Illinois Four-quad Trb Paper 1109 Revision-hellman

1A. D. Hellman, T.G. Ngamdung 2 1 2 3 4 5 6 7 8

ILLINOIS HIGH-SPEED RAIL FOUR-QUADRANT GATE RELIABILITY ASSESSMENT: PRELIMINARY RESULTS

9 10 11Submission Date: November 14, 2009 12 13Word Count: 4,429 14Figures: 7 15Tables: 2 16 17 Adrian D. Hellman (Corresponding author) Volpe National Transportation Systems Center 55 Broadway RVT-62 Cambridge, MA 02142 Phone: (617) 494-2171 Fax: (617) 494-2318 E-mail: [email protected] Tashi G. Ngamdung Volpe National Transportation Systems Center 55 Broadway RVT-62 Cambridge, MA 02142 Phone: (617) 494-2937 Fax: (617) 494-2318 E-mail: [email protected]

1

1A. D. Hellman, T.G. Ngamdung 2

2

1ABSTRACT 2 3 The U.S. Department of Transportation’s (USDOT) Research & Innovative Technology 4Administration’s (RITA) John A. Volpe National Transportation Systems Center (Volpe Center), 5under the direction of the USDOT Federal Railroad Administration’s (FRA) Office of Research 6and Development (R&D), conducted a reliability analysis of the four-quadrant gate/vehicle 7detection equipment based on maintenance records obtained from the Union Pacific Railroad 8(UPRR), the owner and operator of the grade crossings. The results of this analysis were used to 9assess the impact of the equipment reliability on the proposed HSR timetable. 10 11 The Volpe Center study showed that the total average delay to the five scheduled daily 12high-speed passenger roundtrips was an estimated 10.5 minutes, or approximately 1 minute per 13train. Overall, extensive analysis of the trouble ticket data showed that the four-quadrant gate 14and vehicle detection equipment had a minimal direct impact on the frequency and duration of 15grade crossing malfunctions. Moreover, an overwhelming majority of crossing malfunctions 16equally impacted operations of both the entrance and exit gate equipment.

1A. D. Hellman, T.G. Ngamdung 2

3

1INTRODUCTION 2 3 The Chicago high-speed rail (HSR) corridor established a “hub-and-spoke” system 4centered on Chicago, Illinois, with termination points in St. Louis, Detroit, and Milwaukee (1). 5This effort, part of the broader Midwest Regional Rail Initiative (MWRRI), will eventually inter6connect nine states over a 3,000-mile (4,631 km) system shown in Figure 1. The overall goal of 7the MWRRI is to achieve reliable and frequent HSR service with trains operating at speeds 8between 90-110 mph (145 – 177 km/h). The features of this service will include new train sets, 9track infrastructure improvements, four-quadrant gate warning device technology at high-speed 10highway-rail grade crossings, and railroad signals accommodating the increased speed regimens 11(2).

FIGURE 1 Midwest High-Speed Rail Network (2). 13 These HSR systems are being implemented on pre-existing rail corridors, with highway14rail grade crossings that usually cannot be closed or separated. Typically, these crossings are not 15equipped with the risk mitigation technologies recommended by the United States Department of 16Transportation (USDOT) Federal Railroad Administration (FRA) for rail operations in the 8017110 mph (128 – 177 km/h) speed regime (3). These recommendations, although not required, 18include the installation of sophisticated traffic control/warning devices such as four-quadrant 19gates equipped with constant warning time and vehicle detection equipment. 20

1A. D. Hellman, T.G. Ngamdung 2

4

1 Although installed on a limited basis, grade crossings equipped with four-quadrant gates 2and inductive loop vehicle detection have proven to be an excellent solution in situations where 3grade separation or closure are precluded. The Chicago-St. Louis HSR corridor, with 69 four4quadrant gate crossings in a span of 60 miles (97 km), is the first large-scale deployment of this 5treatment type. 6 7 Although four-quadrant gate technology has been extremely successful as a grade 8crossing safety treatment, HSR experience in the United States has so far been limited to the 9Northeast Corridor (NEC) in Connecticut and the Southeast High Speed Rail (SEHSR) corridor 10in North Carolina. The benefits from the use of this treatment at crossings on these corridors, 11which demonstrate the operational range of four-quadrant gate technology, have been well 12documented (4,5). In addition, the Los Angeles County Metropolitan Transit Authority 13(LACMTA) has aggressively deployed four-quadrant gate technology through its light rail transit 14(LRT) system. 15 16ILLINOIS HSR CORRIDOR 17 18 Figure 2 shows the HSR corridor between Chicago and St. Louis. The corridor is owned 19and operated by the Union Pacific Railroad (UPRR), and contains a mix of freight, intercity 20conventional passenger rail, and commuter rail service. The majority of the line is single track 21and at the time the HSR program was initiated, Amtrak operated three daily roundtrip passenger 22trains. In 2006, the frequency was increased to five daily roundtrip trains. 23

24 The initial goal for the HSR service was eight round trips per day between Chicago and 25St. Louis, with one-way end-to-end travel time of approximately 3 ½ - 4 hours. However, lower 26than expected funding precluded any further infrastructure improvements between Chicago and 27Dwight. Furthermore, at the time of this study, fewer high-speed train sets than originally 28anticipated were being procured. As a result, the level of service is now projected at five round 29trips per day, with a one-way trip time of 4 - 4 ½ hours. 30 31 The yellow highlighted portion in Figure 2, between Springfield and Mazonia, has 32undergone extensive track rehabilitation, including construction of 12 miles (19 km) of double 33track and 22 miles (35 km) of freight sidings, and now satisfies the FRA Class 6 track 34regulations for 110 mph (177 km/h) train service. 35 36ILLINOIS SYSTEM OPERATIONAL OVERVIEW 37 38 The four-quadrant gate crossing system is comprised of the exit gate management system 39(EGMS), the vehicle detection subsystem, and the gates. The EGMS is a microprocessor-based 40controller that works in tandem with the vehicle detection system to resolve motor vehicle 41presence within the grade crossing and supply the appropriate gate response. During a train 42event at the crossing, the EGMS prevents both exit gates from lowering until motor vehicle 43traffic is no longer detected in the crossing (6). If the EGMS detects a compromise in the health 44of the vehicle detection equipment, it instructs the exit gates to ascend. 45

1A. D. Hellman, T.G. Ngamdung 2 1 1 2 The 3detection 4consists of 5loops 6within the 7connected to 8vehicle 9unit on the 10single track 11four loops 12in each 13exit highway 14inside the 15crossing 16double track 17crossing 18an extra pair 19installed 20tracks.

5

vehicle technology inductive embedded roadway, a solid state detection wayside. At locations, are installed approach and quadrant grade gates. For grade applications, of loops is between the These

21 configurations FIGURE 2 Illinois HSR corridor (7). 22are illustrated in Figure 3. 23Each loop contains a second “check” loop, which as its name implies, is used to check the 24operation of the primary loop. This self-check operation is performed continuously, as specified 25by American Railway Engineering and Maintenance-of-Way Association (AREMA) guidelines. 26The approach distances are for traditional track circuit operation at a maximum speed of 79 mph 27(145 km/h). High-speed operation of the grade crossing equipment is addressed later in this 28paper. 29 30The EGMS operates in two modes with respect to exit gate operation; “dynamic” and “timed”. 31In dynamic mode, which is the primary operational state, exit gate function is dependent on the 32presence and detection of vehicles within the grade crossing. If no vehicles are detected, the 33entrance and exit gates descend simultaneously. Timed mode is the EGMS backup operational 34state. If there is a failure in the EGMS hardware or the vehicle detection subsystem, exit gate 35descent is delayed until the entrance gates have reached the horizontal position. This operational 36mode provides an exit means for vehicles within the crossing when the warning system is 37activated. 38 The system is designed for fail-safe operation, meaning that the equipment will fail in the 39most restrictive operational state, so as to minimize the increase in risk. Thus, in either dynamic

1A. D. Hellman, T.G. Ngamdung 2

6

1

Southbound Circuit: 475

FIGURE 3 Four-quadrant gate (a) single track and (b) double track configurations. entional track circuit operation ,

N

1A. D. Hellman, T.G. Ngamdung 2

7

1or timed operational mode, a loss of system power will result in the entrance gates descending 2and the exit gates ascending. Like the previous description of timed mode, this condition 3provides a path for vehicles within the crossing to exit if power fails. Because the entrance gates 4remain lowered, the railroad civil speed limit at the crossing is reduced to 15 mph (24 km/h). 5 The basis for the high-speed capability of the four-quadrant gate crossings is the 6advanced activation function, in which an approaching properly equipped train interrogates the 7health status of the grade crossing using a radio-based communication link. This function is an 8overlay to the underlying conventional track circuit crossing activation circuitry. However, 9advanced activation provides vital support to high-speed operations at the four-quadrant gate 10crossings in two key respects; 1) a constant warning time (CWT) for train time to arrival and 2) a 11safe distance for responding to speed restrictions if the crossing health status is compromised. 12 Each properly equipped locomotive contains an onboard computer with a database of 13equipped crossings, including location, track circuit configuration, conventional approach 14speeds, and warning times. As the locomotive approaches a grade crossing, the locomotive 15computer initiates an advanced activation session with the grade crossing control system. Once 16the advanced activation session is established, the locomotive computer transmits the estimated 17train arrival time to the crossing based on its predicted speed, current locomotive control settings, 18and train and track characteristics. The grade crossing equipment processes the information and 19transmits a response to the locomotive that includes the operational status of the advanced 20activation function and the total time the crossing has been activated (8). If the crossing 21equipment is either inoperative or the equipment status cannot be established, the train control 22system will generate a speed restriction equal to the track circuit configuration of the crossing, 23typically 79 mph (127 km/h) (Weber, G., unpublished data). 24 25Vehicle Detection System Operational Sequence 26 27 The four-quadrant gate crossings are designed for a 30 second minimum warning time for 28passenger and freight trains operating at 79 mph (127 km/h) and 60 mph (97 km/h), respectively. 29When a grade crossing is activated, the gate warning lights flash for five seconds. In dynamic 30mode operation, if no motor vehicles are detected and the crossing equipment is operating 31normally, both the entrance and exit gates will descend in tandem. All four gates will arrive at 32the horizontal position within 10 seconds of beginning descent and remain there for a minimum 33of 15 seconds prior to train arrival. This sequence is depicted in Figure 4. During this time, if a 34vehicle is detected within the crossing, the exit gates will cease lowering and begin to ascend. 35Once the detection system verifies the crossing is unoccupied, the exit gates will resume 36descending. When all of the gates reach the horizontal position and the grade crossing island 37circuit is activated by a train, the vehicle detection system is inhibited. This prevents the train 38from being incorrectly detected as a highway vehicle, which would result in the inadvertent 39raising of the exit gates by the EGMS. Once the rear end of the train clears the island circuit, the 40gates begin to ascend and return to the vertical position within 12 seconds. 41 42 Under certain conditions, speed restrictions are issued to the locomotive as it approaches 43a four-quadrant gate grade crossing. The first case involves either a non-equipped locomotive or 44faulty advanced activation functionality at a crossing. In response, the train control system will 45revert to the underlying track circuit signaling system and the corresponding maximum track 46speeds. On the Illinois HSR corridor, these speeds are typically 79 mph (127 km/h) and 60 mph

1A. D. Hellman, T.G. Ngamdung 2

8

Start Adv. Activation Session

TOTAL APPROACH TIME

CROSSING

EQUIPMENT RESPONSE

Advanced Preemption Time = 0

Equipment Response Time

FOUR QUADRANT GATE

TOTAL WARNING TIME

ADVANCED PREEMPTION

30 SEC MINIMUM ALL LIGHTS FLASH

Entrance Gate Delay 3 Sec. Minimum Before Gates Start to Descend (IL = 5 Sec.)

Entrance Gate Descent

Exit Gate Delay EG CT

Entrance Gate Horizontal 15 Sec. Min. Before Train Enter Crossing Exit Gate Descent

Clearance Time (IL = 0)

Buffer Time (IL = 0)

Exit Gate Horizontal

FIGURE 4 Four-quadrant gate crossing system timing. (Illinois Commerce Commission, unpublished data).

2(97 km/h) for passenger and freight operations, respectively. The second scenario is a 3consequence of the crossing equipment being activated more than two minutes. If the activation

1A. D. Hellman, T.G. Ngamdung 2

9

1time falls between two and five minutes, the existing track circuit speed limit is enforced. In 2situations involving the equipment being activated greater than five minutes, a locomotive speed 3restriction of 15 mph (24 km/h) is generated. 4 5 When a crossing alarm is generate, a trouble ticket is automatically issued and a 6maintainer is dispatched to the crossing. Once the issue is resolved, the maintainer updates and 7closes out the trouble ticket. These records are stored electronically by the UPRR at its central 8office in Omaha, Nebraska. Some of the most common alarm triggers include expiration of gate 9activation timeout delays, gate release and response times in excess of defined limit, and detector 10failures (6). 11 12RESEARCH METHODOLOGY 13 14 The evaluation process consisted of 1) identifying and characterizing the malfunction 15types, 2) calculating the probability of occurrence and mean time to repair (MTTR) for each 16malfunction type, and 3) estimating the resulting cumulative delay on the proposed HSR 17timetable. 18 By means of an FRA request, two data sets were provided by the UPRR to the Volpe 19Center for analysis. These were evaluated for trends in malfunction occurrences and 20maintenance downtimes that may impact the future HSR timetable. The first data set was 21derived exclusively from exit gate related trouble tickets. The second set was a synthesis of 22trouble tickets resulting from both entrance and exit gate maintenance calls. For this data set, the 23goal was to determine the effect on the entrance and exit gates from grade crossing equipment 24malfunctions. Both datasets were used to characterize the impact from grade crossing equipment 25malfunction on the future HSR timetable from the types and frequencies of grade crossing 26equipment malfunctions. 27 28 Evaluation Assumptions 29 30• A constant train velocity of 110 mph (177 km/h) was assumed for the entire 31 corridor. 32• The anticipated number of daily roundtrip HSR trains ranges from 6-16, 33 depending on the available level of funding support for the service. 34• The high speed alternative, the most ambitious approach, has a one-way trip 35 travel time of 3.5 hours and a 110 mph (177 km/h) maximum train velocity on most of the 36 corridor with 125 mph (202 km/h) on an 18 mile (29 km) stretch between Lincoln and 37 Springfield, Illinois. The preferred alternative has a one-way trip travel time of 4 to 4.5 38 hours, subject to the extent of the infrastructure upgrades. The maximum train velocity under 39 this option is 79 mph (127 km/h) between Chicago and Dwight, Illinois and 110 mph (177 40 km/h) between Dwight and St. Louis (9). 41• The values employed in this analysis, 5 roundtrip trains per day and a one 42 way travel time of 3 hours and 50 minutes, fall between these two estimates. The high-speed 43 timetable shown in Table 1 reflects these assumptions. 44 45 TABLE 1 Typical Representation of the Chicago-St. Louis 46 High-Speed Rail Timetable 47

1A. D. Hellman, T.G. Ngamdung 2 Mile 0 12 37 74 92 124 156 185 224 257 284

10

Southbound: Read Down

Northbound: Read Up

6:45a

8:15a

3:20a

5:15p

7:15p

10:35a

12:05p

7:10p

9:05p

11:05p

Chicago Summit Joliet Dwight Pontiac Normal Lincoln Springfield Carlinville Alton St. Louis

8:25a

10:25a

12:20a

7:00p

9:00p

4:35a

6:35a

8:30a

3:10p

5:10p

2ANALYSIS OF ENTRANCE AND EXIT GATE DATA 3

W in d Is s u es 2 4%

Ex it Ga te L o op W iringO the r Contr oller Pro c e s s orLo o p 2% 1% 9% 2% De tec to r

G ate Me c han ic al Is s ue s 24%

23 %

Ele c tr on ic G a te Mo nito r 15 %

FIGURE 5 Exit gate issues from Data Set I. 4Data Set I consisted of 93 trouble tickets collected over 726 days between May 2003 and May 52005. This data set was filtered for trouble tickets that were strictly identified with exit gate 6equipment issues. Three malfunction types totaling 27 percent of the trouble tickets – exit gate 7controller (EGC), loop processor, and loop detector failure – were identified as specific to exit 8gate malfunctions. As shown in the Figure 5 pie chart, loop detector equipment accounted for 23 9percent, mostly arising from oversensitive detectors. This condition was typically resolved by 10decreasing detector sensitivity, but maintaining it above the motor vehicle detection threshold. 11

1A. D. Hellman, T.G. Ngamdung 2

11

1 Data Set II entrance and exit gate malfunction data were used to analyze the impact of the 2four-quadrant gate/vehicle detection system on the HSR corridor timetable. The data collection 3period spanned 677 days between February 2004 and December 2005. In total, 889 unique 4trouble tickets were tabulated, equating to an average of 1.31 malfunctions per day. Altogether, 538 different malfunction types were identified as being specific to the grade crossing equipment. 6Analysis of the data showed that ten malfunction types, as depicted by the Pareto distribution in 7Figure 6, contributed to approximately 80 percent of the total number of trouble tickets. The 8other 28 types accounted for the remaining 20 percent.

100

90

9Theoretical Analysis of Average Time to Fix 10 11 Each malfunction type was assigned a weighting proportional to its probability of 12occurrence. These values, along with the MTTR for each malfunction type, were used to 13calculate the impact of the four-quadrant gate/vehicle detection system on the proposed HSR 14timetable. MTTR was calculated as the difference between the time a trouble ticket was opened 15and the time it was closed by a grade crossing maintainer. For normally distributed, symmetric 16data, the arithmetic mean would typically be used. However, analysis of the MTTR data 17revealed a significant time-based component with several orders of magnitude between the 18highest and lowest values. Also, the data sets are positively skewed and bounded by zero. These 19characteristics are associated with a log-normal distribution rather than a normally distributed, 20symmetric distribution. For this type of application, the geometric mean, which is related to the 21log-normal distribution, provides a more realistic model for averaging data. 22

80

1A. D. Hellman, T.G. Ngamdung 2 1 2 3 4 5 6 7 8

Calculating Delay Let MT DT M

= Total number of malfunctions = Total days = Average number of malfunctions per day

Then M =

9 10 11 12 13 14

12

Also, let V f Pf

M T 889 = = 1.31 malfunctions per day DT 677

(1)

= Malfunction code frequency = Probability of malfunction code

Then

Pf =

Vf

MT 15 (2) 16 17the contribution of each malfunction type to M is Pf ⋅ M , (3) 18and the average weighted daily delay (AWDD) resulting from a malfunction type is 19 AWDD = Pf ⋅ M ⋅110 s ⋅ N E , 20 (4) = P ⋅ 1 . 31 ⋅ 110 s ⋅ N 21 f E 22where the number of affected trains per day is N E 23the total schedule delay from a malfunction, assuming it has occurred, is 110 s ⋅ N E . 24 (5) 25 26In calculating N E , the following assumptions were used: 27 28 • The geometric MTTR of each malfunction type equals the average time a malfunction 29 will affect the HSR timetable 30 • The proposed Chicago-St. Louis HSR trip time is 3 hr 50 min 31 • The end-to-end trip time implies an average train speed of 73.2 mph 32 • 10 trains will operate daily between Chicago and St. Louis daily (5 roundtrips) 33 34 Similarly, several assumptions were used in the total schedule delay, 110 s ⋅ N E : 35 36 • Under worst-case conditions, the maximum allowable train speed at a malfunctioning 37 grade crossing is 15 mph (24 km/h). 38 • The train deceleration and acceleration rate is 1 mph/s (0.621 km/s), 39 • Train length is 500 ft (151 m), including the locomotive 40 • A typical crossing is 150 ft (45 m) in length 41

1A. D. Hellman, T.G. Ngamdung 2

13

1 Under these assumptions, a train will require 95 seconds (s) to decelerate from 110 mph 2(177 km/h) to 15 mph (24 km/h), while covering a distance of 8,650 ft (2,621 m). At 15 mph (24 3km/h), the train will require 30s to traverse the crossing and clear the island circuit. The process 4is then repeated as the train accelerates back to 110 mph (177 km/h), for a total time of 220s (G. 5Meyer, unpublished data). This speed profile is illustrated in Figure 7. Conversely, a train that

120 100

i/hr)

FIGURE 7 Speed profile of a 110 mph train approaching a 15 mph grade crossing. 6is not required to reduce speed will traverse the entire 17,950 ft (5,440 m) in approximately 110s. 7The difference between the two times, 110s, is the delay a train will incur due to a single 8malfunctioning crossing. 9 N E for each malfunction type was resolved manually by scanning the proposed HSR 10 11timetable for the interval, equal to the MTTR, in which the maximum number of trains will be 12present on the corridor. Since not all trains on the corridor will be affected by an individual 13malfunction, an approach was developed to estimate the number of trains that will experience a 14delay. The arrival times for each train at the four-quadrant gate crossings were estimated using 15the proposed 3 hour 50 minute schedule. 16 17RESULTS AND DISCUSSION 18 19 In this analysis, 889 unique malfunction tickets were recorded over 677 days, resulting in 20an average of 1.31 malfunctions per day. The probability of each malfunction type, ranging from 21 11 ⋅10 −4 to 23 .2 ⋅10 −2 , was calculated by applying the above delay equations to the frequency 22data for each malfunction type. Additionally, the total average daily delay was 10 minutes and

80

1A. D. Hellman, T.G. Ngamdung 2

14

138 seconds, equivalent to an average of about one minute per train. Thus, on average, the impact 2from the malfunction codes on the HSR timetable was minimal. 3 4 The second and third columns of Table 2 show the malfunction types ranked by event 5frequency and probability of occurrence, unadjusted for MTTR or duration. Two types, No 6Cause Found and Other were typically entered by the grade crossing maintainer when a 7malfunction was reported or an alarm was recorded but could not be duplicated. In many cases, 8the trouble tickets were related to the exit gate equipment, including improper lowering of one or 9more exit gates, and wind issues. Although No Cause Found and Other were not conclusive 10indicators of delay to the HSR timetable, they could not be eliminated as potential predictors of 11future crossing equipment induced delays. The MTTR values, derived from trouble ticket 12open/close times, are found in column four. 13 14 The geometric averaged weighted daily delay for each malfunction type is shown in the 15fifth column of Table 2. These values were calculated from the product of the event probability, 16the number of trains affected per day assuming a 10-train schedule, and the worst-case delay 17experienced by a single train from a malfunction (110 seconds). This calculation yielded a 18probabilistic estimate of the contribution from each malfunction to the average of 1.31 19malfunctions per day. These values, including the total average weighted delay for the entire set 20of malfunction types, are found in the fifth column of Table 2. The sum of the 7 malfunction 21types responsible for approximately 80 percent of the trouble tickets contributed over 8 minutes 22to the total 10.5 minutes of average weighted daily delay. Assuming a 10-train daily schedule, 23this equates roughly to an average of one minute per train. 24 25 The last column in Table 2 is an estimate of the HSR timetable delay attributed to each 26malfunction event, assuming 100 percent probability of occurrence. These results typify the 27expected delay until a malfunction event has been resolved. Of importance is the marked 28difference from the average weighted delay values. More significantly, these results show that 29the average weighted delay, which is directly related to event probability, may not necessarily be 30the best measure of the impact from a malfunction. A new metric for characterizing the impact, 31Delay Index (DI), is presented in column six. This is a measure of the delay incurred on the 32HSR timetable resulting from a particular malfunction type and is analogous to the expression 33for risk in safety-related research. DI is expressed as the product of the event probability, Pf , 34and the average weighted daily delay (eq. 4) resulting from each malfunction type. The formula, 35 DI AWWD = Pf ⋅ AWDD , 36 (6) 37 38is akin to the expression for risk in safety-related research, where AWWD is the severity term. 39Delay is often used as an alternative metric for risk. However, in this study, the delay index was 40employed as a means to measure the delay impact for a grade crossing malfunction, so as to 41maintain the distinction with the traditional definition of risk. As shown in sixth column of 42Table 2, the average delay term weighs the DI ranking such that the Electronics Failure and 43Sand, Rust, Or Other Deposit On Rail terms have a stronger impact on timetable delay than as a 44function of event frequency or probability only. 45 46 TABLE 2 Malfunction Events Ranked By Event Frequency 47

1A. D. Hellman, T.G. Ngamdung 2

15 1

No Cause Found Electronics Failure Gate Mechanical Failure AC Power Failure Sand, Rust, Or Other Deposit On Rail Gate Hung Up In High Wind Bracket/Cantilever Other Not Dispatched Planned Work Gate Arm Re-hung

206 147

23.17 16.54

2:10:59 9:55:26

Average Weighted Daily Delay (mm:ss) 1:40 2:22

98

11.02

2:10:45

0:47

53

5.96

6:53:25

0:42

48

5.40

165:57:41

1:18

Totals Highest 20% Remaining 80% For All Types

Top 20% of Malfunction Events

Event Event MTTR Frequency Probability (hh:mm:ss) (%)

(DIAWWD) 38.62 39.14 8.63 4.17 7.02 2.40

40

4.50

3:34:00

0:32

40 31 24 23

4.50 3.49 2.70 2.59

1:44:34 0:34:43 6:06:28 2:06:31

0:19 0:10 0:19 0:11

710

80

08:20

179 889

20 100.00

02:08 10:38

1.43 0.58 0.86 0.47

1 2 3 4CONCLUSIONS 5 6 The purpose of this research was to evaluate the timetable impact, if any, from the 7reliability of the 69 four-quadrant gate/vehicle detection systems installed on the future Illinois 8HSR corridor. Since a HSR timetable was not in existence, the authors constructed a 9representative “best-guess” estimate that is a synthesis of likely one-way trip timetables and 10daily roundtrip frequencies published in the public domain. 11 12 Equipment reliability was determined using a probabilistic model developed by applying 13statistical analysis to identify trends in grade crossing equipment trouble tickets. The 38 14malfunction types were sorted by type, frequency (which was used to derive probability), and 15resolution time (which was used to calculate MTTR). The weighted probabilities of occurrence 16and MTTR for each malfunction type were used to calculate the impact of the four-quadrant 17gate/vehicle detection system on the proposed high-speed timetable. 18 19 The total average daily delay was found to be 10 minutes and 38 seconds, equivalent to 20an estimated average of one minute per train for a 10-train daily schedule. As such, the 21reliability of the four-quadrant gate/vehicle detection system will incur minimal impact on the 22HSR timetable. However, some interesting insights resulted from this research. First, the

1A. D. Hellman, T.G. Ngamdung 2

16

1majority of trouble tickets were related to the maintenance of railroad signaling system 2components, of which the grade crossing electronics are a part, and not an indication of the four3quadrant gate/vehicle detection system reliability. Moreover, an overwhelming majority of 4crossing malfunctions equally impacted operations of both the entrance and exit gate equipment. 5However, the malfunction or improper operation of a small subset of components was predicted 6to result in potentially prolonged disruptions to passenger rail operations. These low probability 7events may occur concurrently at multiple grade crossings, potentially resulting in an 8amplification of the impact on the HSR timetable. 9 10 Analysis of trouble ticket data can be used to identify recurring maintenance issues that 11may require further study. This may eventually lead to optimization of railroad inspection and 12maintenance procedures. Fortunately, longitudinal analysis of maintenance data will facilitate 13identification of such long-term trends. Based on this research, railroad and state engineers will 14be able to review and, if necessary, modify maintenance procedures to optimize operation of the 15four-quadrant gate technology. 16 17ACKNOWLEDGEMENTS 18 19The authors wish to thank Leith Al-Nazer and James Smailes (retired), U.S. DOT Federal 20Railroad Administration, Office of Railroad Development, for their insight, guidance and 21direction in developing this report. We would also like to thank Stan Milewski and Brian 22Vercruysse of the Illinois Commerce Commission, Rail Safety Section, for providing invaluable 23technical support and data analysis assistance. Special thanks to William Breeden, Director 24Signal Engineering (retired), Union Pacific Railroad, who graciously satisfied all of our data 25requests. We would also like to express our appreciation to Rick Campbell of Campbell 26Technology Corporation and Matthew Ablett of Railroad Controls Limited for insightful 27technical discussions about the Exit Gate Management System. 28 29REFERENCES 30 31(1) High-Speed Ground Transportation for America. FRA, U.S. Department of Transportation, 32 1997. 33 34(2) Midwest Regional Rail System, Executive Report. Transportation Economics and 35 Management Systems, Inc. and HNTB Corp, 2004. 36 37(3) Rail-Highway Crossing Safety Action Plan Support Proposals. FRA, U.S. Department of 38 Transportation, 1994. 39 40(4) Hellman, A.D., Carroll, A.A., and D.M. Chappell. Evaluation of the School Street Four41 Quadrant Gate/In-Cab Signaling Grade Crossing System. Publications DOT/FRA/ORD42 07/09 and DOT-VNTSC-FRA-03-04. FRA, U.S. Department of Transportation, 2007. 43 44(5) North Carolina “Sealed Corridor” Phase I U.S. DOT Assessment Report: Report to 45 Congress. FRA, U.S. Department of Transportation, 2001. 46

1A. D. Hellman, T.G. Ngamdung 2 1(6) 2 3 4(7) 5 6 7(8) 8 9(9) 10 11 12 13 14 15 16 17

17

Exit Gate Management System User’s Manual Version 2.3 Software. Railroad Controls Limited, Benbrook, Texas, 2004. Tse, T. FRA Initiatives - PTC. Presented at National Transportation Safety Board Positive Train Control Symposium. Ashburn, VA, 2005. IDOT PTC Project System Specification Version 3.0. ARINC & CANAC, 2000. Final Environmental Impact Statement Chicago - St. Louis High-Speed Rail Project. FHWA, FRA, and IDOT, U.S. Department of Transportation, and Illinois Department of Transportation, 2003.