CDOT SPF Development and 10 Years of Application. A Practical Approach...

Transcription

1 CDOT SPF Development and 10 Years of Application A Practical Approach Jake Kononov, P.E. Ph.D. Bryan K. Allery, P.E. National SPF Summit Chicago 2009

2 In Order to Manage Safety Effectively, We Need to be Able to Measure it

3 How To Measure Safety? Accident Rate is Still the Most Common Measure of Safety

4 Rate = #Acc x 1,000,000 AADT x 365 x Length Let s s Examine Its Application

5 Year # Acc AADT Rate , , , , Average Rate = 2.28 Gambling Introduced In Miles After Gambling Year # Acc AADT Rate , , , , Average Rate = 1.24

6 Before Gambling Average Rate = 2.28 Highway Alignment and Typical Cross-Section have not Changed After Gambling Average Rate = 1.24

7 After the Introduction of Gambling, the % of Accident Involving Alcohol increased 500%. Is Drinking and Driving in Concert with Gambling Good for Safety? Probably not, but if Accident Rates are Used as a Measuring Device One Would have to Conclude that it is.

8 Colfax Ave MP ,692 Crashes 181,927 AADT Crashes 150,548 AADT Arapahoe Road MP I-25

9 Colfax Ave MP ,692 Crashes 181,927 AADT 1990 Between 1990 and Crashes 150,548 AADT AADT Increased from 150,548 to 181,927 Total Accident Rate Increased by 75% Injury and Fatal Rate Increased by 29% Arapahoe Road MP

10 C-470 Platte Canyon MP I-25 MP Crashes 36,011 AADT Crashes 77,680 AADT

11 Platte Canyon MP I-25 MP Between 1990 and 2004 AADT Increased from 36,011 to 77, Total Accident 58 Crashes Rate Increased 308 Crashes by 146% 36,011 AADT 77,680 AADT Injury and Fatal Rate Increased by 60%

12 Clearly the Rate is Changing with AADT In Order to Understand how the Crash Rate is Changing, We need to Develop a Relationship between Safety and Traffic Exposure

13 This Relationship is Reflected by, Safety Performance Function (SPF)

14 Calibration of Safety Performance Functions in Rural and Urban Environments

15

16

17 Rural Mountainous 2-Lane Highway LOSS/SPF Graph

18 14 Rural 2-Lane 2 Highway APMPY ,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 AADT

19 14 Rural 2-Lane 2 Highway APMPY ,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 AADT

20 14 Rural 2-Lane 2 Highway E = μ No Overdispersion Poission Error Structure 1.5σ 8 APMPY ,000 AADT Corresponds to An Expected Accident Frequency of 8 Accidents Per Mile Per Year 1.5σ ,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 AADT

21 Segment 1 Segment 2 Segment 3

22 Urban 6-Lane 6 Freeway SPF Graph

23 Urban 6-Lane 6 Freeway Each Point Represents Accident Frequency for a Segment of Roadway for a Single Year 100 APMPY ,000 40,000 60,000 80, , , , , , ,000 AADT

24 Urban 6-Lane 6 Freeway Fit a Curve to the Data Using Neural Networks 120 APMPY Mean ,000 40,000 60,000 80, , , , , , ,000 AADT

25 Urban 6-Lane 6 Freeway E > μ Overdispersion Present N-B B Error Structure 1.5σ APMPY Mean ,000 AADT Corresponds to An Expected Accident Frequency of 60 Accidents Per Mile Per Year 1.5σ 0 20,000 40,000 60,000 80, , , , , , ,000 AADT

26 California 6-Lane 6 Freeway

27 California 8-Lane 8 Freeway

28 Texas 8-Lane 8 Freeway

29 120 CO6L (Total) APMPY (Total) ( 56Acc)( 1,000,000) ( 120,000)( 365)( 1) Rate = = 1.28acc/ mvmt APMPY ,000 60,000 80, , , , , , ,000 AADT

30 The Following Assumptions Typical of the Urban Freeway Environment were used to Estimate LOS Boundaries: DHV (Design Hourly Volume) =10 % of AADT for AADT <130,000 DHV (Design Hourly Volume)=8% of AADT for AADT>130,000 PHF (Peak Hour Factor) = 0.9 %Truck during peak period = 2% Terrain Level Lane Width = 12 ft Shoulder Width > 6 ft Interchange spacing =1 Interchange/Mile

31 120 CO6L (Total) APMPY (Total) 100 Sub-Critical Zone Transitional Zone Super-Critical Zone 80 APMPY 60 Super-Critical Density d > 45 pc/mi/ln v < 52 mph 40 d = 16 pc/mi/ln v = 70 mph 20 Critical Density d = 24 pc/mi/ln v = 69 mph 0 40,000 60,000 80, , , , , , ,000 AADT Sigmoid Or Dose-Response Curve

32 CO6L (Total) APMPY (Total) 120 LOS B or Better LOS C LOS D and E LOS F Super-Critical Density APMPY Critical Density 0 40,000 60,000 80, , , , , , ,000 AADT

33 CO6L (Total) APMPY (Total) 120 LOS B or Better LOS C LOS D and E LOS F APMPY Avg Acc Rate 0.64 Avg Acc Rate 0.75 Avg Acc Rate 1.23 Avg Acc Rate 1.45 Super-Critical Density Critical Density 0 40,000 60,000 80, , , , , , ,000 AADT

34 CO6L (INJ+FAT) APMPY (INJ+FAT) LOS B or Better LOS C LOS D and E LOS F APMPY Avg Acc Rate 0.23 Avg Acc Rate 0.23 Avg Acc Rate 0.35 Avg Acc Rate Super-Critical Density 10 5 Critical Density 0 40,000 60,000 80, , , , , , ,000 AADT

35 Relationship Between the Number of Lanes and Safety.... Jake Kononov, P.E. Ph.D. Bryan K. Allery, P.E. National SPF Summit Chicago 2009

36 Urban Freeways Lane Comparison for Total Accidents Lane APMPY Lane 50 Safety Dividend 6-Lane to 8-Lane 4- Lane Safety Dividend 4-Lane to 6-Lane ,000 50,000 75, , , , , , , , ,000 AADT

37 Urban Freeways Lane Comparison for Total Accidents Lane APMPY Lane Lane ,000 50,000 75, , , , , , , , ,000 AADT

38 Urban Freeways Lane Comparison for Total Accidents The Volume Doubled, But the Accident Frequency Tripled Lane APMPY 123 Accidents Lane Accidents 4- Lane ,000 50,000 75, , , , , , , , ,000 AADT

39 Number of Lanes and Conflict Points C C n n 4-Lane Freeway (2 Potential Conflicts) 6-Lane Freeway (7 Potential Conflicts) 8-Lane Freeway (16 Potential Conflicts) = = f ( n) = n( n 1) f ( n) for n = n( n 1) + = 2 n! 3!( n 3)! for n > 2

40 Urban Freeways Lane Comparison for Total Accidents Lane APMPY Lane Lane ,000 50,000 75, , , , , , , , ,000 AADT

41 As the Number of Lanes Increases, the Degree of Freedom for Things to go Wrong also Increases

42 C 5 = 30 C n = f n! 5! ( n) = n( n 1) + = 5(5 1) + = 30 3!( n 3)! 3!(5 3)! C 2 = 2 C 3 = 7 C C C 2 3 = = 2+ 3 f ( n) = n( n 1) f ( n) = n( n 1) + = = 9 = 2(2 1) = n! = 3(3 1) + 3!( n 3)! 3! = 7 3!(3 3)! 2

43 9 is Less Than 30

44 Safety Performance of Signalized Intersections. Jake Kononov, P.E. Ph.D. Bryan K. Allery, P.E. National SPF Summit Chicago 2009

45 26 Year Old Male 164 Lbs Blood Pressure 65 Diastolic 110 Systolic Does he have Hypertension?

46 43 Year Old Male 264 Lbs Blood Pressure Diastolic Systolic Does he have Hypertension?

47 Does anyone here have Training in Internal Medicine or Cardiology?

48 Urban 4 Leg 6-Lane Signalized Fully Actuated Mast Arm Signal Layout Lighted

49 1997 to 2000 Mainline AADT = 51,000 Side Road AADT = 9,000 4 Year Crash History PDO = ( Inj + Fat ) Per Year INJ = (Total ) FAT = 0 Do We Have a Safety Problem?

50 2001 to 2004 Mainline AADT = 63,000 Side Road AADT = 12,000 4 Year Crash History PDO = ( Inj + Fat ) Per Year INJ = (Total ) FAT = 0 Do We Have a Safety Problem?

51 The Safety Performance Function of an Intersection can be viewed Mathematically as a 3-Dimensional 3 Response Surface, where: # Crashes/Year = ƒ( ADT Mainline, ADT ADT Mainline ADT Side Road )

52

53 CDOT SPF Development and 10 Years of Application A Practical Approach Jake Kononov, P.E. Ph.D. Bryan K. Allery, P.E. National SPF Summit Chicago 2009