“How much sewer rehabilitation is enough?” and “What will a
rehabilitation program cost?” are two questions that have plagued municipal
program planners and our industry from the 1970s. The first question depends on
knowing a relationship between the quantity of pipe rehabilitation in a project
(usually full-length lining) and the amount of I/I eliminated.
The second question may be answered easily (in terms of a cost range) once the amount of required sewer rehabilitation is established in response to the first question. Information about the range of unit costs for various sewer rehabilitation materials and processes is readily available. However, (contrary to common thinking) the success of a sewer rehabilitation program for I/I reduction is not dependent on the “right” product or process. Instead, it is more dependent on a sound program strategy, stopping migration of groundwater in the pipe bedding to bypassed defects (a “system” approach), aggressive selection of pipe segments for rehabilitation, and level of intensity of lining. Each of these points is crucial to answering the first question and will be discussed further in this section.
Sewer rehabilitation is often a specific requirement in Commissioner’s Orders and Consent Orders on municipal sewage collection systems with unpermitted overflows. The idea of conducting sewer rehabilitation to decrease the level of I/I (infiltration and inflow) has been a part of EPA’s national strategy since the development of the 201 facility planning process as part of regulations to implement the 1972 Clean Water Act. At times it has been controversial and come under scrutiny. In the early years of the Construction Grants Program, EPA’s “Conklin Report” (Conklin & Lewis, 1981) evaluated work in 18 municipal systems and concluded that generally the efforts to eliminate excessive I/I were not successful. More recently, Lukas, et al (2001) reported on a WERF (Water Environment Research Foundation) study of trends of I/I reduction where 44 utilities had been contacted and data collection efforts were underway. However, the authors pointed out a major problem for accomplishing this important work: “Unfortunately, none of the information between projects was comparable, due to the differences that can exist between two projects. These differences can range from drainage area of the flow meter basin, amount of RDII present prior to rehab, (and) the amount of system rehab performed.”
An evaluation of effectiveness for I/I removal was conducted by WERF (Water Environment Research Foundation) in 2003 (Merrill, et al, 2003) and reported a wide range of levels of effectiveness for twelve RDI/I (rainfall derived infiltration and inflow) reduction projects in six municipal systems. Merrill et al observed that most RDI/I removal projects in the USA are not documented. Second, of those published, few provide good data. Last, data gathering and analysis of data from various programs is hampered by lack of documentation, weaknesses in monitoring techniques, and variation in how information is reported.
I/I reduction information measured in the first Nashville OAP (Overflow Abatement Program 1989-2006) was a significant exception to the problems reported in the WERF study. Early in the program, the managers saw the work in Nashville as an ideal opportunity to develop the necessary documentation and procedures to measure the effectiveness of sewer rehabilitation. Once the procedures were implemented, then rehabilitation effectiveness became a valuable tool to make adjustments in the program and the resulting data were used for modeling and verifying system improvements for overflow reduction. Twenty-seven projects were analyzed with before-and-after flow and rainfall monitoring (Kurz 2012a). This program used ADS flow monitors, which were the only flow monitoring instruments that had been tested by the US-EPA ETV (Environmental Testing and Verification) program (US-EPA, 2002 and Kimbrough, et al, 2003). The flow and rainfall monitoring data were analyzed using a standardized procedure (Kurz, 2003; Kurz, 2005; Kurz, 2010a) – also described as the Linear Regression (24-Hour Rainfall) method (Kurz, et al, 2014). This procedure was designed to minimize or eliminate human analyst bias by requiring use of objective rules for selecting storm events, etc. The quality of the final prediction is defined by two standard statistical measures: the regression coefficient (r) and the 95% confidence interval of the RDI/I vs. rainfall data sets (Kurz, 2010b). I/I was expressed three ways:
Annual I/I reduction data from the 27 Nashville projects (Kurz 1994; Kurz 1997; Kurz et al, 2004) were augmented with similar data from three project areas in the City of Brentwood, which is a satellite collection system to Nashville. This information was comparable since the both programs used a system approach for rehabilitation design decisions (rehabilitate pipes, manholes, and service laterals), ETV verified flow monitoring equipment, and the same method for analysis (Kurz et al, 2012b). A graph of the results is shown in Figure 1.
The significance of this graph is that it shows a relationship between quantities of rehabilitation lining and annual I/I reduction (Kurz et al, 1998). This can be a valuable tool for a municipality trying to plan and develop a budget for a sewer rehabilitation program. On a rough, aggregate basis, this graph may be interpreted to mean that 6 million gallons of annual I/I can be removed by installation of 1,000 linear feet of lining (or replacement pipe). However, the planner must recognize that there are critical conditions associated with this factor. First, the results are only valid for a program that includes rehabilitation of manholes and service laterals (at least to the property line or easement line), as well as public sewer pipes. Second, the work must be focused in project areas to stop migration of groundwater in the bedding material (a system approach) (Hannen & Hollenbeck, 1984). Third, badly deteriorated areas should receive a minimum level of intensity of 20% to 25% lining (e.g. lining 25,000 lf in a basin of 100,000 lf represents 25% intensity). Fourth, the projects used for this graph achieved an aggregate 50% annual I/I reduction (Kurz et al, 2004). Projects requiring a greater level of reduction may need to use a more conservative rate of I/I reduction effectiveness (i.e. the “low hanging fruit” is the easiest to eliminate).
The following is an example of how a planner may use this graph:
Assume: a small community sewer system with 800,000 lf of public sewer. Flow monitoring has established a level of 1,800 MG of annual I/I. The system needs to eliminate 50% I/I.
50% I/I reduction = 900 MG/year = 50% x 1,800 MG/year
The amount of projected lining = 150,000 ft lining = (900 MG/year)/(6 MG/year/1,000 ft lining)
Assume the total program cost is about $0.6 million per mile of lining. This includes the lining, associated manholes, service laterals, flow monitoring, TV inspection, and engineering. (This is an average based on roughly $ 0.7 million in the Nashville program, and about $ 0.5 million per mile in Brentwood).
Program cost = $ 17.05 million = ($ 0.6 million/mile x 150,000 lf lining / 5,280 ft / mile)
A quick check shows that this projected amount of lining is only about 18.8% of the total system (150,000 lf/800,000 lf) and does not meet the 25% criterion. However, experience has shown that usually a whole system is not uniformly deteriorated (Stevens, 1993). Proper flow monitoring must be employed to isolate badly deteriorated areas as well as basins which do not require treatment. Once they have been identified, the annual I/I in the deteriorated basins should be evaluated and the quantity of rehabilitation projected and checked against the suggested 25% threshold.
While this example has been simplified for purposes of illustration, the costs and quantities for the rehabilitation program in Brentwood demonstrate that this example is quite realistic. The City has rehabilitated the following quantities:
REFERENCES
Conklin, G., Lewis, P. (1981) Evaluation of Infiltration/Inflow Program Final Report, US EPA project No. 68-01-4913, Washington, D.C., February, 1981.
Hannan, P. and A. Hollenbeck, (1984) Impact of Groundwater Migration on Sanitary Sewer Rehabilitation," Water Pollution Control Federation 57th Conference, October 1984
Kimbrough, H. R., Stevens, P. L., Davison, P. J., Frederick, R. M. (2003) Comparing Sewer Flow Meter Performance; WEFTEC-03 Los Angeles, California, October, 2003.
Kurz, G., Stonecipher, P., Dillard, C., Ballard, G., Scott, L., Toosi, C., (1994) Successful Sewer System Rehabilitation, Proceedings of WEFTEC-94, October 1994.
Kurz, G., “Predicting I/I Reduction for Planning Sewer Renewal,” ASCE International Pipeline Congress: Trenchless Pipeline Projects Practical Applications, Boston, Massachusetts (1997).
Kurz, G., Thompson, V., Stonecipher, P., (1998) A Strategy for Projecting Costs in a Sewer Rehabilitation Program, WEFTEC '98, Orlando, FL, October 1998.
Kurz, G., Ballard, G., Burgett, M. and Smith, J. (2003) A Proposal for Industry-Wide Standardization of I/I Calculations, Proceedings of WEFTEC-2003; Los Angeles, California, October 13-15, 2003.
Kurz, G., Ballard, G., Stonecipher, P., (2004) Nashville’s Program Removes 3.2 Billion Gallons of I/I – No-Dig 2004 Conference, New Orleans, LA. March 2004.
Kurz, G. (2005) Standardization of I/I Analysis and Calculations for Objective Comparison of Project Effectiveness, Buried Asset Management Institute (BAMI); Atlanta, GA, 29 August 2005.
Kurz, G., (2010a) I/I Analysis Methodology Including Storm Event Selection, Underground Construction Technology Preconference Workshop, Houston, Texas, January 2010.
Kurz, G., Hamilton, W. (2010b) Confidence Levels for Estimating RDI/I, WEFTEC 2010, New Orleans, LA.
Kurz, G. (2012a) An I/I Program Case History: Nashville Removes 3.6 Billion Gallons of I/I and 137 SSOs, UCT 2012 Sewer Strategies Rehab Workshop, January 25, 2012, San Antonio, Texas.
Kurz, G., Milton, C., Colvett, K., Muirhead, D., (2012b) Sewer Rehabilitation Pays in Brentwood Tennessee, No-Dig 2012, March 11-15, 2012, Nashville, Tennessee.
Kurz, G., Nelson, R., Moisio, S., Varghese, V., (2014) Evaluation of Three Methods to Analyze Sewage Flow and I/I, NASTT-No-Dig 2014, Orlando, FL. April 2014.
Lukas, A., Merrill, M. S., Palmer, R., Van Rheenan, N., (2001) In Search of Valid I/I Removal Data: The Holy Grail of Sewer Rehab? (WERF: Predictive Methodologies for Determining Peak Flows After Sanitary Sewer Rehabilitation), WEFTEC-01 Atlanta, Georgia, October, 2001.
Merrill, M., Lukas, A., Roberts, C, Palmer, R., Van Rheenan, N., (2003) Reducing Peak Rainfall-Derived Infiltration/Inflow Rates –Case Studies and Protocol, WERF, Alexandria, Virginia, 2003
Stevens, P., (1993) Basin Size is the Most Effective Variable an I/I Engineer Can Control. Water Environment Federation 66th Conference, October, 1993.
US Environmental Protection Agency (2002) ETV Joint Verification Statement, NSF Report #02/01/EPAWWF399; Ann Arbor, MI, April 2002.
The second question may be answered easily (in terms of a cost range) once the amount of required sewer rehabilitation is established in response to the first question. Information about the range of unit costs for various sewer rehabilitation materials and processes is readily available. However, (contrary to common thinking) the success of a sewer rehabilitation program for I/I reduction is not dependent on the “right” product or process. Instead, it is more dependent on a sound program strategy, stopping migration of groundwater in the pipe bedding to bypassed defects (a “system” approach), aggressive selection of pipe segments for rehabilitation, and level of intensity of lining. Each of these points is crucial to answering the first question and will be discussed further in this section.
Sewer rehabilitation is often a specific requirement in Commissioner’s Orders and Consent Orders on municipal sewage collection systems with unpermitted overflows. The idea of conducting sewer rehabilitation to decrease the level of I/I (infiltration and inflow) has been a part of EPA’s national strategy since the development of the 201 facility planning process as part of regulations to implement the 1972 Clean Water Act. At times it has been controversial and come under scrutiny. In the early years of the Construction Grants Program, EPA’s “Conklin Report” (Conklin & Lewis, 1981) evaluated work in 18 municipal systems and concluded that generally the efforts to eliminate excessive I/I were not successful. More recently, Lukas, et al (2001) reported on a WERF (Water Environment Research Foundation) study of trends of I/I reduction where 44 utilities had been contacted and data collection efforts were underway. However, the authors pointed out a major problem for accomplishing this important work: “Unfortunately, none of the information between projects was comparable, due to the differences that can exist between two projects. These differences can range from drainage area of the flow meter basin, amount of RDII present prior to rehab, (and) the amount of system rehab performed.”
An evaluation of effectiveness for I/I removal was conducted by WERF (Water Environment Research Foundation) in 2003 (Merrill, et al, 2003) and reported a wide range of levels of effectiveness for twelve RDI/I (rainfall derived infiltration and inflow) reduction projects in six municipal systems. Merrill et al observed that most RDI/I removal projects in the USA are not documented. Second, of those published, few provide good data. Last, data gathering and analysis of data from various programs is hampered by lack of documentation, weaknesses in monitoring techniques, and variation in how information is reported.
I/I reduction information measured in the first Nashville OAP (Overflow Abatement Program 1989-2006) was a significant exception to the problems reported in the WERF study. Early in the program, the managers saw the work in Nashville as an ideal opportunity to develop the necessary documentation and procedures to measure the effectiveness of sewer rehabilitation. Once the procedures were implemented, then rehabilitation effectiveness became a valuable tool to make adjustments in the program and the resulting data were used for modeling and verifying system improvements for overflow reduction. Twenty-seven projects were analyzed with before-and-after flow and rainfall monitoring (Kurz 2012a). This program used ADS flow monitors, which were the only flow monitoring instruments that had been tested by the US-EPA ETV (Environmental Testing and Verification) program (US-EPA, 2002 and Kimbrough, et al, 2003). The flow and rainfall monitoring data were analyzed using a standardized procedure (Kurz, 2003; Kurz, 2005; Kurz, 2010a) – also described as the Linear Regression (24-Hour Rainfall) method (Kurz, et al, 2014). This procedure was designed to minimize or eliminate human analyst bias by requiring use of objective rules for selecting storm events, etc. The quality of the final prediction is defined by two standard statistical measures: the regression coefficient (r) and the 95% confidence interval of the RDI/I vs. rainfall data sets (Kurz, 2010b). I/I was expressed three ways:
- projected 24-hour RDI/I (rainfall dependent I/I)
for a standard return interval storm (5 years in the Nashville OAP)
- projected peak-hour RDI/I for a standard return
interval storm, and
- annual I/I adjusted for variation in annual
rainfall.
Annual I/I reduction data from the 27 Nashville projects (Kurz 1994; Kurz 1997; Kurz et al, 2004) were augmented with similar data from three project areas in the City of Brentwood, which is a satellite collection system to Nashville. This information was comparable since the both programs used a system approach for rehabilitation design decisions (rehabilitate pipes, manholes, and service laterals), ETV verified flow monitoring equipment, and the same method for analysis (Kurz et al, 2012b). A graph of the results is shown in Figure 1.
The significance of this graph is that it shows a relationship between quantities of rehabilitation lining and annual I/I reduction (Kurz et al, 1998). This can be a valuable tool for a municipality trying to plan and develop a budget for a sewer rehabilitation program. On a rough, aggregate basis, this graph may be interpreted to mean that 6 million gallons of annual I/I can be removed by installation of 1,000 linear feet of lining (or replacement pipe). However, the planner must recognize that there are critical conditions associated with this factor. First, the results are only valid for a program that includes rehabilitation of manholes and service laterals (at least to the property line or easement line), as well as public sewer pipes. Second, the work must be focused in project areas to stop migration of groundwater in the bedding material (a system approach) (Hannen & Hollenbeck, 1984). Third, badly deteriorated areas should receive a minimum level of intensity of 20% to 25% lining (e.g. lining 25,000 lf in a basin of 100,000 lf represents 25% intensity). Fourth, the projects used for this graph achieved an aggregate 50% annual I/I reduction (Kurz et al, 2004). Projects requiring a greater level of reduction may need to use a more conservative rate of I/I reduction effectiveness (i.e. the “low hanging fruit” is the easiest to eliminate).
The following is an example of how a planner may use this graph:
Assume: a small community sewer system with 800,000 lf of public sewer. Flow monitoring has established a level of 1,800 MG of annual I/I. The system needs to eliminate 50% I/I.
50% I/I reduction = 900 MG/year = 50% x 1,800 MG/year
The amount of projected lining = 150,000 ft lining = (900 MG/year)/(6 MG/year/1,000 ft lining)
Assume the total program cost is about $0.6 million per mile of lining. This includes the lining, associated manholes, service laterals, flow monitoring, TV inspection, and engineering. (This is an average based on roughly $ 0.7 million in the Nashville program, and about $ 0.5 million per mile in Brentwood).
Program cost = $ 17.05 million = ($ 0.6 million/mile x 150,000 lf lining / 5,280 ft / mile)
A quick check shows that this projected amount of lining is only about 18.8% of the total system (150,000 lf/800,000 lf) and does not meet the 25% criterion. However, experience has shown that usually a whole system is not uniformly deteriorated (Stevens, 1993). Proper flow monitoring must be employed to isolate badly deteriorated areas as well as basins which do not require treatment. Once they have been identified, the annual I/I in the deteriorated basins should be evaluated and the quantity of rehabilitation projected and checked against the suggested 25% threshold.
While this example has been simplified for purposes of illustration, the costs and quantities for the rehabilitation program in Brentwood demonstrate that this example is quite realistic. The City has rehabilitated the following quantities:
- 32 miles of pipe (about 20.6% of the system)
- 1400 manholes
- 320 laterals
REFERENCES
Conklin, G., Lewis, P. (1981) Evaluation of Infiltration/Inflow Program Final Report, US EPA project No. 68-01-4913, Washington, D.C., February, 1981.
Hannan, P. and A. Hollenbeck, (1984) Impact of Groundwater Migration on Sanitary Sewer Rehabilitation," Water Pollution Control Federation 57th Conference, October 1984
Kimbrough, H. R., Stevens, P. L., Davison, P. J., Frederick, R. M. (2003) Comparing Sewer Flow Meter Performance; WEFTEC-03 Los Angeles, California, October, 2003.
Kurz, G., Stonecipher, P., Dillard, C., Ballard, G., Scott, L., Toosi, C., (1994) Successful Sewer System Rehabilitation, Proceedings of WEFTEC-94, October 1994.
Kurz, G., “Predicting I/I Reduction for Planning Sewer Renewal,” ASCE International Pipeline Congress: Trenchless Pipeline Projects Practical Applications, Boston, Massachusetts (1997).
Kurz, G., Thompson, V., Stonecipher, P., (1998) A Strategy for Projecting Costs in a Sewer Rehabilitation Program, WEFTEC '98, Orlando, FL, October 1998.
Kurz, G., Ballard, G., Burgett, M. and Smith, J. (2003) A Proposal for Industry-Wide Standardization of I/I Calculations, Proceedings of WEFTEC-2003; Los Angeles, California, October 13-15, 2003.
Kurz, G., Ballard, G., Stonecipher, P., (2004) Nashville’s Program Removes 3.2 Billion Gallons of I/I – No-Dig 2004 Conference, New Orleans, LA. March 2004.
Kurz, G. (2005) Standardization of I/I Analysis and Calculations for Objective Comparison of Project Effectiveness, Buried Asset Management Institute (BAMI); Atlanta, GA, 29 August 2005.
Kurz, G., (2010a) I/I Analysis Methodology Including Storm Event Selection, Underground Construction Technology Preconference Workshop, Houston, Texas, January 2010.
Kurz, G., Hamilton, W. (2010b) Confidence Levels for Estimating RDI/I, WEFTEC 2010, New Orleans, LA.
Kurz, G. (2012a) An I/I Program Case History: Nashville Removes 3.6 Billion Gallons of I/I and 137 SSOs, UCT 2012 Sewer Strategies Rehab Workshop, January 25, 2012, San Antonio, Texas.
Kurz, G., Milton, C., Colvett, K., Muirhead, D., (2012b) Sewer Rehabilitation Pays in Brentwood Tennessee, No-Dig 2012, March 11-15, 2012, Nashville, Tennessee.
Kurz, G., Nelson, R., Moisio, S., Varghese, V., (2014) Evaluation of Three Methods to Analyze Sewage Flow and I/I, NASTT-No-Dig 2014, Orlando, FL. April 2014.
Lukas, A., Merrill, M. S., Palmer, R., Van Rheenan, N., (2001) In Search of Valid I/I Removal Data: The Holy Grail of Sewer Rehab? (WERF: Predictive Methodologies for Determining Peak Flows After Sanitary Sewer Rehabilitation), WEFTEC-01 Atlanta, Georgia, October, 2001.
Merrill, M., Lukas, A., Roberts, C, Palmer, R., Van Rheenan, N., (2003) Reducing Peak Rainfall-Derived Infiltration/Inflow Rates –Case Studies and Protocol, WERF, Alexandria, Virginia, 2003
Stevens, P., (1993) Basin Size is the Most Effective Variable an I/I Engineer Can Control. Water Environment Federation 66th Conference, October, 1993.
US Environmental Protection Agency (2002) ETV Joint Verification Statement, NSF Report #02/01/EPAWWF399; Ann Arbor, MI, April 2002.