&M, and replacement estimates of all alternatives.

System-wide optimization runs were completed for the following scenarios:

Conveyance-Only
-- Optimization without storage or I&I reduction alternatives
-- Conveyance + Storage - Optimization without I&I reduction alternatives
-- All Alternatives (Ultra-Conservative I&I)
-- All Alternatives (Conservative I&I)
-- All Alternatives (Aggressive I&I)

The optimized solution costs for each scenario are compared in Figure 1. The preferred RMP strategy was the All Alternatives (Aggressive I&I) scenario illustrated in Figure 2. The optimization model used for the prioritization task was formulated to select projects from the Optimized RMP (aggressive I&I reduction scenario). The prioritization analysis used the optimization model to determine the sequence of implementation that provides the maximum return on investment with respect to reducing total overflow volume in the 2-year design storms. The preliminary prioritization results are shown in Figure 3.

Conclusion The Jefferson County, Valley Creek Optimization demonstration project and the System-Wide Optimization project demonstrate both significant savings when compared with an apples-to-apples Baseline RMP developed using traditional trial-and-error modelling and highlighted opportunities for incorporating flexibility to allow the RMP to be adapted over time. The prioritization analysis completed for each project consistently demonstrated a high return on investment whereby the vast majority of SSOs could be eliminated at a fraction of the total program cost. This outcome allows Jefferson County to provide customers with immediate and noticeable improvements in system performance during the early stages of the program while also providing an opportunity for the final stages of the program to be adapted and potentially eliminated if a diminishing return on investment can be illustrated as more data is collected. The County considered the optimization projects successful in minimizing capital expenditure required to meet consent requirements and informing decision making when selecting where to focus short-term investments to maximize value provided to customers.

The City of Spokane (City) owns, operates, and maintains the wastewater and stormwater collection systems (collection system) in Spokane. These systems consist of a variety of assets and facilities, and accomplish the important purposes of:
-- Collecting and conveying wastewater from residents and businesses, along with stormwater runoff from areas with combined sewers, to Riverside Park Water Reclamation Facility (RPWRF) for treatment and discharge.
-- Collecting and conveying stormwater runoff from surfaces throughout the City for at least partial treatment and discharge.
Stormwater is discharged to a variety of endpoints, including the Spokane River (as part of the City's municipal separated storm sewer system [MS4]), into the ground (via bioretention and underground injection wells, also known as dry wells), or evaporated. Both the wastewater and stormwater systems have historically been operated with minimal data collection and supervisory or remote control capabilities, primarily because the assets and facilities have generally been achieving their desired performance. However, the City has long recognized the potential for improved performance of the collection system through the implementation of supervisory control and data acquisition (SCADA) for the collection system (referred to as the 'collection system SCADA'). The collection system SCADA will allow for increased monitoring and alarming, improved data management, and the potential application of real-time control (RTC) at combined sewer overflow (CSO) control facilities and interceptor protection tanks (IPTs). These improvements will result in reduced CSOs, reduced risk of sanitary sewer overflows (SSOs), more efficient operations and maintenance (O&M), and an increased understanding of how the system operates. Implementation and operation of the collection system SCADA would also have staffing and cost (capital and O&M) impacts on the City that need to be considered. Now that the City has constructed large and complex CSO control facilities and IPTs throughout the City as part of complying with the City's Waste Discharge Permit for CSOs and the RPWRF, the City has prepared a Collection System SCADA Master Plan (Master Plan), completed in May 2019. The purpose of this Master Plan was to:
-- Identify and evaluate potential opportunities for RTC at CSO control facilities and IPTs that could be implemented in the future, and assess costs and staffing impacts
-- Identify and prioritize monitoring and alarming improvements to the collection system, and assess costs and staffing impacts
-- Identify SCADA network and infrastructure needs to support the implementation and operation of the monitoring, alarming, and RTC improvements identified, and assess costs and staffing impacts

Establish an implementation roadmap for the selected collection system SCADA improvements and assesses the capital, O&M, and organizational impacts of implementing the recommendations As part of the Master Plan, the project team developed a system model of the City's sewer system, and simulated possible RTC scenarios to manage wet weather events. The model results indicated a reduction in CSO volume of between 30 to 40 percent, as shown in Figure 1, depending on the storm and level of RTC implemented. Not only did RTC reduce CSO volumes, but it also allowed for better protection of critical points in the interceptor system that are susceptible to surcharging. In conclusion, this project established a clear and implementable path forward for the City to implement SCADA in their collection system. The SCADA system will provide benefits on many fronts -- CSO reduction through RTC, improved O&M and performance due to real-time data collection, and improved reporting. This presentation will focus on describing the gaps and solutions identified in the Master Plan, with a particular focus on discussing the RTC analysis and results, and presenting the anticipated O&M impacts of making the improvements. This completed project was unique because it focused on using RTC as an adaptive management strategy for completed CSO control facilities. Many other municipalities are now in similar situations, where they have finished building CSO control facilities and are dealing with what to do if the control facilities are not sufficient to meet CSO reduction standards. This project provides an example of retrofitting completed facilities with RTC and monitoring in order to improve performance. Following the completion of this project the City has moved forward with designing and constructing the collection system SCADA, and will be making monitoring improvements at several high-priority pump stations and CSO control facilities.

The sewer collection system is an important infrastructure that protects the public health by safe disposal of wastewater. Sewer lines are designed for a certain carrying capacity, which is reduced by blockages, resulting in the risk of flooding and the occurrence of Sanitary Sewer Overflows (SSOs) that release raw contaminated sewage. According to the US Environmental Protection Agency, around 25% of the 23,000 to 75,000 SSOs are due to sewer line blockages related to fat, oil, and grease (FOG) deposits. Research studies have been conducted to understand the FOG deposit formation mechanism, sources of chemical constituents involved in FOG deposit formation, and engineering solutions to overcome the FOG-related SSO problems (He et al. 2017). These studies have shown that the long-chain free fatty acids (LCFFAs) from FOG discharge undergo a saponification reaction with calcium ion present in wastewater to form insoluble calcium soap known as FOG deposits that can adhere to sewer line surfaces to reduce its carrying capacity. The primary sources of FOG discharge are the food service establishments and domestic kitchen sinks, whereas sources of calcium can be attributed to the background wastewater as well as concrete structure corrosion. Previous research suggests that calcium released from concrete corrosion may have a more adverse impact on the FOG deposits accumulation on sewer line surfaces (Iasmin et al., 2014). The main contributor of calcium from concrete is the cement used as a binder. Research shows that the substitution of Portland cement by Supplementary Cementitious Materials (SCMs) such as Fly Ash (FA) reduces the calcium leaching potential of concrete exposed to corrosive media (Rozière et al. 2009). In the research discussed in this presentation, we will evaluate the use of FA to replace a large volume of cement to produce an alternative binder material for sewer line construction that reduces the calcium leaching from concrete corrosion, therefore, reducing the FOG deposit formation. We will also explore different factors such as- porosity and pore size, surface roughness, and surface pH of different sewer line construction materials to understand the FOG deposits adhesion mechanism. In our study, two High Volume Fly Ash (HVFA) concrete materials, Concrete-50 and Concrete-75, were produced by keeping the water to cementitious material ratio at 0.42 using 50% and 75% replacement of cement by FA, respectively. Additionally, control samples, Concrete-0, were cast with no FA replacement and used as a standard to study the effect of FA replacement. HVFA samples were tested for compressive strength (ASTM C39-15) and chemical durability (ASTM C267) to check their suitability as sewer line construction materials. Concrete samples were also tested for their calcium leaching potential at pH 5 and 7 by submerging them into Deionized Water (DI) for 90 days. The FOG deposit formation test was performed at pH 7 for a 30 days testing period by submerging concrete samples in synthetic wastewater prepared by mixing oleic acid as the LCFFA source, canola oil, distilled water, and calcium chloride as a background calcium source. The compressive strength test results show that after 90 days of sealed curing, Concrete-50 is well above 4 ksi; a threshold compressive strength used for sewer line construction materials. The chemical durability test results showed that 50% FA replacement enhanced the durability of concrete against corrosive media. After 90 days of leaching under corrosive conditions, Concrete-50 and Concrete-75 samples showed 75% and 86% reduction in calcium leaching, respectively, when compared with the standard Concrete-0 sample. The calcium leaching potential from HVFA samples also revealed that the calcium ion diffusion flux of the HVFA samples was almost similar and lower than the Concrete-0 samples. This observation suggests that FA replacement can reduce the porosity of concrete; thereby, reduce the calcium leaching potential from sewer lines under corrosive media. In addition to the calcium leaching potential, HVFA samples were also tested for heavy metals such as Arsenic (As), Cadmium (Cd), Chromium (Cr), Selenium (Se), Mercury (Hg), and Lead (Pb) leaching potential. After 90 days of leaching test, the cumulative toxic ion concentration for Concrete-50 and Concrete-75 samples did not exceed the pollutant discharge limit provided by USEPA clean water act. FOG deposition formation test results showed that 50% and 75% FA replacement can reduce FOG deposit formation by 55% and 67%, respectively. Therefore, results from this study suggests that a significant decrease in calcium release can be achieved through the use of FA as a cement replacement, which can eventually reduce the FOG deposit formation. To understand the FOG deposits adhesion mechanism on different sewer line surfaces, a combined FOG deposit formation test was conducted by submerging Concrete-0, Concrete-50, Concrete-75, Vitrified Clay Pipe, and PVC pipe into the synthetic wastewater for 30 days of the testing period. This test results revealed a similar FOG deposit formation trend (highest FOG deposition on Concrete-0 samples and lowest FOG deposition on Concrete-75 samples); however, no FOG deposits were found on PVC and Vitrified clay pipes surfaces. Hence, we hypothesize that the FOG deposit adhesion phenomena are primarily controlled by the sewer line material's surface properties in addition to the availability of calcium and LCFFAs. In the next phase of this study, to understand the effect of porosity, pore size, and surface pH on FOG deposit adhesion mechanism, porous ceramic materials with controlled pore size will be used for the FOG deposit formation test. Whereas, to study the effect of surface roughness on FOG deposits adhesion phenomena, HVFA and PVC samples will be prepared at different surface roughness and tested for the FOG deposition test. Results from these tests will be able to identify the key factors that affect the FOG deposits adhesion mechanism. According to Clean Watershed Needs Survey (USEPA, 2016), the total need for replacing sewer lines is $42 billion on top of $25 billion and $18 billion in the new collector and new inceptor sewers, respectively. Therefore, the present study will reduce the future sewer line construction and maintenance cost by reusing FA as an alternative sewer line construction material. Additionally, by studying the FOG deposit adhesion mechanism, this research will provide wastewater collection system utilities with strategies to develop new construction materials or potentially design future coatings that can enhance existing alternative materials to limit or reduce the FOG deposit accumulation on the sewer surface.

Purpose of the research and key message and knowledge transfer In summary, the key message from this work is two-fold: Close collaboration with industry enabled the design of a scientifically rigorous and highly focused experimental program in live sewers. The latter provided conclusive answers within four months and identified a way forward leading to technology implementation and identification of researchable gaps needed to make the technology more widely usable. The specific knowledge that this paper will transfer is that reliable humidity sensors to measure relative humidity >98% in gravity sewers can be achieved. This is a tool that will help the water industry to improve the management of its concrete gravity sewers and other infrastructure that is in contact with high concentrations of gaseous H2S in humid air. No commercially available technology can do this. In this work, the authors have demonstrated experimentally, building upon advanced design work, that optical fibre sensor systems represent an innovative, versatile and affordable technology that can be brought to bear very effectively on key monitoring problems in the water industry's infrastructure. Through the work reported and results demonstrated in this paper, the research has very successfully created a synergy of the skills of an academic group and a major water utility to identify, tackle and solve key problems in sewers -- problems that impact directly on consumers and are seen world-wide in the industry. In doing so, in an effective way like this, the paper demonstrates the benefits to be gained from a close collaboration between the two groups. Above all new knowledge has been transferred -- in both directions -- and through that new and more effective approaches to problem solving developed. Specifically, the project is based on new technology for humidity measurement -- this is an essential parameter that needs to be monitored to minimize corrosion of concrete gravity sewers. No humidity sensors that last longer than ~1 week in the aggressive gaseous atmosphere of gravity sewers and can reliably measure humidity above 98% relative humidity are available. Thus supporting better predictive maintenance, in situ monitoring of a sewer and an anaerobic digester is undertaken using innovative optical fiber-based sensor systems: these being chosen because of their inherent safety in the potentially methane rich environment, their ability to be protected against biofouling, their light weight, low power consumption (or battery powered), 4G compatible and ease of use over the long lengths (up to kilometers) needed for some applications. The multi-parameter sensors systems designed and evaluated are also well suited to use with the high levels of hydrogen sulfide, coupled with high humidity seen in sewers and biogas rich environments -- environments where conventional electronic-based sensor systems regularly fail due to acid attack. Benefits of Presentation and why this presentation should be selected and will provide a benefit to our industry Sydney Water's current management strategy relies on reducing gas phase H2S by the addition of chemicals (ferrous chloride, magnesium hydroxide and calcium nitrate) to the waste water. This strategy, implemented in a number of waste-water systems in Sydney, is costly (about $4M p.a.) and its effectiveness in reducing corrosion is unknown, due to an inability to directly measure concrete corrosion and deterioration of rehabilitation materials. No matter how effective this program is in slowing the corrosion rate, there will always be concrete corrosion and a method for prediction and detection of imminent pipe failure is needed because of the enormous costs of such failures. In particular, it is necessary to determine how much of the ubiquitous H2S and H2O can be left in the air in order to reduce corrosion rates to less than what Sydney Water defines as a sustainable 0.5 mm/year. Gravity sewers are buried infrastructures that undergo chemical deterioration because of microbiological induced corrosion (MIC). Microbiologically induced oxidation (MIC) of gaseous H2S into sulfuric acid results in concrete sewer corrosion. MIC costs billions of dollars annually to the water industry and has been identified as a main cause of global sewer deterioration. Sydney Water spends about A$60-80M annually repairing and protecting concrete gravity sewers and has a regular sewer inspection program. Currently, sewer condition is determined by man entry which is inconvenient and hazardous, has major safety issues and is expensive and so there is a need to minimize man entry by reducing damage from MIC, which can be reduced by controlling moisture in the gravity sewers to values approximately below 85% RH. No conventional humidity sensors that can last longer than one week exist under these aggressive conditions. Tackling this, this work describes the first and successful attempt to use optical fiber-based photonics sensors to monitor humidity in the range >98%RH in the overhead space in gravity sewers. This information is fundamental to develop reliable models that can predict the end-of-service life (EOSL) of concrete sewers. With the ageing of an asset, there is a point in its life, termed the End of Service Life (EOSL) and by doing so, it will be possible to save hundreds of millions of dollars that would have to be spent if collapses occur or premature repairs carried out to prevent a potential collapse. The approach developed together thus supports Sydney Water's corrosion management strategy, which has been created as an outcome of over $3M in research projects striving to better understand the processes that lead to the formation of H2S dissolved in the waste-water. Status of Completion The work done together by the authors has enabled the project to reach a high level of completion and results of the use of the systems in sewers and biodigesters have been reported in the major international media. Thus to date, three trials of tailor-made fiber optic-based sensor systems to monitor humidity in the range >98% RH in sewer air that contains high gaseous H2S concentrations of up to around 400 ppnv have been successfully completed. Based on this Sydney Water is implementing the photonic sensors technology in small catchments. For extensive implementation, Sydney Water is supporting extra research to develop 'low cost' interrogators that would make the technology more cost effective, well suited to long term, in-the-field research. Conclusion and take-away message The key 'take-away' message is that this is a significant innovation in optical fiber and photonic sensors that they can now be used in these 'dirty' environments. This has also allowed in-situ long-term monitoring under conditions where current technology fails. The major outcomes of the series of experiments carried out over the last five years is clear; fiber optic sensors systems can operate effectively, in the long term, in remote locations under battery power, saving of millions of dollars.

Purpose of the Presentation
The purpose of the presentation is to demonstrate the innovative approach to assessing sanitary sewer overflow compliance being used by Christchurch City Council (CCC), New Zealand. Continuous simulation of long-term rainfall data is used by CCC to assess overflow compliance rather than using a synthetic rainfall pattern that is intended to represent the target return period. Furthermore, collection system planning is based on a design storm that was generated based on statistical analysis of the 25-year rainfall continuous model simulation results. The presentation will show that SSO frequencies are sometimes not well represented by rainfall events that have an equivalent return period and that antecedent rainfall conditions and hydrologic and hydraulic (H&H) routing of flows unique to each collection system can help to select a better design storm to simulate overflow return periods of interest.

Benefits of the Presentation
This presentation should be selected because it will provide the following benefits to the audience and industry:
-- it will illustrate the difference in overflows predicted using a synthetic return period rainfall event and the overflow frequencies calculated based on continuous simulation of long-term rainfall.
-- it will show how the environmental regulators for Christchurch, New Zealand require continuous simulation of long-term rainfall to validate overflow compliance.
-- it will demonstrate the value in reviewing continuous simulation overflow results to develop more accurate design storms to be used for infrastructure planning.
-- it will show how a utility can use an improved approach to developing design storms to ensure investments are made in new infrastructure and infrastructure rehabilitation programs that are targeted in the relevant area of the collection system in order to meet consent conditions effectively.
-- it will illustrate how three different 15-year windows of rainfall from a 25-year period of rainfall can affect the outcome of the overflow frequency assessment but that the climatic variations do not have a significant bearing on the outcome of the overflow compliance.

Project Background
Christchurch City Council provides drinking water, stormwater, and wastewater across five council areas in the North Island, New Zealand. Christchurch has a complex wastewater network serving 373,000 people. Christchurch has a large and complex wastewater network serving approximately 373,000 people with a single wastewater treatment plant capable of treating 650 megaliters/day (MLD). Christchurch has two main river systems: the Avon River and the Heathcote River, which both flow into the Avon-Heathcote Estuary which then flows into the Pacific Ocean (see Figure 1). Christchurch City Council (CCC) holds a resource consent (permit) for the overflows to waterways. The resource consent allows an overflow frequency to each of these receiving environments which decreases over time to a 2 year ARI, based on 15 years of long time series modelling. In addition, no overflow site may overflow more than every six months on average, based on the same long time series modelling. The Christchurch Overflow Compliance Assessment project was commenced and completed in 2020. The project was completed using the latest model calibration completed in 2020 and was an update to the overflow compliance assessment completed based on the model previously calibrated in 2015. The overflow consent conditions require that continuous simulation of the most recent 15-years of rainfall be completed to assess the system performance. This project included continuous simulation of the most recent 25-years of rainfall and analysis of overflow results from three different 15-year windows to investigate whether the overflow frequencies are significantly affected by the 15-year window of rainfall selected. Overflow control measure alternatives to achieve the 2-year overflow return period target were previously assessed using a synthetic rainfall pattern. To determine whether this rainfall pattern would be appropriate for future planning studies, the long-term continuous simulation results were used to predict the location of overflows occurring more frequently than once every two years and then compared to the locations predicted using the synthetic rainfall pattern. This assessment demonstrated that the synthetic event did not provide a good representation of actual overflows and a new design storm was developed based on historic events.

Project Details
The objective of the wet weather overflow compliance assessment was to determine wastewater overflow compliance with the consent conditions stipulated by the environmental protection agency, Environment Canterbury (ECAN). This was undertaken using the calibrated hydraulic model to simulate system performance based on continuous rainfall data from the period of 1995 to 2020. To determine whether the rainfall window could have an impact on any of the consent compliance results, the assessment of overflow frequency and volume was performed for the following time periods:
1. 1995 to 2020
2. 1995 to 2010
3. 2000 to 2015
4. 2005 to 2020

The compliance assessment results summarised in Table 1 and Table 2 below, show the level of variation in overflow response according to the rainfall window chosen. While the variations are significant in some time windows, the compliance outcomes would not have been affected. Table 1: Overflow volume in each catchment for different time periods Table 2: Overflow frequency in each catchment for different time periods The objective of the design storm review was to ensure the design storm used for wastewater infrastructure planning is representative of a 2-year return period based on the 25-year continuous simulation results. The assessment was based on overflow volume, peak and spatial distribution. The 25-year continuous simulation results were reviewed to identify historical events that corresponded to an overflow return period of approximately 2 years. Due to a high degree of rainfall spatial variance in the historical events that were close to a 2-year return period, it was decided to repeat the overflow statistical analysis by receiving environment rather than system-wide. Based on the results from the overflow statistical analysis, several 2-year design storm alternatives (synthetic and historical) were shortlisted. The preferred 2-year design storm, presented in Figure 1, was the August 5, 1995 (northern basins) and April 17, 2014 (southern basins) composite event. This event provided the best representation of 2-year ARI overflow volumes, peaks and spatial distribution and comprised events of sufficiently short duration to be convenient for planning.

Conclusion
The results from this study demonstrate a novel and improved approach to developing design storms to ensure investments are made in new infrastructure and infrastructure rehabilitation programs that are targeted in the relevant area of the collection system in order to meet design criteria and/or compliance targets.

Purpose This presentation will share how Jefferson County accomplished the following tangible results: A) shifted from being nearly 100% reactive to 70% proactive with its sewer cleaning program (see graphic 1), and B) reduced dry weather SSOs by 44% with no increase in O&M budget (see graphic 2.) How did we do it? There are three legs to our strategy stool: first, we structured the dry weather program effectively, second, we use data to allocate resources based on risk, and finally, we balance crew productivity and quality expectations. The County's O&M structure includes three key elements. The structure in and of itself does not reduce SSOs, success is driven by how the structure is implemented, but the key elements of the structure must be in place to be successful. The three elements include:

1. Reactive -- the goal is to quickly and effectively respond to service requests (e.g. customer calls) and to also collect the right data to support root-cause-analysis to minimize the chance of re-occurrence.
2. Preventive -- the goal is to identify pipes with O&M risk and clean them at the right frequency, which is not too little, and not too much (e.g. the goldilocks frequency.) If implemented correctly, this will prevent SSOs on pipes with known risk without 'cleaning clean pipe.'
3. Proactive -- the goal is to assess the rest of the system that has no known O&M risk (i.e. finding needles in the haystack.)

The County uses an acoustic inspection technology, which is a low cost, high production rate method. The second leg of the stool involves using data to allocate resources based on risk. This starts with collecting the right data from our three primary sources of O&M condition assessment: sewer cleaning, CCTV, and SL-Rat. All three sources use 'code-based' (instead of narrative) data collected on each pipe asset. Data that is collected in this structure allows it to be both collected and used efficiently and effectively. We then mined all the historical data sources in order to determine which pipes need to be set on a recurring, standard cleaning frequency. Jefferson County then implemented a decision support system (DSS) to apply consistent, objective business rules via an algorithm in order to make recommendations to optimize sewer cleaning as new data points are collected. The DSS generates recommendations that a planner/scheduler then reviews. If the planner/schedule accepts the recommendation, the DSS automatically updates the Cityworks database. The types of recommendations include: increase/decrease the frequency, add/remove a pipe to the schedule, and accelerate/push out the next scheduled cleaning date. The third leg of the stool focuses on achieving the right level of production from crews, while not sacrificing quality. The quality part of the equation started with developing best practice SOPs and providing extensive training. The next phase of the quality program will focus on quantitative measurement and feedback. The productivity part of the equation is helped by the strategy itself: as the preventive and proactive portions of the program produce SSO reductions, it allows shifting from reactive to planned work. Planned work is much more efficient than reactive work, therefore crews can do more preventive and proactive work -- which, further reduces reactive work. It's a perpetual improvement cycle. Other mechanisms to increase crew production include effectively using GIS to make spatially efficient work packages and setting and tracking appropriate production goals.

Benefits of Presentation
Many utilities across the nation have goals to significantly decrease the number of Sanitary Sewer Overflows (SSOs) in their collection systems. Often, 50-90% of SSOs are caused by maintenance-related issues such as grease, roots, non-'flushable' wipes, and debris. Optimizing the maintenance program can be the most cost-effective method to drive rapid SSO reduction. These reductions are pennies on the dollar as compared to CIP intensive work like rehabilitation and capacity upgrades. This presentation includes a unique application of a Decision Support System (DSS) that only Jefferson County and one other County have implemented. The other County has achieved over a 50% reduction in dry weather SSOs with no increase in budget. The presentation also includes best practices around a relatively new acoustic monitoring technology. This presentation will touch on a highly structured pilot of the acoustic technology resulting in a statistically significant result for the best trigger score and its highest value usage in the program. Finally, this presentation will put a different spin on old concept -- sewer cleaning programs are as old as sewer systems themselves, but this approach takes advantage of modern techniques and technology that are structured and coordinated to maximize tangible benefit.

Status of Completion
Nearly all the work described above has been completed, which has led to the 44% reduction in dry-weather SSOs. Completed items include:
1. Best practice SOPs have been developed and training has been conducted
2. Configuration updates to Cityworks have been made and high-quality data is being collected
3. The DSS has been implemented to optimize cleaning
4. A planner/scheduler function was implemented

Like other high performing utilities, Jefferson County seeks to continually evolve, and has several 'next steps' its moving forward on. These include:
1. Revamped SSO review meeting to perform root cause analysis and corrective measures to prevent re-occurrence of SSOs.
2. FOG program assessment to apply similar asset management principles to the commercial program (food service establishments) and public outreach.
3. Sewer cleaning QAQC - quantitative measuring and feedback loop

Conclusions
The primary conclusion is that Jefferson County's approach works, and its very cost effective. The great thing is that anyone can do it if they bring the right best practices, the right tools, and implement them correctly. An asset management-centric O&M program needs to be planned and designed comprehensively -- there are a LOT of moving parts. Applying asset management principles to O&M happens in our industry, but it's not nearly as common as applications on the structural and capacity portions of programs. But its 'pennies on the dollar' to achieve tangible benefits on the O&M side vs the more capital centric portions of a program.