Skip to contentUnited States Department of Transportation - Federal Highway AdministrationSearch FHWAFeedback
Highways for LIFE

Arrow California Demonstration Project: Pavement Replacement Using a Precast Concrete Pavement System

Data Acquisition and Analysis

Data on safety, traffic flow, quality, and user satisfaction before, during, and after construction were collected to determine if this project met the HfL performance goals. The primary objective of acquiring these types of data was to quantify project performance and provide an objective basis from which to determine the feasibility of the project innovations and to demonstrate that the innovations can be used to do the following:

  • Achieve a safer work environment for the traveling public and workers.
  • Reduce construction time and minimize traffic interruptions.
  • Produce a high-quality project and gain user satisfaction.

This section discusses how well the Caltrans demonstration project met the HfL performance goals related to these areas.

Safety

This portion of I-15 is considered one of the busiest in California because of the proximity to Ontario International Airport and the Port of Los Angeles. The higher volume of traffic associated with this stretch of highway is prone to higher crash rates than the statewide average for a similar type of facility.

The project included the HfL performance goal of achieving a work zone crash rate equal to or less than the existing conditions. Caltrans’ crash records before the start of construction, between 2006 and 2009, indicate a crash rate of 1.030 per million vehicle-miles traveled (MVMT). Table 9 lists the breakdown of the crashes recorded before construction.

Table 9. Breakdown of the type and amount of crashes before construction.
Year Fatality Injury Property Damage Only Total
2006 0 22 41 63
2007 0 25 49 74
2008 0 17 32 49
2009 0 6 8 14
Total 0 70 130 200

During construction, from 2009 to 2010, the crash rate dropped by half to 0.490 per MVMT, meeting the HfL goal. Table 10 lists the breakdown of the crashes recorded during construction.

Table 10. Breakdown of the type and amount of crashes during construction.
Year Fatality Injury Property Damage Only Total
2009 0 15 25 40
2010 1 7 17 25
Total 1 22 42 65

The reason for the reduced crash rate during construction was not clear because traditional rehabilitation methods took place concurrently with the installation of PCPS.

The project included the performance goal of achieving an incident rate for worker injuries of less than 4.0 based on the OSHA 300 rate. The contractor indicated that the firm does not maintain records on a project basis and instead aggregates safety performance on an annual basis for all construction projects.

Caltrans does not anticipate achieving a 20 percent reduction in fatalities and injuries in this section of I-15 because this is a rehabilitation project and the geometrics and other major features of the facility will remain the same at completion. Caltrans does anticipate a less tangible impact in reduced pavement maintenance costs because the higher quality PCPS is projected to have three times the lifespan of traditional RSC. This will result in less exposure of maintenance personnel to traffic, which further reduces worker injuries and construction work zone incidents.

Construction Congestion—Effect of Ramp Closures on Travel Times

Introduction

Freeway-to-freeway connectors and ramps were planned to be closed for 55-hour weekend periods over about 35 weekends to accommodate work on the connectors and ramps and within the mainline weaving areas. The I-15 Ontario corridor has consistently high weekday commuter traffic and similar volumes on weekends when leisure travelers from Los Angeles head to and from Las Vegas and resort locations along the Colorado River. In 2009, the annual average daily traffic (AADT) volumes on I-15 near the I-10 interchange were about 214,000 vehicles per day (vpd). About 40 to 50 percent of the I-15 traffic exits to I-10 from each direction. As stated earlier, rehabilitation around this interchange was the most critical for construction congestion and estimating delay times. Therefore, the analysis was performed when the ramp was closed during a 55-hour weekend closure.

To assess the impacts of the ramp closures, the project team conducted a series of travel time runs to determine the additional time required to traverse the detour routes (compared to the normal travel route along I-15) and the total hours of vehicle delay per day that resulted from that detour. Travel time studies were conducted before closure of the northbound I-15 ramps to I-10, on April 24 and 25, 2010. Researchers returned to the site and collected travel times on July 10 and 11, 2010, with the exit ramps from northbound I-15 to eastbound and westbound I-10 closed.

Data Collection

The floating vehicle method was used to collect travel times, which attempts to mimic the typical driving speed of other vehicles along the various roadway segments of the detour route. During the April 2010 data collection, the exit ramp from southbound I-15 to Pomona Freeway (Route 60), which included a dropped lane on I-15, was closed. This closure created delays on southbound I-15. There were no closures in the northbound direction at this time. Data were collected again in July 2010 while the exit ramps from northbound I-15 to both eastbound and westbound I-10 were closed simultaneously, with separate detours for each movement. During this time, the exit ramp from southbound I-15 to Pomona Freeway was open. Data were collected only during daytime hours, since traffic demands were lower at night and thus any effects of the total roadway closure were smaller. Specifically, on Saturdays data were collected from 9 a.m. to noon and 3 to 6 p.m., and on Sundays from 10:30 a.m. to 3 p.m. A minimum of three travel time runs were made over each segment each day in each direction.

Figure 21 identifies key nodes used in the travel time data collection process within the study region. Table 11 identifies the travel distance between nodes and the typical average speed on each segment during the April 2010 data collection. The analysis was based on the desire to compare travel times between northbound I-15 at Cantu-Galleano Road to westbound I-10 at Archibald Avenue and northbound I-15 at Cantu-Galleano Road to eastbound I-10 at Cherry Avenue.

Figure 21. I-15 and I-10 ramp closure analysis region.

Figure 21. I-15 and I-10 ramp closure analysis region.

The normal routes for these segments in April 2010 were simply along the northbound I-15 direct connect ramps to I-10 in either direction. For the northbound I-15 to westbound I-10 detour route in July 2010, travelers continued along I-15 (past closed direct connect ramp) to the 4th Street exit, then along 4th Street westbound to the I-15 South entry ramp, then onto southbound I-15 to the exit ramp to westbound I-10. This detour added 1.8 mi to the segment length. For the northbound I-15 to eastbound I-10 detour, travelers exited at Jurupa Street, followed Jurupa Street eastbound to Etiwanda Avenue, turning left at the traffic signal onto northbound Etiwanda Avenue to the eastbound I-10 entry ramp. This detour added only 0.3 mi to the segment length, but travelers also incurred delay because of the signalized intersection at Jurupa and Etiwanda.

Analysis of the April 2010 data indicated significant travel time variations during each day for the northbound segments, so the data from the two days were analyzed separately. A similar result was obtained for much of the travel time data collected in July 2010. Consequently, those travel times were also analyzed separately by segment. A summary of distances and speeds is shown in table 11

Table 11. Summary of distances and speeds on route segments in analysis region.
Movement Distance
(mi)
Saturday
Average
Speed(mi/h)
Sunday
Average
Speed(mi/h)
April 2010
NB I-15 through movement 7.1 61 67
NB I-15 to WB I-10 6.9 40a 64
NB I-15 to EB I-10 7.3 49a 66
EB I-10 to NB I-15 5.3 61 58
WB I-10 to NB I-15 5.7 65 62
SB I-15 through movement 7.2 46b 46b
July 2010
NB I-15 through movement 7.1 61 67
NB I-15 to WB I-10 Detour 8.7 29 44
NB I-15 to EB I-10 Detour 7.6 41 46
EB I-10 to NB I-15 5.3 35a 50
WB I-10 to NB I-15 5.7 30a 46
SB I-15 through movement 7.2 74 74
aSpeed reductions on northbound I-15 attributed to congestion near 4th Street exit ramp
bSpeed reductions on southbound I-15 attributed to closure of ramp to Pomona Freeway (Route 60)

Overall, average speeds for the primary movements being evaluated decreased when the detours were in place. For the northbound I-15 to westbound I-10 movement, speeds decreased 28 percent and 31 percent on Saturday and Sunday, respectively. This is not surprising because the geometric conditions on the 4th Street detour include two loop ramps with small radius curves. For the northbound I-15 to eastbound I-10 movement, speeds decreased 16 percent and 30 percent on Saturday and Sunday, respectively. This is also not surprising, given that traffic speeds include time stopped at any of the six signalized intersections along this detour route.

For the northbound I-15 to westbound I-10 movement, a comparison of Saturday segment travel times between the two data collection periods is presented in table 12. Sunday segment travel times for this same movement are provided in table 13. Similarly, tables 14 and 15 provide travel times for Saturday and Sunday, respectively, for the northbound I-15 to eastbound I-10 movement.

Table 12. Saturday travel time comparisons for NB I-15 to WB I-10.
Segment Travel Time, minutes
April 2010 July 2010 Difference
Cantu Galleano Road to Route 60 1.4 0.9 -0.5
Route 60 to Jurupa Street 3.3 3.6 0.3
Jurupa Street to I-10 exit ramp 1.0 2.4 1.4
I-10 ramp (April) or 4th Street detour (July) 0.9 6.9 6.0
I-10 entry ramp to Milliken Avenue 0.3 0.3 0.0
Milliken Avenue to Archibald Avenue 1.7 1.5 -0.2
TOTAL 8.6 15.6 7.0

Table 13. Sunday travel time comparisons for NB I-15 to WB I-10.
Segment Travel Time, minutes
April 2010 July 2010 Difference
Cantu Galleano Road to Route 60 0.9 0.9 0.0
Route 60 to Jurupa Street 1.7 1.7 0.0
Jurupa Street to I-10 exit ramp 1.0 0.9 -0.1
I-10 ramp (April) or 4th Street detour (July) 0.8 5.8 5.0
I-10 entry ramp to Milliken Avenue 0.3 0.3 0.0
Milliken Avenue to Archibald Avenue 1.6 1.5 -0.1
TOTAL 6.3 11.1 4.8

Table 14. Saturday travel time comparisons for NB I-15 to EB I-10.
Segment Travel Time, minutes
April 2010 July 2010 Difference
Cantu Galleano Road to Route 60 1.4 0.9 -0.5
Route 60 to Jurupa Street 3.3 2.8 -0.5
Jurupa Street to I-10 exit ramp (April) 1.0 -- 3.2
I-10 ramp (April) 0.5 --
I-10 entry ramp to Etiwanda Avenue (April) 0.8 --
Jurupa Street to I-10 @ Etiwanda Avenue (June) -- 5.5
Etiwanda Avenue to Cherry Avenue 1.7 1.5 -0.2
Total 8.7 10.7 2.0

Table 15. Sunday travel time comparisons for NB I-15 to EB I-10.
Segment Travel Time, minutes
April 2010 July 2010 Difference
Cantu Galleano Road to Route 60 0.9 0.9 0.0
Route 60 to Jurupa Street 1.7 1.6 -0.1
Jurupa Street to I-10 exit ramp (April) 1.0 -- 3.3
I-10 ramp (April) 0.5 --
I-10 entry ramp to Etiwanda Avenue (April) 0.8 --
Jurupa Street to I-10 @ Etiwanda Avenue (June) -- 5.6
Etiwanda Avenue to Cherry Avenue 1.5 1.5 0.0
Total 6.4 9.6 3.2

Overall, the detour for northbound I-15 to westbound I-10 created an average of 5.9 additional minutes of weekend delay per vehicle when the ramp was closed between 9 a.m. and 6 p.m. This corresponds to an average 80 percent increase in travel time. In addition, the detour for northbound I-15 to eastbound I-10 added an average of 2.5 minutes of delay when the ramp was closed. This corresponds to an average 36 percent increase in travel time.

Delay Analysis Results

The total amount of delay incurred over the weekend closure was the sum of the additional travel times (and distance) for traffic normally exiting at the I-15/I-10 interchange (northbound on this particular weekend) and the traffic remaining on I-15 delayed by congestion when it occurred (in the northbound direction, it appears that delays were created at the 4th Street exit during the day on Saturday, but not on Sunday).

Ramp ADTs for the interchange indicate that 54,000 vpd exit northbound. Typically, the number of weekend trips is slightly less than weekday trips. Although it was noted earlier that I-15 is used extensively on the weekends for travel between Los Angeles and Las Vegas, use of the northbound exit ramp from I-15 to I-10 was assumed to not maintain that same level of use on the weekend. Data from freeway facilities in Texas and elsewhere suggest that Saturday volumes are typically 92 percent of AADT values and Sunday volumes are typically 80 percent of AADT values. Consequently, it was assumed that 0.92 x 50,500 = 46,460 vehicles used the exit ramp on Saturday, and 0.80 x 50,500 = 40,400 vehicles used the ramps on Sunday. An additional 10 percent of daily traffic was assumed to normally use the ramps on Friday night (9 p.m. to midnight) and 7 percent was assumed to use the ramps on Monday morning (midnight to 5 a.m.). These two periods added 0.16*50,500 = 8,080 vehicles. For the entire closure period, a total of 94,940 vehicles had to divert to exit to I-10 from I-15 northbound. It was assumed that traffic split equally between eastbound and westbound on I-10.

Based on these assumed ramp volumes, a total of (94,940/2*1.8) + (94,940/2*0.3) = 99,687 additional vehicle-miles traveled were incurred during the weekend closure. For travel time delays created, traffic normally exiting to I-10 westbound incurred (94,940/2*5.9/60) = 4,668 vehicle-hours of additional delay. For traffic normally exiting to I-10 eastbound, a total of (94,940/2*2.5/60) = 1,978 vehicle-hours of additional delay were incurred.

In addition to the delays incurred by exiting traffic, traffic remaining on I-15 through the interchange experienced congestion-related delays during part of the weekend closure. Based on travel time study data, it was assumed that congestion-related delays were experienced only on Saturday during daytime hours (9 a.m. to 7 p.m.). On Saturdays, about 60 percent of ADT traffic occurs during that time period. Assuming a conservative value of 92 percent of AADT traffic occurring on Saturdays, traffic heading northbound beyond the 4th Street exit was approximated as (0.92*0.60*214000/2 – 46460/2*.60) = 45,126 vehicles. During the hours of congestion, table 10 implies that northbound traffic was delayed an average of 4.1 minutes per vehicle at the 4th Street exit. Multiplying this value by the 45,126 vehicles indicates an additional 3,084 vehicle-hours of delay. Table 16 summarizes these numbers.

Table 16. Additional travel distance and delays incurred.
Measure Value
Additional Travel Distance:
  • Exit to I-10 Eastbound
  • Exit to I-10 Westbound
  • TOTAL

14,241 vehicle-miles
85,446 vehicle-miles
99,687 vehicle-miles
Additional Travel Time Delay:
  • Exit to I-10 Eastbound
  • Exit to I-10 Westbound
  • I-15 Northbound
  • TOTAL

1,978 vehicle-hours
4,668 vehicle-hours
3,084 vehicle-hours
9,730 vehicle-hours

Quality

Pavement Test Site

Researchers collected sound intensity (SI) and smoothness test data from a 2,000-ft section of the outermost lane of northbound I-15 beginning at the ramp to I-10 eastbound. Comparing this data before and after construction provides a measure of the quality of the finished pavement.

Sound Intensity Testing

Researchers recorded SI measurements using the current accepted OBSI technique described in American Association of State Highway and Transportation Officials (AASHTO) TP 76-10, which includes dual vertical sound intensity probes and an ASTM-recommended standard reference test tire (SRTT). SI data collection occurred before construction and on the new pavement shortly after opening to traffic. The SI measurements were recorded and analyzed using an onboard computer and data collection system. The two SI probes simultaneously captured noise data from the leading and trailing tire–pavement contact areas. Figure 22 shows the dual probe instrumentation and the tread pattern of the SRTT.

Figure 22. OBSI dual probe system and the SRTT.

Figure 22. OBSI dual probe system and the SRTT.

The average of the front and rear OBSI values was computed to produce the global SI level. Raw noise data were normalized for the ambient air temperature and barometric pressure at the time of testing. The resulting mean SI level was A-weighted to produce the SI frequency spectra in one-third octave bands, as shown in figure 23.

Figure 23. Mean A-weighted SI frequency spectra before and after construction.

Figure 23. Mean A-weighted SI frequency spectra before and after construction.

Global SI levels were calculated using logarithmic addition of the one-third octave band frequencies across the spectra. The global SI value was 108.3 dB(A) for the existing pavement and 102.4 dB(A) for the new pavement. While not meeting the HfL goal of 96.0 dB(A), the 5.9 dB(A) drop in SI is a significant improvement. Overall, each frequency was reduced, indicating the absence of the distinct tone or whine common to concrete pavement with a transverse or overly aggressive surface texture.

Smoothness Measurement

Smoothness testing was done in conjunction with SI testing using a high-speed inertial profiler attached to the test vehicle. The smoothness, or profile, data were collected from both wheelpaths and averaged to produce an IRI value. A low value is an indication of higher ride quality (i.e., smoother road). Figure 24 shows the test vehicle with the profiler positioned in line with the right rear wheel. Figure 25 graphically presents the IRI values for the preconstruction and newly constructed pavement. Two bridge decks were excluded from the data set. The existing distressed pavement had a 225 in/mi value and the new pavement was 66 in/mi. Again, while not meeting the HfL goal of 48 in/mi, the 66 in/mi is much smoother than the existing pavement and a noticeable improvement.

Figure 24. High-speed inertial profiler mounted behind the test vehicle.

Figure 24. High-speed inertial profiler mounted behind the test vehicle.


Figure 25. Mean IRI values before and after construction.

Figure 25. Mean IRI values before and after construction.

User Satisfaction

The HfL requirement for user satisfaction includes a performance goal of 4 or more on a Likert scale of 1 to 7 (in other words, 57 percent or more participants showing favorable response) for the following two questions:

  • How satisfied is the user with the new facility compared with its previous condition?
  • How satisfied is the user with the approach used to construct the new facility in terms of minimizing disruption?

Instead of the HfL questions, Caltrans posed 12 alternative questions to roadway users, asking them to rate their responses or to select an answer category. The survey questions and responses are documented in an Appendix to this report.

Thirty-six survey responses were received and analyzed. Most respondents (23 out of 36 or 63 percent) either agreed or strongly agreed with question 12, indicating they were satisfied with the rehabilitated pavement. This question is similar to the HfL question on how satisfied the user is with the new facility. Twenty-six out of 36 (or 72 percent) of respondents to question 11 supported Caltrans' measures to minimize traffic impact, indicating a favorable response to the HfL question gauging how satisfied the user is with the approach used to construct the new facility in terms of minimizing disruption. Complete survey results are in the appendix.

While the survey was not formatted to a 7-point Likert scale, the favorable response to both HfL questions were 63 percent or better, which meets the goal of 4 or more points on a 7-point Likert scale or the equivalent of 57 percent of users being satisfied.

More Information

Events

Contact

Mary Huie
Highways for LIFE
202-366-3039
mary.huie@dot.gov

Updated: 11/29/2012

FHWA
United States Department of Transportation - Federal Highway Administration