Development of Real-Time Predictive Analytics Tools for Small Water Distribution System
| Main Author: | |
|---|---|
| Language: | English |
| Published: |
University of Cincinnati / OhioLINK
2017
|
| Subjects: | |
| Online Access: | http://rave.ohiolink.edu/etdc/view?acc_num=ucin1504802657161527 |
| id |
ndltd-OhioLink-oai-etd.ohiolink.edu-ucin1504802657161527 |
|---|---|
| record_format |
oai_dc |
| spelling |
ndltd-OhioLink-oai-etd.ohiolink.edu-ucin15048026571615272021-08-03T07:04:08Z Development of Real-Time Predictive Analytics Tools for Small Water Distribution System Woo, Hyoungmin Environmental Engineering EPANET-RTX SCADA Dynamic Time Warping Pump Valve Data Clustering One of the uncertain and variable components of a water distribution system model is the pump-head discharge curve, which plays a significant role in regulating the head and flow within the system. SCADA information can be expected to contain a rich trove of data relating to the head-discharge variation associated with individual pumps and pumps operating in specific configurations. To date, such data are typically ignored when developing or updating hydraulic network models. This research proposed a new algorithm for pump curve estimation using commonly available SCADA information (real-time or historical). The results showed the benefits of a new concept for constant-speed pump using SCADA database without a specially designed field test. The recursive least squares method is presented as a strategy for automated calibration of pump curves at a pump station using real-time SCADA data. The proposed algorithm is extended to analyze the characteristics curves of variable speed pumps in parallel operations. The dimensionless speed quotient was introduced to a linear model to consider the speed variation effect and energy calculation. The method was applied to a pump station in a water utility, KY. It revealed that even the identical pumps have different efficiency curves and the current control operating strategy does not guarantee the best energy operating point in need of sophisticated operation strategy to save energy more. Two data clustering methods (DBSCAN and k-means clustering) were applied to analyze the valve throttling for pump operation. DBSCAN can discriminate data clusters in nonlinearity condition better than k-means clustering. Then dimensionless valve head loss coefficient was calculated with the estimated pump curve.On the other hand, a quantitative analysis method was newly proposed for tracer study. The tracer studies provide significant information to assess the ability of a given network model to represent the underlying hydraulic and transport characteristics of the network. In most cases, the quality of the network model predicted representation of the observed water quality time series is assessed primarily by visual inspection of the data. The use of standard quantitative metrics, such as arrival times, sum-of-squared errors (SSE), and correlation analysis at different time lags to assess the differences between the observed and predicted time series provide some information but are not entirely appropriate for the paired data signals. In this study, the use of Dynamic Time Warping (DTW) is presented as a method for quantitative analysis. DTW uses dynamic programming to match the elements of two time series, in a sequential approach, to minimize the SSE of the two signals. While the SSE provides one goodness-of-fit metric, the resulting length of the warping path also provides additional information as to the degree of the shift between the two data streams. Finally, the recently developed technology of EPANET-RTX, the linkage of the hydraulic network model to SCADA data, was applied to a small utility in OH. A pilot field study was implemented to collect data for calibration of the real-time model that predict hydraulics and water quality. 2017 English text University of Cincinnati / OhioLINK http://rave.ohiolink.edu/etdc/view?acc_num=ucin1504802657161527 http://rave.ohiolink.edu/etdc/view?acc_num=ucin1504802657161527 unrestricted This thesis or dissertation is protected by copyright: all rights reserved. It may not be copied or redistributed beyond the terms of applicable copyright laws. |
| collection |
NDLTD |
| language |
English |
| sources |
NDLTD |
| topic |
Environmental Engineering EPANET-RTX SCADA Dynamic Time Warping Pump Valve Data Clustering |
| spellingShingle |
Environmental Engineering EPANET-RTX SCADA Dynamic Time Warping Pump Valve Data Clustering Woo, Hyoungmin Development of Real-Time Predictive Analytics Tools for Small Water Distribution System |
| author |
Woo, Hyoungmin |
| author_facet |
Woo, Hyoungmin |
| author_sort |
Woo, Hyoungmin |
| title |
Development of Real-Time Predictive Analytics Tools for Small Water Distribution System |
| title_short |
Development of Real-Time Predictive Analytics Tools for Small Water Distribution System |
| title_full |
Development of Real-Time Predictive Analytics Tools for Small Water Distribution System |
| title_fullStr |
Development of Real-Time Predictive Analytics Tools for Small Water Distribution System |
| title_full_unstemmed |
Development of Real-Time Predictive Analytics Tools for Small Water Distribution System |
| title_sort |
development of real-time predictive analytics tools for small water distribution system |
| publisher |
University of Cincinnati / OhioLINK |
| publishDate |
2017 |
| url |
http://rave.ohiolink.edu/etdc/view?acc_num=ucin1504802657161527 |
| work_keys_str_mv |
AT woohyoungmin developmentofrealtimepredictiveanalyticstoolsforsmallwaterdistributionsystem |
| _version_ |
1719452928981336064 |