Strategic Approach for Enhancing Reliability of Traffic Signals in an Urban Environment Using Big Data Analytics-Case Study from Washington, DC, SS Dey, D Patterson, BO Pérez, R Jain, H Alexander

Tags: BBS, intersections, generator, traffic signal, Power Outages, generators, Federal Highway Administration, DDOT, power outage, DC, cost savings, American Association of State Highway and Transportation Officials, Backup Power, VA, Manual on Uniform Traffic Control Devices, National Oceanic and Atmospheric Administration, deployment, Center for Infrastructure and Transportation Studies, Matthew Hill, dark signal, opportunity cost, Geographical Distribution, intersection, traffic control, Stephanie Dock, signal heads, signal power, types of intersections, Rensselaer Polytechnic Institute, Tammie Jones, Michael Kinney, maintenance, National 50 Association of City Transportation Officials, Michael Miller, Traffic Signals, preventive maintenance program, operating costs, Thomas Moore, Craig Moran, Robert Waller, Mo Zhao, portable generators, backup systems, Rahim F. Benekohal, Montgomery County Department, Homeland Security Grant Program, Traffic Signal Energy, Negin Askarzadeh, Robert Lewis, Michael Keatley, communications system
Content: 1 Strategic Approach for Enhancing Reliability of Traffic Signals in an Urban Environment
2
Using Big Data Analytics ­ case study from Washington, D.C.
3
4
5 Soumya S. Dey, P.E., PMP (Corresponding Author)
6 District Department of Transportation
7 55 M Street, SE, Washington, DC 20003
8 Tel: 202-671-1369; Fax: 202-727-1078; Email: [email protected]
9
10 Diane Patterson
11 District Department of Transportation
12 55 M Street, Southeast, Suite 600
13 Washington, DC 20003
14 [email protected]
15
16 Benito O. Pйrez, AICP
17 District Department of Transportation
18 55 M Street, Southeast, Suite 500
19 Washington, DC 20003
20 Telephone: (202) 671-1597; Fax: (202) 727-1078; [email protected]
21
22 Rahul Jain
23 District Department of Transportation
24 55 M Street, Southeast, Suite 600
25 Washington, DC 20003
26 Telephone: (202) 741-5337; Fax: (202) 727-1078; [email protected]
27
28 Harvey Alexander
29 District Department of Transportation 30 2000 14th St NW
31 Washington, DC 20009
32 Telephone: (202) 671-1495; Fax: (202) 727-1078; [email protected]
33
34 Serdar Senyurt
35 M.C. Dean, Inc
36 22980 Indian Creek Drive
37 Dulles, VA 20166
38 Telephone: (571) 237-0487; Fax: (703) 421-4670; [email protected]
39
40
41 Word count: 5284 word text + 8 tables/figures x 250 (each) = 7,284
42
43
44
45
46 Submission Date: August 1, 2016
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
2
1 ABSTRACT 2 This paper discusses strategies implemented by the District Department of Transportation (DDOT) to 3 enhance the reliability of its signal system. The reliability measures discussed in this paper include 4 deploying backup generators and installing uninterruptable Power supply (UPS) battery backup systems 5 (BBS) for critical intersections to minimize disruptions during power outages. The measures have 6 yielded beneficial results. The benefits were apparent not only during severe weather events such as the 7 2012 derecho and Hurricane Sandy, but also under normal circumstances. The paper quantifies benefits 8 by analyzing data on signal outage and run-time of generators and BBS. Each deployment has 9 implications not only on traffic flow and safety, but also on city resources (a dark signal at a critical 10 intersection typically requires a response from police or traffic control officers). The paper makes the 11 business case for investing in systems that enhance system operations through increased reliability. The 12 paper also discusses how big data analytics can help jurisdictions devise their strategy on technology 13 selection and deployment. 14 15 Keywords: Reliability, signal system, arterial operation, performance measurement, data based decision 16 making.
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
3
1 INTRODUCTION
2 The District Department of Transportation (DDOT) plans, designs, builds, maintains, and operates the
3 transportation infrastructure within Washington, DC. DDOT is unique: it has characteristics of both a
4 state and local/municipal department of transportation. DDOT's transportation assets are valued at over
5 $50 billion and are utilized by 650,000 residents, 500,000 commuters, and an average of 125,000 visitors
6 on a daily basis. DC's travel characteristics are very multi-modal. It has the second highest transit usage
7 in the nation, a large percentage of walk and bike trips, and an increasing trend in shared transportation
8 services (such as bikeshare, traditional and point-to-point car share, etc.).
9
Washington, DC has primarily an arterial system comprised of 1,600 signalized intersections; less
10 than one percent of roadway mileage is freeways. Because of this, the efficiency of the transportation
11 network is dictated by the reliability of the signal system.
12
This paper discusses strategies that DDOT has put in place to enhance its signal reliability. Signal
13 power is an important component of reliability, as dark signals and in-flash malfunctions represent safety
14 hazards and cause traffic delays. In addition, electrical surges can damage signal equipment. Over the past
15 several years, DDOT has invested in in portable backup generators and uninterruptible power supply
16 (UPS) battery backup systems (BBS) for traffic signals, and plans to expand the backup network while
17 maximizing efficiency and agency funding.
18
The authors employ a data-based approach to analyzing outages in the DC signal system,
19 demonstrating the benefits realized as a result of backup deployments, and provide a foundation for
20 prioritizing future backup installations. Practitioner telephone and online surveys, in addition to a
21 literature review, were conducted to determine the current state of practice and deployment strategies. The
22 paper provides a case study and builds on the small but growing body of guidelines for how jurisdictions
23 can best determine the configuration and deployment of power backups to increase the reliability of their
24 signal systems.
25
26 STATE OF THE PRACTICE
27
In order to learn about current practices regarding traffic signal power backup systems, DDOT
28 conducted a literature review, in addition to online and telephone surveys during the summer of 2016; the
29 results are shown below in Figure 1.
30
50.0% 45.0% 40.0% 35.0% 30.0%
Respondent Jurisdiction Locations
Figure 1 (A) Online Survey Respondent Jurisdiction Locations
25.0%
20.0%
15.0%
10.0%
5.0%
0.0%
FL VA CA MD WA IL GA BC NJ IN OH ID NC WY NS CO TX MN MS MI AZ ME
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
Outages are a Significant Issue
60.9%
39.1%
Have Backup Power 6.1% 27.3% 66.7%
4 Figure 1 (B & C) Online Survey ­ Jurisdictions where outages are significant issues & Jurisdictions with backup power
Yes No
Yes No No, but plan to
Backup Types
Other Generators
Figure 1 (D) Online Survey - Backup Power Types in Use
Generators and UPS
UPS
0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0%
50
45
40
35
30
25
20
15
10
5
0
Have Power Backups
Dedicated funding for backup maintenance
Dedicated funding for capital expansion
Dedicated funding for life cycle costs
Studies on benefits of backups
Figure 1 (D) Online Survey ­ Backup Power Funding & Study Results Yes No
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
5
Popular Prioritization Criteria
Other
54.8%
Outage History
23.8%
Figure 1 (E) Online Survey ­ Most Popular Backup Power Deployment Prioritization Criteria
Congestion
40.5%
Traffic Volume
71.4%
0
5
10
15
20
25
30
35
Backup Power Prioritization Criteria
Complex intersection geometry
Preemption
Congestion Crash history Distance from signal shop Left turn bays, multiple left turns Intersections with other ITS Assets (CCTV, etc) Interstate ramp approach Evacuation routes/corridors
Proximity to essential services (hospitals, schools, airports etc.) Proximity to military bases Proximity to other signalized intersections Request from police, others Reversible lanes Railroad intersections Signal warranted by MUTCD
Figure 1 (E) Literature Review, Online Survey, & Telephone Survey Power Backup Prioritization by Intersection Criteria Types
Large power loads Number of approaches Outage/power issue history Pedestrian signals/high pedestrian volume
Significant roads Speed limits on approaches Traffic volume Truck route
1 FIGURE 1 LITERATURE REVIEW & SURVEY RESULTS
2
3 Prioritization Factors
4
Figure 1 also contains a list of criteria based on our surveys and literature review that agencies
5 take into consideration for prioritizing backup power deployment.
67
While providing signal backup power is becoming common practice in the U.S., very few are
8 employing robust data analysis to determine the appropriate backup configuration. The only requirements
9 in the Manual of Uniform Traffic Control Devices (MUTCD) for power backup installations are
10 intersections with light rail, railroad preemption, or flashing light signal systems [1]. While many
11 jurisdictions have criteria lists for backup deployment, CalTrans and the British Columbia Ministry of
12 Transportation and Infrastructure applied more methodical approaches by creating scoring systems for
13 their intersections [2].
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
6
1
The Guidelines for Traffic Signal Energy Back-Up Systems developed for the New York State
2 DOT took this scoring system further by analyzing outages and vehicle crash data to develop their
3 prioritization matrix (7). Mo Zhao et al. created statistical models using outage history and power load
4 data from signalized intersections in Indiana to determine the best BBS battery capacities for various
5 types of intersections. Their analysis was based on three different budget scenarios and can be used to
6 ensure that the right resources are deployed to the intersections with the greatest need [3].
7
8 Types of Backup Power Sources
9
The most thorough review of all traffic signal backup power technologies for can be found in the
10 2009 Guidelines for Traffic Signal Energy Back-Up Systems [2]. These guidelines describe the
11 advantages and disadvantages for each type, as well as Case Studies of jurisdictions that have used them.
12
There are two main types of backup power sources: fixed and portable. Table 1 summarizes the
13 types of technologies currently in use.
14
15 TABLE 1: TYPES OF SIGNAL POWER BACKUPS IN USE [4, 5, 6, 7, 8]
Type Description
Advantages
Disadvantages
Local
Jurisdictions
Using
Portable Backups
Portable - gasoline or diesel - mobile
- deployment
- Baltimore
Generators powered
- easy to refuel
personnel costs
city, MD
- manually
- greater output enables - deployment delay - Montgomery
deployed
operations with both
- exhaust/emissions County, MD
- good option for incandescent and LED - noise
- Arlington,
medium to long- signal heads
- risk of theft
VA
term outages
- can recharge BBS
- cannot address short- - Alexandria,
term outages
VA
- higher maintenance - Hampton, VA
and operating costs - Newport
News, VA
- VDOT
Vehicle- - converts a
- mobile
- deployment
Testing in
Mounted vehicle's power
- easy to refuel
personnel costs
Montgomery
Power
from DC to AC
- greater output enables - deployment delay
County, MD
Inverters - good option for operations with both
- exhaust/emissions
medium to long- incandescent and LED - noise
term outages
signal heads
- risk of theft
- can recharge BBS
- higher maintenance
and operating costs
Fixed/Permanent/Uninterruptable Power Supply (UPS) Backups
Battery
- can directly run - addresses both voltage - reliant on LED signal - Anne Arundel
Backup
signal operations reductions & power
heads for maximum County, MD
Systems and recharge on outages
capacity
- Baltimore
(BBS)
grid power, or
- automatic deployment - some battery models city, MD
remain on standby - can address longer-
have risk of bleeding - Howard
until outage occurs term outages if employ - occupies cabinet/curb County, MD
- good option for battery swapping
space
MD SHA
short to medium-
- battery capacity
- Arlington,
term outages
affected by low
VA
temperatures
- Alexandria,
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
7
VA
- Chesapeake,
VA
- Virginia
Beach, VA
- Newport
News, VA
- VDOT, VA
Fixed
- natural gas,
- addresses both voltage - monthly fees for
- Suffolk, VA
Generators propane, or diesel reductions & power
natural gas
- Testing in
powered
outages
- direct hit to cabinet MD SHA
- good option for - automatic deployment potentially very
all power outages - natural gas option
dangerous
- integrated with provides unlimited
BBS for initial
power supply
activation
- diesel & propane can
be refueled
Renewables - wind, solar or a - extends life of BBS
- wind power causes - Testing in
hybrid of both
- can sell power back to noise and potentially Montgomery
- directly mounted utility company
bird kills
County, MD
onto traffic signal - reduced carbon
- increases height,
- Testing in
poles
footprint
weight, and surface MD SHA
- integrated with - increased
area of pole
BBS & power grid functionality with new - reduced functionality
to recharge
48 volt cabinets
in urban, dense areas
batteries during
- available technology
outage
can only power low
loads
1
2 DDOT'S POWER BACK UPS
3 DDOT implements contingency measures to minimize the impacts of power outages to signals when they
4 do happen, through the deployment of emergency back-up generators and BBS devices.
5
6 Emergency Back-Up Generators
7
In 2005, DDOT received a $1.6M competitive grant from the Department of Homeland
8 Security's (DHS) Urban Area Security Initiative (UASI) program [9] to procure 200 emergency
9 back-up portable generators and in 2006 DDOT replaced all cabinets on its evacuation routes
10 with ones with generator ports. The focus of the grant was emergency preparedness; hence the
11 initial deployment was on cabinets along evacuation routes [10]. DDOT cabinet specifications
12 were subsequently modified and now call for all future cabinet installations to be outfitted with
13 the DDOT standard generator interface.
14
The generator is built around a Honda 12.0 in3, 4-stroke gasoline engine with electric start, and
15 housed in a custom aluminum enclosure designed with input from DDOT personnel and can support the
16 weight of an average sized man. The generator dimensions are 19.5 inches wide, 22.5 inches deep, and
17 22 inches tall. Each generator weighs 140 pounds, requires two people to set in place, and is designed
18 with a custom mechanism to secure it to the side of a traffic control cabinet. Because a generator is
19 attached to the side of a cabinet when deployed, theft has not been a problem. The generators have a fuel
20 tank capacity of approximately 3.5 gallons, have a rated output of 3.0 kVA, and can typically power an
21 intersection for five to six hours.
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
8
1
While generators are deployed, the technicians' standard operating procedure is to check them
2 every two hours and refuel as needed. Traffic cabinets outfitted with generator ports include an automatic
3 transfer switch that isolates it from the PEPCO power grid when the generator is running, and
4 automatically reconnects the intersection to PEPCO when power is restored. Reliability of generators and
5 the power switch over is nearly 100%.
6
Between 2008 and 2015 DDOT deployed the generators 739 times with a total run time of 15,601
7 hours (Figure 2).
8
Number of Signal Power Outages & Generator Deployments 500
Figure 2 (A) Number of Outages vs. Generator Deployments
400
Number (Outages/Deployments)
300
200
100
0
2008 2009 2010 2011 2012 2013 2014 2015 Year
Outages
Generator Deployment
Signal Power Outage & Generator Run
14000
Time
12000
10000
8000
Hours
6000
4000
2000
0 2008 2009 2010 2011 2012 2013 2014 2015
Year
Hours of Power Outage
Generator Run Time
9 FIGURE 2 OUTAGE VS. GENERATOR DEPLOYMENT 10 11
Figure 2 (B) Outage Hours vs. Generator Run Time
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
9
1 Battery Back Up (BBS)
2
Since 2011, DDOT has invested in BBS at nearly 400 signalized intersections (approximately
3 25% of system). The program was kick-started using funding from American Reinvestment and Recovery
4 Act (ARRA) [11]. The BBS can power a typical intersection for approximately six hours. DDOT now
5 allocates funds annually to increase BBS coverage. Additionally, as of 2013 BBS are part of DDOT's
6 signal specifications so new signals or signal modifications include this backup as part of the construction
7 cost. DDOT has been adding approximately 100 BBS units each year.
8
The initial BBS deployment has been based on a list of "critical intersections" that was developed
9 by the metropolitan police department (MPD) in consultation with DDOT. The BBS communication
10 system is not fully integrated with the traffic signal system, and staff are not able to monitor battery
11 activation and duration in real-time. Thus, a technician is unaware of any issue with a signal if its BBS is
12 activated until the battery power starts to run down and the signal goes into flash mode, or if the battery's
13 power runs out entirely and the signal goes dark.
14
For this paper, DDOT analyzed data from the 131 BBS (approximately 44% of the current
15 installed base) that could be accessed from a centralized location from 2012-2014. The remaining BBS
16 devices store data locally. Figure 3 (A) below shows the total and networked base of BBS, the number of
17 times BBS was utilized, and the number of unique intersections at which they were activated. It also
18 shows the total time during which signals were powered by BBS. The system-wide BBS usage was
19 estimated based on the ratio of networked to total BBS for the specific calendar year, using the data from
20 the networked BBS. It was interesting to note that the average BBS usage per event was significantly
21 lower than generators.
22
Statistics
2012
2013
2014
Figure 3 (A) Summary of
Total BBS Deployed
140
199
272 BBS Usage
Networked BBS
51
94
131
2012-2014
% Networked
36%
47%
48%
# of times networked BBS utilized
70
230
665
# of unique intersections (networked)
33
71
96
Total time networked BBS used
13:57 75:34 240:34
Estimated system-wide hours
38
159
499
Estimated # of times BBS used (system-wide)
192
487
1381
Average BBS usage per event (minutes)
1.1
19.5
21.6
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
Distribution of Usage Time for UPS
15-30 min, 1%
5-15 min, 4%
31-60 min 2%
1-5 min, 12%
1-4 hrs 3%
4-12 hrs, 1% 12-24 hrs 1%
<30 min, 86% <1 min, 69%
Over 24 hrs 6%
10 Figure 3 (B) Distribution of BBS Time Length Usage
1 FIGURE 3 SUMMARY AND TIME DISTRIBUTION OF BBS USAGE
2
3
An analysis of the usage data, depicted in Figure 3(B), shows that 69% of BBS usage was for less
4 than 1 minute; 12% between one and 5 minutes. This implies that BBS usage is being pre-dominantly
5 triggered by brown outs or lack of clean electricity. Without a BBS, the signals would have gone into the
6 flash mode. The presence of BBS is ensuring uninterrupted signal operation.
7
8 ANALYSIS OF SIGNAL POWER OUTAGES
9 DDOT's signal maintenance data is stored in a centralized work order system called the Traffic
10 Infrastructure Maintenance Management Software (TIMMS). The work order system maintains an
11 inventory of all the traffic signal and intelligent transportation system (ITS) assets. It also has time-
12 stamped entries for all work assigned to the contractor and the contractor's response, resolution, and
13 materials used. TIMMS is primarily used as a work order management tool, but also provides
14 information about the state of the signal system, the contractor's performance, and provides the basis for
15 billing and verifying monthly invoices. TIMMS has information dating back to 2000 and contains a
16 wealth of information about DC's signal system. This paper analyzed the data in TIMMS to provide
17 insight into signal system.
18
Though power outages represent only 5% of malfunctions in Washington DC, they account for
19 87% of signal performance disruption time, as shown in Figures 5 (A) and (B). Moreover, from a public
20 safety standpoint, a power outage has a higher degree of risk for the traveling public than other
21 malfunction types. Therefore, to enhance the reliability of the signal system, this paper focuses on
22 addressing signal power outages.
23
Power outages are most pronounced during weather emergencies. As an example, Hurricane
24 Isabel in 2003 left one-third of the signals in Washington, DC dark over an extended period of time (up to
25 one week). More recently, Hurricane Sandy and the derecho in 2012 also caused significant power
26 outages [12,13], as shown in Figure 4 below. A derecho is a widespread, long-lived wind storm that are
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
11
1 associated with bands of rapidly moving showers or thunderstorms variously known as bow echoes,
2 squall lines, or quasi-linear convective systems [14].
3
Derecho
Event:
Derecho Dark Signal Locations by Outage Time:
Derecho
Dates of Impact: June 29 ­ July 3, 2012
Total # of signals out: 94
Total Hours of Outage: 4906 hours
Average outage time: 52 hours
Percent of 2012 outage: 66%
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
12
Event: Hurricane Sandy
Hurricane Sandy Dark Signal Locations by Outage Time:
Dates of Impact: October 29-31, 2012
Total # of signals out: 19
Total Hours of Outage: 160 hours
Average outage time: 8.5 hours
Percent of 2012 outage: 2%
1 FIGURE 4 SUMMARY OF SIGNAL POWER OUTAGE DURING HURRICANE SANDY
2 AND THE DERECHO
3
4
Power outages are a major source of disruption to the signal system during non-weather related
5 events as well. As an example, power outages in downtown DC on June 13, 2008 affected 30 signals, 6
6 Metrorail stations, and 5,000 customers, including the White House [15].
7
8 Annual Trends in Power Outages
9
Over the last eight years (2008-2015), signals have been dark because of power outages 2,403
10 times for a total of 50,895 hours. This translates to an average of 300 outages per year for an average
11 annual outage length of 6,361 hours. This accounts for 10% of the operating hours of the signal system..
12
Figure 5 (C) shows the total number and hours of outages due to power failures. The years 2010,
13 2011, and 2012 had particularly high numbers for power outage-related signal disruptions.
14
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
13
1 Distribution of Outage Durations
2
As shown in Figure 5 (D), approximately 4% of the outages in DC are under 1 hour. The typical
3 response time to deploy a generator is one hour. So a "generator only" solution will not be able to
4 address 4% of the outages in DC. A typical generator can run an intersection for five to six hours on one
5 tank of fuel. This accounts for approximately 45% of the outages that are between 1 to 6 hours. For
6 outages beyond six hours, generators need to be refueled to ensure continued operation. Sixteen (16)
7 percent of the outages in DC last for more than 24 hours.
8 4% 1% 1%
Signals in Flash
Figure 5 (A). Causes of Signal Malfunctions
19%
45%
Power Outage Twisted Signals
25%
LED replacement Signal head replacement
Pole replacement
5%
Cabinet Replacement
Figure 5 (B). Outage Time for Different Types of Signal Malfunctions
Hours of Power Related Signal Outages Number of Power Related Signal Outages
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
Disruptions to Signals Due to Power Failures
14000
500
12000
400
10000
8000
300
6000
200
4000
2000
100
0
0
2008 2009 2010 2011 2012 2013 2014 2015
Outages
Year Hours of Power Outage
Distribution of Outage Duration
>24 hrs
16% 1-3 hrs
12-24 hrs
20%
12%
< 1 hr 4%
6-12 hrs 23%
3-6 hrs 25%
14 Figure 5 (C). PowerRelated Signal Outages Figure 5 (D). All Signal Outages by Length of Time
Distribution of Outages by Time of Day
30%
25%
20%
15%
10%
5%
0%
Overnight AM (5:30 to Mid Day (10:00 PM (14:30 to Evening (19:00
(midnight to
10:00)
to 14:30)
19:00)
to Midnight)
5:30)
Figure 5 (E). Signal Outages by Time of Day
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
15
2008-2016 No. of Outages by Month 160 140 120 100 80 60 40 20 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
2008 2009 2010 2011 2012 2013 2014 2015 2016
Figure 5 (F). Number of Outages by Month of Year
Hours
2008-2016 Outage Length by Month
9000 8000 7000 6000 5000 4000 3000 2000 1000 0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
2008 2009 2010 2011 2012 2013 2014 2015 2016
Figure 5 (G). Duration of Outages by Month of Year
1 FIGURE 5 OUTAGE CHARACTERISTICS
2
3 Time of Outages
4
Figure 5 (E) shows the distribution of outages by time of day. Approximately 75% of the outages
5 occur from 7 PM to 5:30 AM and at mid-day. Analysis of the time of outages is important because the
6 impact of an outage will be different based on the time of outage; an outage at 1 AM will have
7 significantly lower impact than one at 1 PM. The other advantage of such an analysis is that it helps an
8 agency assess whether resources are available during periods of high outages. By adjusting work
9 schedules, an agency can reduce response times and Labor Costs (by minimizing overtime pay).
10
11 Outage by Month of Year
12
Figures 5 (F) and (G) show the number and duration of outages by month. Summer months have
13 high outages, when the load on the utility system overall is the highest. Moreover, though thunderstorms
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
16
1 occur throughout the year, 75 percent occur from May-August. These storms frequently bring damaging
2 winds, tornadoes, hail, and lightning. Most are associated with cold fronts that sweep in from the west or
3 with daytime development over mountainous terrain. The most severe thunderstorms, which account for
4 about 20 percent of annual storms in the DC region, occur mostly during the spring and early summer
5 when the atmosphere is more unstable [16].
6
7 Geographical Distribution of Outages
8
Figure 6 (A) below shows the total power-related signal outage time per intersection between
9 2008 and 2016. The figure also shows the locations of signals with BBS and signals that are capable of
10 taking a generator. It appears that the back-up systems are covering areas of high outages (the darker
11 colors). There are a few obvious "gaps" that are highlighted. These locations would benefit from future
12 BBS installations or retrofitting the cabinet to take a generator.
13
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
17 Figure 6 (A) Distribution of Total Signal Power Outage Time 2008-2016
Signals needing additional power back-up
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
18 Figure 6 (B) Power Outage Trends over Time 20082016
1 FIGURE 6 DISTRIBUTION OF POWER OUTAGES
2
3
Figure 6 (B) shows signal power outage hotspots over time from 2008-2016; the boxed areas
4 show the most problematic sections of the city: Takoma (purple), Columbia Heights (orange), and Capitol
5 Hill/Southwest Waterfront (green). As many of the intersections in these neighborhoods are along
6 evacuation routes, a majority already have one form of backup, mostly generator ports but some BBS too.
7 These hot spots would benefit from greater BBS installations, and perhaps greater battery capacity for the
8 existing BBS installations.
9
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
19
1 COST AND BENEFITS OF GENERATOR AND BBS DEPLOYMENT
2
This section makes the business case for including generators and BBS as part of the signal
3 system management and operation toolkit.
4
5 Cost
6
There are both capital and operating costs associated with generators and BBS, which are
7 provided below in Table 2. Part of the operating costs are fixed (such as storage and routine
8 maintenance), while the remainder is based on the number of deployments (in case of generators).
9
10 Benefits
11
There are several tangible and intangible benefits. Tangible benefits include savings in personnel
12 cost and user costs. Intangible benefits include increased safety for city personnel (such as police and
13 traffic control officers) and motorists, in addition to savings in emissions and fuel consumption due to
14 better traffic flow.
15
16 Savings in Personnel Cost
17
The impact of a signal outage is dependent on the location, time, and duration of the outage and is
18 a function of the traffic demand. In the District, motorists are required to treat a dark signal as a 4-way
19 stop controlled intersection. DDOT's standard operating procedure is to first and foremost deploy
20 portable stop signs when a signal loses power. DDOT's signal technicians, roadway operations patrol,
21 signal contractors, and public space inspectors carry portable stop signs for rapid deployment as
22 necessary. For critical intersections during busy time periods, either a police officer or TCO is dispatched
23 to direct traffic at dark signals. This deployment is dependent on weather conditions, location, and time
24 of day. As an example, a TCO will rarely be deployed an intersection outside of downtown or after dark
25 out of safety concerns for the officer due to poor visibility in a black-out situation. police officers have
26 slightly higher visibility because of the flashing lights on their squad cars.
27
Powering a signal using a backup helps DC avoid the cost associated with deployment of police
28 or traffic control officers, although generator backups incur a personnel cost as well. Applying standard
29 hourly rates for MPD and TCOs, if resources had to be deployed at intersections powered by generators
30 or BBS the additional cost would have been $850,000 for the six year period between 2008 and 2013.
31 There is also an opportunity cost of not having to deploy a police officer at an intersection, which has not
32 been quantified.
33
34 Savings in User Cost
35
The user cost benefits were estimated by calculating the difference in vehicle delay based on
36 regular signal operation (when powered by generator or BBS) and four way stop (when a signal is dark).
37 SYNCHRO was used to model the delay under the two scenarios. Multiple intersections were modeled
38 based on the functional classes of intersecting roadways. All the intersections powered by backups were
39 categorized based on the functional classes of the intersecting roadways. The delay savings were applied
40 to an average traffic volume for the approaches based on their functional classes [17] to derive total
41 savings in delay. The delay savings were monetized based on the methodology suggested by Federal
42 Highways Administration [18] and median income data for the DC metropolitan area.
43 The user cost savings as a result of keeping signals operating using a generator was estimated to be of the
44 order of $28 million over a six year period. Table 2 below summarizes the costs and benefits of generator
45 and BBS deployment.
46
47
48
49
50
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
20
1 TABLE 2 SUMMARY OF COSTS AND BENEFITS
Generators
BBS
Costs
Capital Cost
$8,000/unit
$8,000/unit
Fixed Operating $400/unit/year : Storage
$20/unit /year : Maintaining
Costs
$600/unit/year : Preventative
battery cache and monitoring
maintenance
battery levels
Deployment
$275/unit: Deployment
$25/unit: Battery replacement
Costs
92: average # of generator
(labor)
deployments/year (2008-2015)
$175/unit: Battery replacement
$26,300/year: Deployment cost (battery capital cost)
Benefits
Savings in
$141,700/year
personnel cost ­ -$26,000/year
Deployment cost $115,700/year
Savings in user cost Total Savings
$4.67 million/year $4.79 million/year
Comments Assumes TCOs and police would each cover half of the outages Sketch planning level estimate
Intangible benefits
· Personnel safety · Public safety · Reduced emissions · Lower fuel consumption · Opportunity cost of assigning officers to tasks other than traffic control
2
4 SUMMARY OF FINDINGS & LESSONS LEARNED
5 DC's investments in more proactively maintaining and managing its signal assets seem to be paying
6 dividends. Since 2008, generators have powered traffic signals for more than 16,000 hours or 31% of the
7 time that signals were dark due to power outages. In calendar year 2013 alone, BBS powered signals for
8 over 500 hours. These measures reduce motorists' confusion, enhance safety and operations, and have
9 positive environmental impacts (reduced emission and fuel consumption). They also help minimize lost
10 productivity for travelers (user cost savings) and eliminate the need for deploying TCOs and police
11 officers at dark intersections to direct traffic. DDOT will use the analysis in this paper to inform future
12 backup installations.
13
This paper focused on power outage characteristics to prioritize which intersections should
14 receive backup systems. However, the authors suggest that further analysis be conducted to evaluate the
15 effects of signal power issues on the following intersection criteria: crash volumes and characteristics, trip
16 volume (vehicle, bicycle & pedestrian), preemption, geometry, cabinet load, proximity to emergency
17 services, proximity to important institutions, and reversible lanes. Other jurisdictions may wish to
18 consider intersections with railroad crossings and approach speed limits for their backup prioritization
19 analysis. DC does not contain any railroad crossing intersections, and has a relatively low maximum
20 speed limit compared with other U.S. cities. Critical thresholds should be determined over the course of
21 future analysis in order to assign points or weights to these criteria.
22
In addition to the backups themselves, there are several lessons learned around backup
23 implementation in DC that other jurisdictions could leverage. These include:
24 · Adopt a data-based procurement strategy ­ Agencies should adopt a procurement strategy that is
25
based on the following factors:
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
21
1
o Outage distribution ­ BBS can be automatically deployed, but portable generators offer the
2
flexibility of being applied to multiple intersections. So, as a general rule of thumb, if a
3
jurisdiction experiences outages at a few intersections, BBS might be a logical choice.
4
However, if outages are uniformly distributed over a large percentage of the system,
5
generators might be more cost effective.
6
o Outage duration - BBS has the flexibility of detecting power outages and automatically
7
switching to the alternative power source. Generators need to be warehoused, maintained,
8
driven to the site, and deployed following a power outage, resulting in a delayed response
9
time. So, as a general rule of thumb, BBS make sense for shorter outages; generators for
10
larger.
11
o Cost ­ Capital and operating costs are important factors. Operating costs include costs for a
12
preventive maintenance program for both BBS and generators. Generators have additional
13
operating costs such as storage/warehousing and deployment. The other cost parameter to
14
consider is the cost to provide system-wide coverage. As an example, given DC's price
15
points, the capital investment for complete BBS coverage on a system with 1000 signals will
16
be approximately $13 million. However, the jurisdiction might decide based on its risk
17
management strategy that procuring 100 portable generators for $0.8 million provides
18
sufficient coverage.
19 · Expand coverage programmatically ­ After settling on a technology and establishing that it adds
20
value, expand the coverage programmatically. One way to do this is to ensure that the specifications
21
for the backup become part of the standard signal specifications. This way, the cost of the back-up
22
devices can be included as part of the capital cost of new signal construction or signal
23
upgrades/modifications.
24 · Budget for maintenance ­Ensure that there is sufficient budget for maintenance and storage so that
25
these assets are available and functional when needed.
26 · Reduce response time by fine tuning deployment strategy ­ DDOT reduced deployment time for
27
generators by ensuring that the vehicles used by signal technicians are loaded with generators always.
28
DDOT also stores generators in different facilities for rapid deployment.
29 · Form alliances with non-DOT stakeholders ­ Rapid response to signal outages require
30
coordination and cooperation between non-DOT stakeholders such as police, fire/emergency medical
31
services, urban forestry, and utility companies. Ensure that the standard operating procedures include
32
the non-traditional stakeholders and that the respective roles and expectation have been vetted.
33 · Upgrade communications networks together with backup installations ­ The time constraints
34
associated with grants used for procurement of BBS precluded DDOT from deploying a centralized
35
communications system during initial installation. Staff had to retroactively install twisted copper
36
connections. Real-time and centrally-network communications are key to utilizing staff time
37
efficiently. In addition, fiber networks greatly enhance the amount of data that can be extracted from
38
backup systems, including battery levels and temperatures, which can help implement more targeted
39
and efficient maintenance programs.
40
41 ACKNOWLEDGEMENTS
42
The authors would like to acknowledge the contributions of Stephanie Dock, Robert Waller,
43 Tammie Jones, Eugene Patrick, Negin Askarzadeh, Amy Liang, Kathleen Crabb, and Craig Moran of
44 DDOT, as well as Michael Keatley and Matthew Hill of M.C. Dean. The authors would also like to thank
45 all the survey respondents for taking the time to share their experiences, including Keith Riniker (Sabra
46 Wang), Sivasailam Daivamani (MWCOG), Michael Miller (VDOT-Eastern Region), Greg Sawyer
47 (Virginia Beach, VA), Amit Sidhaye (Arlington), Ed Rodenhizer (MDSHA), Robert Lewis (Suffolk),
48 Jackie Kassel (Newport News), Michael Kinney (Montgomery County), Matt Milkerson (Alexandria),
49 Mark Hagan (VDOT-Northern Region), and Thomas Moore (Nevada); in addition to the National
50 Association of City Transportation Officials (NACTO) and the Research Advisory Committee of the
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
22
1 American Association of State Highway and Transportation Officials (AASHTO) for helping distribute 2 the online survey. 3 4
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
23
1 REFERENCES
2 1. Manual on Uniform Traffic Control Devices. Federal Highway Administration, U.S. Department of
3
Transportation. 2009 Edition including Revisions dated May 2012.
4 2. Wallace, W. A., J. Wojtowicz, D. Torrey, N. Renna, and J. Tan. Guidelines for Traffic Signal Energy
5
Back-Up Systems. Technical Report 09-05. Center for Infrastructure and Transportation Studies,
6
Rensselaer Polytechnic Institute, Troy, N.Y., 2009.
7
https://www.dot.ny.gov/divisions/engineering/technical-services/trans-r-and-d-repository/C-06-
8
08_final%20report_August%202009.pdf
9 3. Zhao, Mo, Anuj Sharma, Edward Smaglik, and Tim Overman. Traffic Signal Battery Backup
10
Systems: Use of Event-Based Traffic Controller Logs in Performance-Based Investment
11
Programming. 2015. http://trrjournalonline.trb.org/doi/pdf/10.3141/2488-06
12 4. Addressing Long Term Power Outages at Traffic Signals. Montgomery County Department of
13
Transportation. December 2012.
14 5. Alternate Power Sources for Traffic Signals and Street Lights. Memorandum, Miami-Dade County
15
Board of Commissioners. May 16th, 2006.
16
http://www.miamidade.gov/govaction/legistarfiles/Matters/Y2006/060690.pdf
17 6. Johnson, Candus. Howard Co. lights its intersections - even in outages. Baltimore Sun. July 29, 2012.
18
http://articles.baltimoresun.com/20120729/business/bsmdtrafficsignalbackup20120712_1_intersectio
19
nsbatterybackupsignals. Accessed on July 2, 2016.
20 7. Mo Zhao, Anuj Sharma. Optimal Deployment of Hybrid Alternative power system at Signalized
21
Intersections. 2014. http://trrjournalonline.trb.org/doi/abs/10.3141/2438-08
22 8. Chitturi, Madhav V., and Rahim F. Benekohal. Effect of High and low temperatures on UPS
23
systems for Intersection Traffic Signals. Department of Civil and environmental engineering,
24
University of Illinois at Urbana-Champaign. September 20, 2005
25 9. US Department of Homeland Security. Fiscal Year 2005 Homeland Security Grant Program
26
http://www.fema.gov/pdf/government/grant/hsgp/fy05_hsgp_guidance.pdf. Accessed March 20,
27
2015.
28 10. http://ddot.dc.gov/sites/default/files/dc/sites/ddot/publication/attachments/emergency_preparedness_2
29
011.pdf
30 11. Federal Highway Administration. Office of Highway Policy Information. American Recovery and
31
Reinvestment Act (ARRA). http://www.fhwa.dot.gov/policyinformation/arra.cfm. Accessed March
32
22, 2015.
33 12. Press Release. "Mayor Gray Outlines Hurricane Sandy Recovery for the District". October 30, 2012.
34 13. US Department of Energy. A Review of Power Outages and Restoration Following the June 2012
35
Derecho. August 2012.
36 14. National Oceanic and Atmospheric Administration (NOAA)
37 15. Dvorak, P & L.H. Sun. "Power Failures Cause Havoc Downtown." Washington Post. June 14,
38
2008. http://www.washingtonpost.com/wp-dyn/content/article/2008/06/13/AR2008061301257.html.
39
Accesed June 27, 2014.
40 16. Mogil, M. and Kristen Seaman. The Climate and Weather of Delaware, Maryland, and Washington,
41
D.C. Weatherwise. http://www.weatherwise.org/Archives/Back%20Issues/2009/July-
42
August%202009/full-mogil.html. Accessed March 21, 2015
43 17. Federal Highway Administration. Highway Functional Classification Concepts, Criteria and
44
Procedures. 2013.
Dey, Pйrez, Patterson, Jain, Alexander, and Senyurt
24
1
http://www.fhwa.dot.gov/planning/processes/statewide/related/highway_functional_classifications.
2
Accessed March 22, 2015.
3 18. Federal Highway Administration. Work Zone Road User Costs ­ Concepts & Applications. 2011.
4
http://ops.fhwa.dot.gov/wz/resources/publications/fhwahop12005/index.htm#toc. Accessed March
5
23, 2015.

SS Dey, D Patterson, BO Pérez, R Jain, H Alexander

File: strategic-approach-for-enhancing-reliability-of-traffic-signals.pdf
Title: TRB template
Author: SS Dey, D Patterson, BO Pérez, R Jain, H Alexander
Author: Stephanie Dock
Published: Mon Aug 1 19:56:04 2016
Pages: 24
File size: 2.07 Mb


, pages, 0 Mb
Copyright © 2018 doc.uments.com