Reliability Guide for the HCM Concepts & Content SHRP 2 Project L08: Incorporation of Travel Time Reliability into the Highway Capacity Manual July 2012 1 Research Objectives The objectives of Project L08 are to: (1) Determine how non-recurrent congestion impacts can be incorporated into HCM procedures (2) Develop methodologies to predict travel time reliability on freeway facilities, urban streets, and freeway-arterial corridors (3) Prepare draft material (Reliability Guide) for possible future incorporation into the HCM 2 1
Anticipated Products Computational procedures for assessing travel time reliability Freeway facilities Urban streets Draft guidebook for possible inclusion in the HCM 3 Introduction to Reliability SHRP 2 Project L08: Incorporation of Travel Time Reliability into the Highway Capacity Manual 4 2
Travel Time Reliability: Basic Concepts Measured by how travel time of a trip varies from one time period to another In other words, reliability is measured as the variability of travel times How long will my trip take today compared to the same trip at the same time on any average day? this implies Travelers have the ability to predict travel time for a trip and to arrive at destination within an on-time window 5 Definitions and Concepts Temporal Dimension Each cell is one analysis period of an analysis segment. Study Period 18:00 15:00 66 66 69 70 63 66 66 66 66 68 68 65 69 63 63 63 68 66 60 67 63 39 64 64 64 70 70 65 38 39 67 67 62 64 68 40 18 37 69 69 66 66 69 64 70 70 63 37 66 14 66 14 66 40 65 65 66 68 68 69 65 39 69 25 63 21 63 16 63 37 69 69 68 66 60 66 67 65 63 38 39 13 64 11 64 37 70 70 64 70 70 68 65 63 38 62 39 40 67 18 67 38 67 67 62 64 68 63 40 63 18 62 37 68 69 40 69 37 68 68 66 66 69 64 70 70 63 37 64 66 14 61 66 14 65 66 40 62 65 61 65 39 61 61 66 68 68 69 65 39 69 25 63 63 21 63 63 16 60 63 37 65 69 67 69 63 63 63 68 66 60 66 67 65 63 38 65 39 13 70 64 11 64 64 37 63 70 67 70 64 64 64 64 70 70 68 65 63 38 62 39 40 67 18 67 38 67 67 66 66 69 70 63 66 66 66 62 64 68 63 40 63 18 62 37 68 69 40 69 37 68 68 66 68 68 65 69 63 63 63 64 70 37 64 14 61 14 65 40 62 65 61 65 39 61 61 68 66 60 67 63 39 64 64 69 39 25 63 21 63 16 60 37 65 69 67 69 63 63 63 64 70 70 65 38 39 67 67 66 66 66 65 38 65 13 70 11 64 37 63 70 67 70 64 64 64 62 69 64 70 68 63 40 66 18 66 37 66 69 69 66 68 68 63 62 40 18 38 67 67 64 68 70 65 37 69 14 63 14 63 40 63 65 65 68 66 63 63 62 68 40 37 68 68 69 60 39 67 25 63 21 39 16 64 37 64 69 69 64 70 64 61 65 62 61 39 61 61 66 70 65 65 38 38 13 39 11 67 37 67 70 70 62 64 63 63 60 65 67 63 63 63 68 68 63 40 62 18 40 37 18 69 38 69 67 67 64 70 65 70 64 63 67 64 64 64 63 37 63 14 62 14 68 40 40 65 37 65 68 68 69 39 64 25 61 21 65 16 62 37 61 69 39 69 61 61 66 65 63 38 63 13 60 11 65 37 67 70 63 70 63 63 68 63 65 62 70 40 64 18 63 38 67 67 64 67 64 64 63 63 62 68 40 37 68 68 64 61 65 62 61 39 61 61 63 63 60 65 67 63 63 63 65 70 64 63 67 64 64 64 Reliability Reporting Period Spatial Dimension Study Section 6 3
What Causes Unreliable Travel? Traffic Control Devices 1 Daily/Seasonal Variation Special 2 Events 3 determine Planned Emergencies Physical Capacity 4 interacts with Demand Volume lowers capacity and changes demand Roadway Events Weather 5 can cause Base Delay ( Recurring or Bottleneck ) can cause Incidents 6 can cause Event Related Delay Work Zones 7 n = Source of Congestion Total Congestion 7 What Features Should Reliability Performance Measures Have? Related to measuring variability of travel times Can be communicated to wide variety of audiences Useful in specific types of analyses that use HCM analyses as their basis Capable of discerning trends, i.e., sensitive to changes in the underlying sources of unreliable travel Many potential measures have these features can we limit? 8 4
Performance Measures are Defined by the Travel Time Number of Trips (in Thousands) Free Flow 400 Mean 350 Semi-Standard Deviation 95 th Percentile 99 th Percentile 300 250 Misery Time 200 150 Buffer Time 100 Planning Time 50 Standard Failure 0 Deviation Measure 4.5 9.5 14.5 19.5 24.5 29.5 Travel Time (in Minutes) 9 Recommended Performance Measures Reliability Performance Measure Core Measures Planning Time Index 80 th Percentile Travel Time Index Semi Standard Deviation Failure/On Time Measures Supplemental Measures Standard Deviation Misery Index (Mod) Definition 95 th percentile Travel Time Index (95th percentile travel time divided by the free flow travel time) 80th percentile travel time divided by the free flow travel time One sided standard deviation, with reference point at free flow rather than mean Percent of trips with space mean speed less than 50 mph; 45 mph; and 30 mph Usual statistical definition The average of the highest five percent of travel times divided by the free flow travel time 10 5
Use Cases for Travel Time Reliability 1. Assess existing reliability of facilities. 2. Predict future reliability of facilities. 3. Identify options to improve reliability. 4. Estimate benefits of improvements. 5. Prioritize reliability improvement options. 6. Set local standards for acceptable reliability. 7. Improve demand modeling. 11 Use Cases for SHRP2-L08/HCM 1. Assess existing reliability of facilities. 2. Predict future reliability of facilities. 3. Identify options to improve reliability. 4. Estimate benefits of improvements. 5. Prioritize reliability improvement options. 6. Set local standards for acceptable reliability. 7. Improve demand modeling. 12 6
Integrating Safety Into the Project Development Process James A. Bonneson Kittelson & Associates July 2012 Highway Safety & Design What is Safety? Safety in the Project Development Process 1
Is This Road Safe? What is Safety? Is This Road Safe? What is Safety? There are many deficient design elements that make it risky, but is that the same as safe? Safety is most clearly defined in terms of crash frequency and severity Crash frequency combines risk and exposure (e.g., AADT) Is a Yes or No answer sufficient? There are degrees of safety Estimating the expected crash frequency is a more reliable indicator (as opposed to the previous year s crash count) A road with 14.9 cr/yr is safer than one with 15.2 cr/yr 2
Substantive Safety Kinds of Safety Expected crash frequency (i.e., long run average number of crashes per year) Substantive safety is a continuous variable Useful for comparing alternatives Nominal Safety A road that conforms to the agency s policy, guidelines, and warrants is nominally safe A road either is, or is not, nominally safe Kinds of Safety Safety Comparison (NCHRP Report 480) 3
Quantifying Safety Highway Safety Manual Safety Prediction Model C = base crash rate volume length CMF Crash Modification Factor (CMF) CMF used to estimate change in crashes due to a change in geometry (CMF = C with /C without ) Example: CMF add bay = 0.70 C no bay = 10 crash/yr C with bay = C no bay CMF add bay = 7 crash/yr Safety Indicators Highway Crashes U.S. Alaska Diff. Fatal crashes/100 mvm 1.02 1.09 +7% Fatalities/100 mvm 1.11 1.17 +5% Ped. fatalities/100,000 pop. 1.38 0.84 34% 56 fatalities/yr on Alaska highways Data for Year 2010 4
Highway Crashes Contributing Factors Driver Age, gender, skill, fatigue level, alcohol, etc. Vehicle Type, age, maintenance, etc. Roadway Geometric design, traffic control devices, lighting Environment Day/night, precipitation, fog, temperature, etc. HSM Focus Roadway ROADWAY and ENVIRONMENT Highway Crashes HSM Chapter 3 5
Safety in the Project Development Process Geometric Design Impacts Safety Environment Traffic operations Right of way Construction costs Goal Find best design that offers minimal total impact Cost benefit analysis is one way to evaluate Project Development Process Stages Planning and programming Alternative identification and evaluation Preliminary design Final design Where to Evaluate Safety? All four stages How to Incorporate Safety? Adhere to design criteria and guidelines Conduct explicit safety evaluation, as appropriate 6
Alternatives Evaluation Stage Elements Traffic analysis Evaluate crash data, identify existing patterns Preliminary alternatives Develop design schematic Adhere to design criteria and guidelines Evaluate safety of alignment alternatives Technical studies Environmental impact study Traffic impact analysis Cost benefit analysis Which Alternative Is Best? 7
Alternatives Evaluation Stage One Approach (no information on safety impact) Impact Area Measure Units Impact Existing Alt. 1 "Benefit" Safety Compliance n.a. partial full ** Operations Motorist delay veh-h/day 290 37 1872 "Cost" R.O.W. Property acquired acres 0 3 6 Access Accessibility Lost access pts. 0 5 7.5 Environment Area restored acres 0 1 2.3 Construction Cost $/1000s 0 1000 1000 Benefit/Cost Ratio 1.8 Alternatives Evaluation Stage Better Approach (with information on safety impact) Impact Area Measure Units Impact Existing Alt. 1 "Benefit" Safety Crashes crash/yr 16 4 1200 Operations Motorist delay veh-h/day 290 37 1872 "Cost" R.O.W. Property acquired acres 0 3 6 Access Accessibility Lost access pts. 0 5 7.5 Environment Area restored acres 0 1 2.3 Construction Cost $/1000s 0 1000 1000 Benefit/Cost Ratio 3.0 8
Elements Preliminary Design Stage Preliminary design conference Identify complex or atypical conditions Preliminary design development Develop design to 30% level Adhere to design criteria and guidelines Evaluate safety effect of key design elements Evaluate safety of features with complex conditions Design Exception Request Evaluate safety effect of deviation from criteria Preliminary Design Stage Key Design Elements that Influence Safety Design speed Lane width Shoulder width Median width and type Bridge width Structural capacity Horizontal alignment Vertical curvature Grade Stopping sight distance Cross slope Superelevation Vertical clearance Length of speed change lane Horizontal clearance Guardrail length Checked items can be evaluated using HSM 9
Safety Conscious Design AASHTO Guidance Consistent adherence to minimum [design criteria] values is not advisable Minimum design criteria may not ensure adequate levels of safety in all situations The challenge to the designer is to achieve the highest level of safety within the physical and financial constraints of a project Highway Safety Design and Operations Guide, 1997 Point / Counterpoint Point Knowledge of a substandard road condition exposes an agency to a suit Accepting design exceptions, or not adding all known safety features, will expose an agency to suit Counterpoint Courts do not expect agencies to upgrade every road as conditions change Quantifying safety impacts and documenting findings shows prudence and yields defensible decisions 10
Additional Resources AASHTO HSM Website http://www.highwaysafetymanual.org FHWA Highway Safety Manual Support http://safety.fhwa.gov/hsm/ TRB Highway Safety Performance Committee http://www.safetyperformance.org FHWA CMF Clearinghouse http://www.cmfclearinghouse.org Questions Comments? 11
Urban Street Safety and Operation: Evaluating Performance Today and Tomorrow James A. Bonneson Kittelson & Associates, Inc. July 2012 Overview Background on Urban Street Evaluation Urban Street Operation Urban Street Safety Combined Operation and Safety Evaluation Urban Street Reliability Evaluation 2 1
Background Observations Newly released Highway Safety Manual (HSM) provides a means to quantify road safety Reliability of service (e.g., travel time reliability) has emerged as an important descriptor of operational performance Reliability is influenced by events (e.g., weather, crashes) that occur over the year Decisions to improve operation and reliability also influence safety Decisions to improve safety also influence operation and reliability Presentation Basis SHRP 2 Project L08: Incorporation of Travel Time Reliability into the Highway Capacity Manual (HCM) Scope Prediction (as opposed to measurement) to inform system design, operation, and management Background Reliability Travel time distribution is used to define system reliability Distribution of performance during year Reflects effect of safety and weather on operations Large variability (i.e., spread) suggests an unreliable operation Weather Incident 2
Urban Street Operation Traditional Operational Evaluation Predictive methods used for evaluation Highway Capacity Manual Synchro, etc. Analysis period Design hour Predict performance measures for design hour Assume: if proposed facility adequately serves design hour, then it will provide adequate service during most other hours in year Operation influenced by design and control choices Intersection control mode: signal vs. stop control Left-turn mode: permitted, protected-permitted, protected-only Change interval duration Median type: raised curb, TWLTL, no median (undivided) Turn bay presence Urban Street Operation Operational Evaluation HCM Approach 2010 Highway Capacity Manual Urban street facilities Urban street segments Signalized intersections Two-way stop-controlled intersections All-way stop-controlled intersections Roundabouts Interchange ramp terminals Urban street segments performance measures Travel time Travel speed Stop rate Through delay 3
Urban Street Safety Traditional Safety Evaluation Historically, safety prediction methods not available Highway Safety Manual is now available for safety prediction Safety audits used to subjectively identify safety issues Adherence to minimum criteria used to ensure nominal safety of alternatives Urban Street Safety Safety Evaluation HSM Approach Predictive method used for safety evaluation Highway Safety Manual Analysis period One year Predict performance measures for whole year Prediction for selected time periods during year not available Safety influenced by design and control choices Intersection control mode: signal vs. stop control Left-turn mode: permitted, protected-permitted, protected-only Change interval duration Median type: raised curb, TWLTL, no median (undivided) Turn bay presence 4
Urban Street Safety Safety Evaluation HSM Approach 2010 Highway Safety Manual Urban street facilities Urban street segments Signalized intersections Two-way stop-controlled intersections All-way Stop-Controlled Intersections Roundabouts Interchange ramp terminals (forthcoming) Urban street performance measures Expected crash frequency Fatal and injury, single vehicle Fatal and injury, multiple vehicle Property damage only, single vehicle Property damage only, multiple vehicle Combined Operation and Safety Evaluation Consideration of Operations and Safety Effect of design and control choices can be quantified Effect on operational performance Effect on safety performance HSM makes combined, quantitative assessment possible 5
Combined Operation and Safety Evaluation Issue 1 Time Domain of Evaluation Operations: design hour Safety: one year How well can the design hour reflect year-round performance? Safety Evaluation Hour 24 Traditional Operational Evaluation Use HCM to estimate travel time for design hour. Hour 24 Use HSM to estimate expected crash frequency for year. 0 0 365 Day of Year 0 0 365 Day of Year Combined Operation and Safety Evaluation Issue 2 Operations and Safety Not Independent Design and traffic control decisions to improve operation... May also influence safety Do not necessarily coincide with improvements in safety Design and traffic control decisions to improve safety... May also influence operation Do not necessarily coincide with improvements in operation 6
Combined Operation and Safety Evaluation Whole-Year Analysis A way to resolve these issues Time domain issue Evaluate operations for every hour of the year Or, perhaps just selected hours of each day (e.g., pm peak period) Interaction of operations and safety issue Quantify effect of crashes on operational performance Number of lanes closed Duration of lane closure Quantify effect of congestion and speed on safety Expected crash frequency Proportion severe crashes Combined Operation and Safety Evaluation Whole-Year Analysis Computation time issue Evaluation of one, one-hour period takes fraction of a second In batch mode, can evaluate all one-hour periods during one year in just a few minutes Demand variation during year Volume of each one-hour period is varied to reflect hour of day, day of week, and month of year patterns Capacity variation during year Capacity of each one-hour period is varied to reflect effect of external events on safety and operation Weather events Crashes Non-crash incidents (e.g., breakdown, debris in road, etc.) Infrastructure failure (e.g., signal malfunction, dangerous potholes) Train presence 7
Combined Operation and Safety Evaluation Whole-Year Analysis Forecasting weather events Hour 24 s Use historic weather records to determine... Number of events per month Probability of event (rain or snow) by month Average Duration by month Use Monte-Carlo methods to randomly determine which days of month have an event, when it starts, when it ends, whether it is rain or snow s Note: hours and location may change each year but travel time distribution Whole-Year Operational for whole-year Evaluation is fairly stable year after Safety year Evaluation Use HCM to estimate travel time for every hour of year.* r r r Hour 24 Use HSM to estimate expected crash frequency for year. 0 0 s r r r 365 Day of Year 0 0 365 Day of Year Combined Operation and Safety Evaluation Whole-Year Analysis Forecasting crash events Hour 24 Use HSM to estimate expected crash frequency and severity Use Monte-Carlo methods to randomly determine which hours have a crash and where crash is located on street Procedure increases chance of crash during poor weather or high volume Note: hours and location may change each year but travel time P distribution Whole-Year Operational for Evaluation whole-year is fairly stable year Safety after Evaluation year P Use HCM to estimate travel time for every hour of year.* P P P F P P P Hour 24 Use HSM to estimate expected crash frequency for year. 0 0 365 Day of Year 0 0 365 Day of Year 8
Combined Operation and Safety Evaluation Whole-Year Analysis - Summary Evaluate every hour of interest during the year Predict effect of cashes on operations Predict effect of congestion and speed on crash frequency Combined Operation and Safety Evaluation Whole-Year Analysis One Last Thought Feasible with 2010 HCM and 2010 HSM But, prediction of expected crash frequency for selected time periods during year not available (yet) So, we can t get distribution of hour expected crash frequency (yet) Whole-Year Operational Evaluation Safety Evaluation Hour 24 Use HCM to estimate travel time for every hour of year.* Hour 24 Use HSM to estimate expected crash frequency for year. 0 0 365 Day of Year 0 0 365 Day of Year 9
Combined Operation and Safety Evaluation Issue 3 Work Zones and Special Events Design and control decisions made to accommodate work zones When, and how long, to accommodate work zone also a factor All can be evaluated using whole-year analysis Combined Operation and Safety Evaluation Payoff of Whole-Year Analysis More complete picture of overall facility performance Distribution of performance during year Reflects effect of safety and weather on operations Travel Time Weather Incident Lane Hours Blocked Incidents per Day Severe crash, many lanes 10
Combined Operation and Safety Evaluation Payoff of Whole-Year Analysis Implement strategies to reduce incidents or mitigate their impact Example: add shoulder (disabled vehicle refuge) No change in operation during design hour, but Average travel time decreases and operation is more reliable Travel Time Lane Hours Blocked Combined Operation and Safety Evaluation Payoff of Whole-Year Analysis Implement strategies to increase speed and reduce delay Example: convert from protected-only to protected-permitted lefts Travel time decreases during design hour (+ other hours), but Crashes increase, Average travel time is unchanged, and operation is less reliable Travel Time Incidents per Day Disclaimer: Trends are illustrative. One design or control change may not produce such dramatic distribution shifts but a combination of changes could produce such shifts. 11
Questions or Comments? Urban Street Reliability Evaluation SHRP 2 Research Project L08 Overview Framework Illustrative Analysis Capabilities 24 12
Project L08 Overview SHRP 2 Project L08 Incorporation of Travel Time Reliability into the Highway Capacity Manual Objectives Determine how non-recurrent congestion impacts can be incorporated into HCM procedures Develop methodologies to predict travel time reliability on freeway facilities and urban streets Prepare a guide with that is suitable for potential inclusion in a future update of the HCM 2010 Project schedule Ends in Fall 2012 Framework Development Goals for Urban Streets Quantify the effect of non-recurring congestion sources Weather Demand variation Incidents Work zones Special events Minimize the amount of required input data Assemble a set of nationally-representative default values Terms Scenario a unique combination of volume and traffic control conditions for one analysis period (e.g., 1 hour) Study period one or more consecutive scenarios during a day (e.g., 3-hour period from 3:30 pm to 6:30 pm) 26 13
Framework Manual Software Work Flow Start with the input data used to evaluate an urban street facility using the 2010 HCM Urban Streets methodology Enter the data in the urban streets engine and save it to a file If desired, enter and save data for each work zone or special event Read the file and use it as a basis for scenario generation Work day-by-day, analysis-period-by-analysis-period in chronologic order through the year... Predict weather events Predict incident events Adjust speed and saturation flow rate based on events Adjust demand volumes using hourly, weekly, monthly factors Save one revised file for each analysis period Submit each revised file to the urban streets engine Collect performance measures and compute reliability statistics 27 Framework Flow Chart 28 14
Framework Input Data Nearest city Functional class Analysis period duration (0.25 hr or 1.0 hr) Study period duration (e.g., 7:00 am, extend for 3 hours) Reliability reporting period (e.g., 1/1/2011, extend for 365 days) Days of week considered (Su, M, Tu, W, Th, F, Sa) Crash frequency by Segment Intersection If work zone or special event present Operating period (e.g., 4/1/2011, extend for 30 days) Crash frequency adjustment factors Framework Work Zones and Special Events Each is dealt with as a unique case for a unique time period Identify volume, geometry, and signal timing for each case Include specific changes due to work zone or special event» Lane closures» Alternative lane assignments» Special signal timing Create one urban streets engine input file for each case Establish operating period (e.g., 4/1/2011, extend for 30 days) Determine crash frequency adjustment factors Traffic demand changes Not predicted If analyst can estimate demand shifts, they can be reflected in volumes entered in the input file 15
Urban Streets Reliability Engine Welcome 31 Scenario Generation Set Up 32 16
Scenario Generation Set Up 33 Scenario Generation 34 17
Performance Summary 35 Performance Summary Performance Measures By Direction EB/NB WB/SB By System Component Facility Segment By Performance Measure Travel time Travel speed Stop rate Running time Through delay Examples EB direction 36 18
Performance Summary Facility Travel Time (s) PTI = 2.9 Note Maximum travel time 800 s Two crashes at same time but different intersections Planning Time Summary Statistics Scenario evaluation interval: 1 Average: 120.38 5th percentile: 91.29 Base free-flow speed, mi/h: 40.78 Standard deviation: 38.34 10th percentile: 92.45 Base free-flow travel time, s: 60.20 Skewness: 7.72 80th percentile: 139.79 Median: 109.73 85th percentile: 143.76 Number of obs.: 3120 95th percentile: 172.59 37 Performance Summary Facility Travel Speed(mi/h) Note 10 percent of analysis periods have LOS D, E, or F F E D C B A Summary Statistics Scenario evaluation interval: 1 Average: 21.52 5th percentile: 14.22 Base free-flow speed, mi/h: 40.78 Standard deviation: 4.29 10th percentile: 16.06 Base free-flow travel time, s: 60.20 Skewness: -0.58 80th percentile: 25.91 Median: 22.37 85th percentile: 26.25 Number of obs.: 3120 95th percentile: 26.89 38 19
Urban Streets Reliability Methodology Framework Overview of next three parts Scenario Generation Facility Evaluation Performance Summary Illustrative Analysis Capabilities 39 Illustrative Analysis Capabilities Alternatives Analysis Compare base condition to alternative condition Analysis Scenarios Work zones and special events Alternative start dates and durations Alternative lane closures and signal timing strategies Weather Examine operational effects of strategies that reduce weatherrelated crashes or crash severity (i.e., snow removal, resurfacing) Incidents Examine operational effects of strategies that reduce incident duration Evaluate benefit of providing shoulder for stalled vehicle refuge Design or Operation Evaluate alternative signal timing plans Evaluate intersection lane allocations or segment geometry 40 20
Closure Questions or Comments? 21