Optimal Risk-Based Life-Cycle Cost Scheduling of Water & Sewer Main Replacements

Transcription

1 Optimal Risk-Based Life-Cycle Cost Scheduling of Water & Sewer Main Replacements

2 The firefighter who was driving the firetruck when it fell into the Valley Village sinkhole caused by a broken water line walks around the scene. (Al Seib / Los Angeles Times / September 8, 2009)

3 Key Questions How long does a pipe last? How many pipes should I replace each year? Where should I start? Which pipes do I replace first? Is now the best time? When is it better to rehab a pipe rather than replace it? When do I replace a pipe that could continue to provide service for another 10+ years?

4 US Water System Under Stress 30 the US water supply network is 30 times the length of the US National Highway System 47 years - average age of a broken water main 240,000 water main breaks/year in USA 7,000,000,000 The total water leaked across the US every day (gallons)

5 We Need to Replace More Pipe GAO Report to US Senate, March 2004

6 Sample Deterioration Curve Pipes do not deteriorate at a constant rate Variables: Material Soil condition Wrapping/Lining External Loading Excavation Activity Corrosion Protection Pipe Depth Pipe Pressure

7 Avg. Pipe Age Keeps Increasing

8 What is the Answer? Practice asset management Prioritize investments based on risk Replace pipes at the optimal time Life Cycle Cost Analysis Detroit, Feb 2014 (30-inch main break)

9 Asset Management Drivers 1. Need to replace, upgrade, expand infrastructure 2. Better manage risk of failure 3. Balance O&M and CIP budgets 4. Maximize value of aging pipelines 5. Spill/overflow tracking & reporting 6. Address EPA fines, orders, consent decrees 7. Enhance public reputation 8. Transfer knowledge of aging/retiring workforce 260 ft

10 Pipe Replacement Without AM Leak of the hour, flavor of the month Hydraulic bottlenecks Inspection results would take the utility It Political more than 300pressures years to replace all of the Oldest city's water pipe pipes atthat is the current rate. most accessible (Southern California Public Radio) July 30, 2014 UCLA, 75,000 gpm (30-in)

11 Case Study East Bay Municipal Utilities District (EBMUD) Oakland, CA

12 East Bay MUD Background SACRAMENTO Hetch Hetchy 90 Miles to East Bay

13 EBMUD Distribution System Serves 1.4 million customers in over 20 cities 4,200 miles of Pipeline 400 mi Transmission Pipe (>20 ) 3,800 mi Distribution Pipe ( 20 ) ~25% in Fault Zone or Landslide Zone

14 Pipeline Inventory Pipeline Length Summary (4,200 mi total) Pipeline Leak Summary

15 80 miles installed < miles >100 years old

16 5-year EBMUD average: 23 leaks/100 miles/year In 2015, a total of 1,155 leaks were recorded (28 leaks/100 miles)

17 Bottom Line The distribution network is aging faster than we can replace it. - East Bay MUD

18 EBMUD Prior Pipe Replacement Program (PRP) + Seismic Hazards > 1.0, more beneficial to replace pipe + Customer Impacts + Claims + Pipe Crew Input + Creek Crossings + Pavement Moratoriums

19 Prior Pipe Replacement Program Advantages: Pipes leaking at the highest rate replaced Cost effective Disadvantages: Mostly pipes in short, cul-de-sacs triggered a high ratio Consequence of failure (COF) not considered in equation Few clustering opportunities

20 Miles of Pipeline Replaced per Year Historical Pipeline Replacement Projected Pipeline Replacement Projected Rebuild Miles Budgeted for 9 In-House District Crews Install ~15 miles/year Which pipes do we replace and how do we prioritize??

21 New Risk-Based Pipe Replacement Program Risk=(Likelihood of Failure) x (Consequence of Failure) LOF COF Pipe Age Total Number of Leaks Leaks per Foot Pipeline Criticality Customer Impacts Environmental Impacts Traffic Impacts Difficulty of Repair

22 More Data => Better Decisions 1. Pipeline data (age, material, etc) 2. Hydraulic model results 3. GIS environmental data 4. Critical facility locations 5. Leak locations (CMMS) 6. Valve locations

23 Likelihood of Failure W 1 LOF 1 E1 + W 2 LOF 2 E2 + W = weight E = exponent LOF = 1-10 ranking of: Infrastructure/Asset Data (Age, Material, Dia.) Hydraulic Condition (Pressure, Flow, & Velocity) Soil Characteristics Seismic Faults Railroad Intersection Traffic Defect History Joint Type Etc.

24 Hazards Increase Likelihood of Failure Fault Lines in EBMUD Service Area

25 LOF Condition Related

26 LOF Hazard Related

27 Likelihood of Failure (LOF) 40% Ratio of TotalLeaks Per Foot 40% Total Number of Leaks 10% TotalLeaks in Past 5 Years 10% Age of Pipe Similar to old Cost Benefit Ratio Future consideration Fault Zones Landslide Zones Liquefaction Zones Leaks per foot favors shorter pipelines Total Leaks favors long pipelines Past 5 Years can reveal accelerated frequency in main breaks Age of pipe does not necessarily correlate with leaks

28 Final LOF

29 Consequence of Failure W 1 COF 1 E1 + W 2 COF 2 E2 + W = weight E = exponent COF = 1-10 ranking of: Flow Supplied (Demand) Population Density Served Critical Facilities Served Outage/Isolation Analysis Traffic Etc.

30 Consequence of Failure (COF) 20% Diameter 15% Backbone Pipelines 10% Critical Customers 10% Consumption 20% Creek Crossing 4% Highway Crossings 3% Railroad Crossings 3% Major Road Crossings 10% Access 5% Slope Pipeline Criticality Engineering Customer Impact Community Affairs Environmental Impact Regulatory Compliance Office Traffic Impact Community Affairs & Repair Crews Difficulty of Repair Repair Crews Attempting to make all stakeholders happy

31 COF

32 COF

33 LOF: Low to High COF: Low to High

34

35 LOF: Low to High COF: Low to High

36 Map of Results Extreme (F) High (D) Moderate (C) Low (B) Negligible (A)

37 Risk Analysis GIS Display

38 Risk Model Results A B C D F Miles 2, % of Network 62% 20% 9% 7% 2.5% At 15mi/yr, would take crews 6.5 years to replace

39 Everett St Cluster Risk Feet % F 7,500 58% D 3,600 28% C 550 4% B 0 0% A 1,100 9% Total 12,750 The goal is for projects to have 80% F & D pipe 2.4 miles (1.3 years per crew)

40 Next Steps LOF COF Risk analysis Criticality analysis Time-based degradation analysis Life cycle cost analysis (LCCA) Formalize a decision model September 9, 2009 Los Angeles - 95-year-old main break

41 Pipe Criticality Trace every pipe to its boundaries (valves) Identify pounds, or shutdown areas Count: Pipe segments Valves Laterals Meters Summation Value Quantifies COF Example output for the pink shutdown area shown:

42 Pipe Criticality GIS map showing shutdown areas Pipes sorted by criticality

43 Reliability Analysis (Decay Curves) Models how pipes degrade over time Predicts service lifespan, survival, and failure probability 6 statistical models: Herz (cohort) Weibull (cohort) Cox regression LEYP (Linear Extended Yule Process) NHPP (Non-Homogeneous Poisson Process) NHMC (Non-Homogeneous Markov Chain)

44 Life Cycle Cost Analysis (LCCA) Financial & Economic tool Uses engineering economic principles Total Ownership Cost Up-front cost O&M costs End of life costs Allows accurate estimates of the lifetime performance of pipes Supports long-term rate planning and investment options Optimum date of replacement

45 LCCA Origins Analysis tool from the transportation & construction sectors Determines best pavement option Metropolitan Transportation Commission (San Francisco Bay Area) Using data-driven economic analysis, including LCCA MTC saved hundreds of millions of dollars by not investing in a project that was not worth the full cost. CalTrans Evaluating the long term cost effectiveness of alternative designs for new and existing pavements. The pavement alternative with the lowest life-cycle is viewed as having the lowest impact to Caltrans even if it has a higher initial cost.

46 Defender vs Challenger Source: AWWA OpFlow December 2016

47 OPTIMAL TIMING OF ASSET REPLACEMENT PRESENT VALUE AGE TO REPLACEMENT Optimal Replacement Age Time Operating Cost Maintenance Cost Risk Cost Total Marginal Cost Capital Cost Total Ownership Cost 1 $0 $ 500 $ - $ 500 $ 8,100 $ 8,600 2 $200 $ 692 $ 433 $ 1,125 $ 4,206 $ 5,531 3 $325 $ 880 $ 854 $ 1,734 $ 2,910 $ 4,969 4 $450 $ 1,062 $ 1,264 $ 2,326 $ 2,264 $ 5,040 5 $550 $ 1,238 $ 1,661 $ 2,899 $ 1,878 $ 5,327 6 $650 $ 1,410 $ 2,048 $ 3,458 $ 1,622 $ 5,730 7 $800 $ 1,578 $ 2,425 $ 4,003 $ 1,440 $ 6,243

48 Results

49 LCCA Results

50 LCCA Results Before Optimal Date

51 LCCA Results Past Optimal Date

52 Overall Life Cycle Budget

53 Optimal (blue) vs. Repair-Only (red) In this example, a major expenditure is required right away, this tells us that we have a large number of pipes that are past their optimal economical age and the utility has a large backlog of pipes that require replacement or renewal.

54 Should I Replace or Just Repair? In the short-to-immediate term, there will be higher costs associated with replacements and renewals (blue line), but over time, by making this initial investment, we see that future costs remain more manageable while the repair only option (red line) continues to climb steeply.

55 Optimal Expenditures by Cost Category Operations cost (blue) and maintenance cost (red) continue to climb at linear rate. However, risk cost begins to climb steeply. This is due to increasing brake rates on pipes as they age.

56 Common Questions After an Asset Failure Captures utility s unique logic Defensible Repeatable Decision Tree

57 Key Indicators (KPI s)

58 Future Enhancements to EBMUD Model LOF factors reduced and looks good Need to fine tune Consequence of Failure Too many factors Include Significant Incident Alerts (SIA) as a factor 50+ customers out of service Property Damage Traffic Impact RCO Notification Media Contact or Presence on Site Incorporate valve criticality, reliability, and LCCA

59 Conclusion: Advantages of Risk-Based LCCA Look at Big Picture (holistic) Prevents over-reaction to leak of the hour or political drivers Incorporates COF in addition to LOF More defensible Easily identifies clustering opportunities using GIS mapping Realistic Improves communication between Engineering, Operations, Finance, Management, Community Affairs, Regulatory Compliance

60 Thanks for attending! Peter Martin, Innovyze Sacramento, CA Tel: (916)