Pavement Management Techniques to Improve Network Performance

Transcription

1 Pavement Management Techniques to Improve Network Performance Jeremy Gregory and CSHub Research Team Ohio Transportation Engineering Conference October 10, 2017

2 The US is not sufficiently investing in its ailing road system 21% of the US highways are in poor condition Slide 2

3 Governments are being forced to do more with less Feb 14, 2013 Governments Look for New Ways to Pay for Roads and Bridges Slide 3

4 Infrastructure spending is at an all-time low Slide 4

5 Life cycle cost analysis is a key element of addressing infrastructure funding gap Reduce life cycle costs by 50% by 2025* Areas of focus Resilience Innovation Life cycle costs Performance standards *$4.6 trillion needed in infrastructure investment by 2025 $2 trillion is unfunded Slide 5

6 Asset management allocation tools are critical to economically-efficient infrastructure Use asset management best practices to prioritize projects and improve the condition, security, and safety of assets while minimizing costs over its entire life span. Slide 6

7 Tools for economically-efficient management of pavement networks Competitive Paving Prices Life Cycle Cost Analysis Asset Management Slide 7

8 Does the presence of competition between material substitutes impact pavement material prices? Statistical analyses of Oman BidTabs* data using these parameters: Paving volume on job Annual spending on paving Number of bidders on job (intra-industry competition) Share of spending on AC vs. PCC (inter-industry competition) Presence of price adjustment clauses 5-Year Average Percent Spending on AC * , 47 states, 298k pay items, 164k jobs Slide 8

9 Significant factors and their effect on bid paving prices Significant Factors Market size (annual spending) Project size (paving volume) Inter-industry competition (percent spending on concrete) Intra-industry competition (number of bidders) Presence of price adjustment clauses Impact on Asphalt $ Impact on Concrete $ N/A Key: = an increase in (or presence of) the factor decreases paving prices + = an increase in the factor increases paving prices more markers indicates more significance. Slide 9

10 Viable competition is an opportunity For an average state spending 5% on concrete... Asphalt Concrete 10% 15% 20%...increasing to this level of concrete spending... 8% 6% 4% 2% 0%...decreases paving prices by this amount Slide 10

11 Tools for economically-efficient management of pavement networks Competitive Paving Prices Life Cycle Cost Analysis Asset Management Slide 11

12 LCCA Life-cycle cost analysis: Method for evaluating total costs of ownership Transform individual pavement expenditures over time into total life-cycle cost B Real Cost A B Life-Cycle Cost A 2 B B B 2 0 A A A 0 Present Value Slide 12

13 Pavement design is iterative, but tools are not: Accelerated feedback more testing, more improvement Design Proposal & Context Layers Traffic Climate 10.0 JPCP w/ 1.25 Dia Dowels 6.0 Agg Subbse Subgrade N Analyze Using ME Design Principles Adequate Performance Y Final Design 8.0 JPCP w/ 1.25 Dia Dowels 6.0 Agg Subbse Subgrade Develop Lifecycle Bill of Activities Evaluate LCCA / LCA Slide 13

14 Pavement design is iterative, but tools are not: Accelerated feedback more testing, more improvement Design Proposal & Context Layers Traffic Climate 10.0 JPCP w/ 1.25 Dia Dowels 6.0 Agg Subbse Subgrade Analyze Using ME Design Principles Final Design 8.0 JPCP w/ 1.25 Dia Dowels 6.0 Agg Subbse Subgrade Develop Lifecycle Bill of Activities MIT research aims to integrate these activities Evaluate LCCA / LCA N Adequate Performance Y Slide 14

15 CSHub created linkage between design tools and evaluation Pavement has associated activities; activities translate to cost & impact Cash flows 7.0 JPCP w/ 1.25 Dia Dowels 6.0 Agg Subbse Subgrade Pavement Design & Context MEPDG Pavement Performance Performanceto-activity modeling Bill of activities Material quantities Construction activities Maintenance timing Logistics LCCA Model LCA Model Magnitude Timing Life cycle inventory Material & energy inflows Emissions outflows Slide 15

16 CSHub created probabilistic cost estimates for entire life-cycle Materials Construction Use Maintenance End-of-Life Decisions long before construction Long life-cycle Cost Uncertainty in initial construction costs User Uncertainty in pavement deterioration & future costs Agency Uncertainty & Risk Uncertainty in excess fuel consumption costs Uncertainty in traffic delay costs Construction Materials & Construction UseOpe r ati omaintenance n EOL Slide 16

17 CSHub conducted LCCAs for a wide range of scenarios 4 Locations 3 Traffic Levels CO: Dry freeze Rural local street/highway Rural state highway Urban interstate AZ: Dry no freeze MO: Wet freeze Several framing conditions Pavement designs Maintenance schedules Pavement design and maintenance schedules developed by Applied Research Associates (ARA), Inc FL: Wet no freeze Design life Analysis period Slide 17

18 Life cycle matters Future costs can be significant Total life-cycle costs for a state highway in Florida Future maintenance and rehabilitation costs 53% Initial construction costs 47% Flexible pavement design developed by Applied Research Associates (ARA), Inc,: AADTT 1k/day; 4 lanes; Wet-no-freeze-FL; FDOT-based rehabilitation schedule; Analysis period = 50 years. Slide 18

19 Life cycle perspective alters relative competitiveness Initial Cost % difference ~0% difference A B Real Price LCC A B 0 (millions $) (millions $) 0 Slide 19

20 User costs matter in pavement life cycle costs Excess fuel consumption drives user costs M & R 16% User Cost 22% Traffic Delay 4% Initial 62% Excess Fuel Consumption 96% Pavement design developed by MNDOT; concrete pavement with asphalt shoulders Slide 20

21 Context matters Costs vary with location, traffic level, & pavement design Interstate, rigid design Rehab costs 2% Initial costs 98% Local highway, rigid design Rehab costs 11% Initial costs 89% Interstate, flexible design State highway, flexible design Rehab costs 21% Initial costs 79% Rehab costs 53% Initial costs 47% Slide 21

22 Probabilistic analysis provides insight on relative risks Frequency Probabilistic Life-Cycle Cost Including Price Projections 10% 100% 75% 50% 25% B 8% mean 0% Net Present Value of LCC (million $) A 30% Analysis can answer: What is statistical significance of difference between alternatives? What are key drivers of difference? Slide 22

23 There is significant variation in initial costs Event Occurrence as a Percentage 30% 20% 10% 0% Distribution of Unit Price of Concrete for Pavement Projects 12 months of project bid data from the CalTrans database What is driving variation in costs? Unit Price of Concrete ($/CY) Source: Caltrans Slide 23

24 Capture drivers of initial cost and variation through statistical models 8 Concrete material prices highly dependent on quantity used on job Log Unit-Price ($/CY) Other factors can also be captured in statistical models Log Quantity (Cubic Yards) Slide 24

25 Key learnings from LCCA research Define functional equivalence for design life & analysis period Pavement designs Maintenance and rehabilitation activities Create models to capture variation in initial and future costs Drivers may include quantity, location, competition Price projections accomplished through analyses of historical prices Conduct probabilistic analysis Captures insight on relative risk Context matters Costs vary with location, traffic level, & pavement design Slide 25

26 Tools for economically-efficient management of pavement networks Competitive Paving Prices Life Cycle Cost Analysis Asset Management Slide 26

27 FHWA has issued new performance management rules due to MAP-21 FHWA motivation: improve decision-making through performance-based planning and programming Key elements of asset management plans: Life cycle planning Risk management analysis 10-year financial plan Slide 27

28 Pavement network performance management process Set targets Measure performance Implement strategies to meet targets % Good % Poor Pavement Management System Which strategies? At what cost? How to allocate funds to obtain best performance at lowest cost? Slide 28

29 Many approaches to allocate funds Pavement Segment Pavement Condition Index A 45 B 47 C 51 D 52 E 56 F 62 G 67 How to prioritize which segments to repair? Will targets be met? Which strategies should be used? Many short term fixes? Few long-term fixes? An optimization modeling approach is required to answer these questions: Performance-Based Planning Slide 29

30 Goal of MIT asset management research: improve allocation decisions Key elements of pavement management systems for performance-based planning Objective: prioritize projects that maximize performance and minimize cost Current & historical pavement network data Projection of network performance Allocation of resources Key Considerations Analysis period Available maintenance, rehabilitation, and reconstruction activities Uncertainty in prices and deterioration Optimization algorithm Optimization metrics Scale of network Slide 30

31 Key elements of the MIT allocation model Key Considerations Analysis Period Conventional models Short analysis periods (~5 yr) MIT model Long analysis period (20+ yr) Available MRR activities Uncertainty in prices & deterioration Prescriptive decision rules for limited MRR alternatives Deterministic data & analysis Dynamic decision rules for diverse set of MRR alternatives Risk analysis including uncertainty Slide 31

32 Example with VA interstate network Expected roadway expenditures over 50 year analysis period Budget=$67m/yr MRR activities including concrete overlays Low expenditures per unit area Medium expenditures per unit area High expenditures per unit area Objective: minimize trafficweighted roughness Slide 32

33 Allocation strategy diversifies after 50 years Current Composition of System Expected Composition in Year 50 AC AC PCC PCC Fixed budget ($67m) Slide 33

34 Diversity of MRR alternatives is a means to enhance network performance Cumulative Probability Single material network performs worse than a diversified network Diversified Network Concrete only Asphalt only Key Considerations: Analysis period Available maintenance, rehabilitation, and reconstruction activities Uncertainty in prices and deterioration 0 ($67m budget for scenarios) Average Traffic-Weighted IRI (in./mi.) over 50-year analysis period Slide 34

35 Longer analysis period improves network performance 1.5 Network Traffic weighted IRI (m/km) Segment analysis period = 5 yr 30 yr 15 yr Network analysis period (year) Analysis period influences mix of fixes Slide 35

36 Key considerations for performance-based planning Analysis period Explore impacts of longer segment and network analysis periods Cost and deterioration data Include uncertainty Evaluate data pedigree (time period and source of training data) MRR alternatives Include concrete overlays Slide 36

37 Thank you