1 Detection of misconnections using DTS (Distributed Temperature Sensing) Jes Vollertsen Professor, Aalborg Universitet M. Uggerby, EnviDan A/S, Silkeborg, Denmark J. Vollertsen, Aalborg University, Denmark A. Laustsen, Aarhus Vand A/S, Aarhus, Denmark C. Jensen, Per Aarsleff A/S, Åbyhøj, Denmark R. Schilperoort, Royal HaskoningDHV, Nijmegen, The Netherlands (and others, but the list would get too long)
2 Why bother about misconnections Storm drains connected to sanitary sewers Sanitary sewers are designed for dry-weather flow Misconnected storm drains cause sanitary sewer overflow Misconnected storm drains cause sanitary sewer flooding Sanitary drains connected to storm sewers Discharge of pathogens to sensitive waters Discharge of organic matter to sensitive waters Discharge of nutrients to sensitive waters
3 How to find sanitary sewer misconnections Run CCTV during rain, check if water comes from sanitary service laterals Problem: Difficult to plan being at a location during rain Add water to all storm drains and roof drains while inspecting the sanitary service lateral with CCTV Problem: Requires access to private property Problem: All drains must be found and accessible Add electronic marker chips to drains and detect if they pass the pipe further downstream Problem: Requires access to private property Problem: All drains must be found and accessible
4 DTS is an alternative hereto Temperature sensitive glass fiber cable Typical data for a medium sensitive setup: ± 0.1 C over several km of cable length Logging every minute Logging every meter
5 Pull the cable through the sewer Royal HaskoningDHV, 2013 Royal HaskoningDHV, 2013
Up side connections to get all of the pipe 6
7 Connect the cable to the DTS data acquisition computer Trailer for ease of installation Standard computer to upload data to an ftp server DTS
8 Test catchment and cable installation Water was added to all storm drains and roof drains while inspecting sanitary service laterals with CCTV Houses 10, 47, and 93 were identified having misconnections One we made ourselves
Raw dry weather data 9
12-Aug-2013 08:00:37 30 28 Dividing the cable into the different stretches 26 12-Aug-2013 07:30:37 22 12-Aug-2013 07:00:37 20 18 12-Aug-2013 06:30:37 16 14 12-Aug-2013 06:00:37 1150 1170 1190 1210 1230 1250 1270 1290 1328 1348 1368 1388 1408 1428 1448 1468 87 127 167 347 387 476 763 803 1138 1478 12-Aug-2013 08:00:37 12-Aug-2013 08:00:37 12-Aug-2013 07:30:37 12-Aug-2013 07:30:37 12-Aug-2013 07:00:37 12-Aug-2013 07:00:37 12-Aug-2013 06:30:37 12-Aug-2013 06:30:37 12-Aug-2013 06:00:37 826 846 866 886 906 926 946 966 991 1011 1031 1051 1071 1091 1111 1131 12-Aug-2013 06:00:37 489 509 12-Aug-2013 08:00:37 12-Aug-2013 08:00:37 12-Aug-2013 07:30:37 12-Aug-2013 07:30:37 12-Aug-2013 07:00:37 12-Aug-2013 07:00:37 12-Aug-2013 06:30:37 12-Aug-2013 06:30:37 12-Aug-2013 06:00:37 401 421 446 466 12-Aug-2013 06:00:37 168 529 549 569 188 589 208 228 1518 1558 1598 609 636 656 248 265 1638 1678 676 285 1718 696 716 305 1758 12 1798 736 325 756 345 Temperature ( C) 24
Analyzing data manually Lateral inflows Flow direction Cable whined up and hanging in the manhole Cable creeping out of pipe invert, not being submerged fully when there is flow 11
Rainstorm on July 27, between 5:58 and 6:28 gave heavy inflow Street 1 Street 2 Known misconnection Unknown misconnection Street 3 Street 4 27-Jul-2013 07:13:52 20 27-Jul-2013 06:58:52 19 Date 17 27-Jul-2013 06:28:52 16 27-Jul-2013 06:13:52 Temperature ( C) 18 27-Jul-2013 06:43:52 15 27-Jul-2013 05:58:52 14 27-Jul-2013 05:43:52 168 188 208 228 248 265 285 305 325 345 401 421 446 466 489 509 529 549 569 589 609 636 656 676 696 716 736 756 Street 5 826 846 866 886 906 926 946 966 991 1011 1031 1051 1071 1091 1111 1131 13 Main street 27-Jul-2013 07:13:52 27-Jul-2013 06:58:52 Date 27-Jul-2013 06:43:52 27-Jul-2013 06:28:52 27-Jul-2013 06:13:52 27-Jul-2013 05:58:52 27-Jul-2013 05:43:52 11501170 1190 1210 1230 1250 1270 12901328 1348 1368 1388 1408 1428 1448 1468 87 107 127 147 347 367 387 476 763 783 803 11381478 1498 1518 1538 1558 1578 1598 1618 1638 1658 1678 1698 1718 1738 1758 1778 1798 12
Automating the search for misconnection Algorithm that identifies any inflow event Identification in space as well as time Both dry and wet weather inflows Database with information on known connections Identified by CCTV inspection data Identified as locations with inflow Database with local rain events Rain gauge mounted on the DTS trailer Time, duration and intensity of events Linking databases to identify misconnections For example: Does a known lateral always show inflow when it rains heavily? Then it is a misconnection 13
14 The state of the algorithm Concept A rapid change in temperature in both time and space identifies an inflow event Approach A numerical filter identifies such locations One issue Not all identified inflow events are true inflows. Some are caused by the cable not lying straight in the pipe invert (Loch Ness Monster)
15 Concluding remarks DTS can identify misconnections without having to move onto private property More inflows were found than what a very detailed traditional inspection could identify Future work focuses on automating the routine, and finding the right balance between automation and manual data analysis Future work focuses on building an user-friendly, web-based platform showing DTS data and analysis results
16 That s it, folks
17 That s it, folks Royal HaskoningDHV, 2013