
[Federal Register Volume 80, Number 224 (Friday, November 20, 2015)]
[Notices]
[Pages 72779-72781]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 2015-29641]


-----------------------------------------------------------------------

DEPARTMENT OF TRANSPORTATION

Federal Railroad Administration


Information and Guidance on the Inspection, Testing, and 
Maintenance of Emergency Window Exits on Railroad Passenger Cars

AGENCY: Federal Railroad Administration (FRA) Department of 
Transportation (DOT).

ACTION: Notice.

-----------------------------------------------------------------------

SUMMARY: FRA has become aware of occurrences when emergency window 
exits on passenger cars did not operate

[[Page 72780]]

as intended because the emergency pull handle became detached from the 
window gasket when pulled, the gasket tore into multiple pieces, or the 
gasket was otherwise difficult to remove. While investigating these 
occurrences, FRA discovered that some railroads were not following, or 
did not clearly understand, the existing Federal regulations on the 
inspection, testing, and maintenance (ITM) of these window exits, 
particularly the requirement that a railroad must utilize a test 
sampling method that conforms with a formalized statistical test 
method. FRA does not believe any of these occurrences involved 
passengers or precluded passengers from opening a window in an 
emergency situation. However, in light of these concerns, FRA is 
issuing this document to provide information and guidance to railroads 
operating passenger train service on the existing regulatory 
requirements regarding ITM of emergency window exits.

FOR FURTHER INFORMATION CONTACT: Mr. Daniel Knote, Staff Director, 
Passenger Rail Division, Office of Railroad Safety, FRA, 1200 New 
Jersey Avenue SE., Washington, DC 20590, (631) 965-1827; or Mr. Michael 
Hunter, Trial Attorney, Office of Chief Counsel, FRA, 1200 New Jersey 
Avenue SE., Washington, DC 20590, (202) 493-0368.

SUPPLEMENTARY INFORMATION: 

I. Historical Background on Existing Requirements

    The current ITM requirements for emergency window exit operability 
are found in Title 49 Code of Federal Regulations (CFR) 238.113(e) and 
238.307(c)(4)(i)(B).\1\ These sections require each passenger railroad 
to test (at an interval not to exceed 184 days, as part of the periodic 
mechanical inspection) a representative sample \2\ of its passenger car 
emergency window exits to determine they ``operate as intended'' and 
``properly operate,'' respectively. Title 49 CFR 238.113(e) further 
requires the sampling method to ``conform with a formalized statistical 
test method.''
---------------------------------------------------------------------------

    \1\ Before FRA's November 29, 2013, Passenger Train Emergency 
Systems II final rule (78 FR 71786), the requirement to test a 
representative sample of emergency window exits was in 49 CFR 
239.107(b)(2) and required each passenger railroad ``to verify that 
they are operating properly.''
    \2\ The requirement to test a representative sample of emergency 
window exits, which was based in large part on Emergency Order No. 
20 (EO 20), was codified by FRA's May 4, 1998, Passenger Train 
Emergency Preparedness final rule (E-Prep final rule). See 63 FR 
24630, 24669-24670; EO 20, Notice No. 1, 61 FR 6876, 6881, Feb. 22, 
1996, and Notice No. 2, 61 FR 8703, Mar. 5, 1996.
---------------------------------------------------------------------------

    As FRA explained in Emergency Order 20 (EO 20), a February 16, 
1996, passenger train accident in Silver Spring, Maryland, involving a 
cab car on fire that filled with smoke, raised concerns that at least 
some of the train occupants could not exit through the windows.\3\ This 
accident demonstrated why emergency windows must be readily 
identifiable and operable when needed.
---------------------------------------------------------------------------

    \3\ The National Transportation Safety Board's (NTSB) Railroad 
Accident Report on this accident reported that it took a Safety 
Board investigator several minutes to remove the left-side, front 
emergency window exit of the last passenger coach in the train's 
consist. See NTSB/RAR-97/02 report at 17 (July 3, 1997). An NTSB 
investigator could not remove the same car's right-side, rear 
emergency window exit, which was later removed by another 
investigator after approximately 3 minutes of physical exertion. The 
report further noted that the lubricant used to install these 
windows had hardened over time.
---------------------------------------------------------------------------

    FRA has continually reminded railroads that these windows ``provide 
an additional means of egress in life-threatening situations requiring 
very rapid exit, such as an on-board fire or submergence of the car in 
a body of water.'' See Passenger Train Emergency Systems (PTES) II 
final rule (78 FR 71786, 71802). In FRA's February 1, 2008, PTES final 
rule, FRA reminded railroads of the requirement to test emergency 
window exits using commonly accepted sampling techniques \4\ to 
determine how many windows to test. See 73 FR 6370, 6384. In doing so, 
FRA reemphasized that sampling should be conducted to meet a 95-percent 
confidence level that no defective units remain after completing the 
tests for the windows in the sample. See id. Further, in the Passenger 
Train Emergency Preparedness (E-Prep) final rule, FRA stated that each 
railroad should ``properly consider the nature and characteristics of 
its operations and passenger equipment to plan for routine and 
scheduled inspection, maintenance, and repair.'' 63 FR 24669. FRA also 
made clear its expectations regarding the inspection and maintenance of 
emergency exits:

    \4\ Railroads should conduct their sampling under either 
Military Standard MIL-STD-105(E), ``Sampling for Attributes'' 
(formally cancelled by the U.S. Department of Defense, but still 
acceptable for FRA's representative sampling purposes) or acceptable 
non-Government, standard sampling procedures and tables for 
inspection by attributes, such as the American National Standards 
Institute (ANSI)/ASQC Z1.4-1993, ``Sampling Procedures for 
Inspections by Attributes.'' See 73 FR 6370, 6384.

    Visual inspections must be performed periodically to verify that 
no emergency exit has a broken release mechanism or other overt sign 
that would render it unable to function in an emergency. 
Maintenance, including lubrication or scheduled replacement of 
depreciated parts or mechanisms, must be performed in accordance 
with standard industry practice and/or manufacturer recommendations. 
All emergency exits that are found during the course of an 
inspection or maintenance cycle to be broken, disabled, or otherwise 
incapable of performing their intended safety function must be 
repaired before the railroad may return the car to passenger 
---------------------------------------------------------------------------
service.

Id.

II. FRA Review of Railroads' Emergency Window Testing Programs

    When FRA reviewed various railroads' emergency window exit testing 
programs, it discovered that some railroads were not following, or did 
not clearly understand, the Federal regulations on the ITM of emergency 
window exits. This was particularly true with respect to adopting a 
sampling method that conforms with a formalized statistical test method 
and to recording window test failures. As a result, FRA is providing 
this guidance to ensure all railroads have in place an appropriate 
window testing program and understand which window tests they must 
record as failures.
    Specifically, FRA considers a window to have failed testing if the 
window or a window component (e.g., gasket, pull handle) does not 
operate as intended, considering both the window design and whether the 
window removal was ``rapid and easy'' when opened in a manner 
simulating a passenger trying to remove the window in an emergency 
(e.g., to escape a car on fire). Examples of window test failures some 
railroads were not categorizing as such include situations where the 
emergency pull handle separated from the gasket, or where the gasket 
tore or needed to be removed in multiple pieces.\5\ In addition, FRA 
observed one railroad testing its windows by carefully pulling out the 
window gasket to try to avoid detaching the handle or damaging the 
gasket. FRA recognizes that many railroads prefer to reinstall the same 
gaskets and handles for the emergency windows after performing the 
tests. However, FRA makes clear it does not consider such a careful 
test to be properly conducted because a passenger would not act that 
way in an emergency.
---------------------------------------------------------------------------

    \5\ FRA makes clear that for any window that is intentionally 
designed with one or more counter-intuitive features (such as an 
emergency pull handle that separates from the gasket when pulled, or 
a gasket that needs to be removed in multiple pieces), the railroad 
must ensure that such features are clearly explained in the required 
operating instructions posted for the affected emergency window 
exits.
---------------------------------------------------------------------------

    FRA also discovered that some railroads believed they were not 
required to formally adopt a sampling program because they were testing 
100 percent of their emergency window exits over a 1- to 2-year period. 
FRA

[[Page 72781]]

appreciates these railroads' efforts for what they believed was going 
above and beyond what is considered a reasonable sample size. However, 
FRA makes clear that for a railroad to truly test 100 percent of its 
windows, the railroad would need to test all of the emergency windows 
in each of its cars at least once during a 184-day period. FRA also 
clarifies that simply testing 100 percent of the emergency window exits 
does not necessarily ensure that the windows will operate as intended 
when needed in an emergency situation. As discussed in this document, 
it is how a railroad characterizes the results of those tests and what 
a railroad does with the results of those tests that will help ensure 
the windows will operate as intended.
    Choosing the number of windows to test (whether it is 20 percent or 
100 percent) is only the first step. Second, if testing fewer than 100 
percent of the windows in a 184-day period, railroads must also ensure 
the sample is representative of the various window types in its fleet 
or fleets.\6\ Third, even if a railroad is testing 100 percent of its 
emergency window exits, it must have a program in place that requires 
monitoring of the tests to determine whether the test results 
demonstrate a 95-percent confidence level that all emergency window 
exits operate as intended. Although EO 20, Notice No. 1, would have 
required testing all window exits on a specific series or type of car 
if one such car had a defective window exit, the amended order, Notice 
No. 2, permitted the use of commonly accepted sampling techniques to 
determine how many additional windows to test. See 61 FR 8703, 8705. In 
general, these principles require that the greater the percentage of 
windows initially found defective, the greater the percentage of 
windows the railroad will have to test.
---------------------------------------------------------------------------

    \6\ Railroads can easily set up a simple spreadsheet (using off-
the-shelf software) to generate a random sample that includes 
windows representing all of the window types in a railroad's fleet 
or fleets.
---------------------------------------------------------------------------

    FRA expects all railroads to: (1) Conduct periodic reviews of 
records of window testing using an acceptable attribute sampling method 
to determine whether they are achieving a 95-percent confidence level 
that no defective units remain; \7\ (2) assess the probable cause of 
any window test failures; and (3) address any such failures. In setting 
up their testing programs, railroads must set the confidence level of 
the sample at 95 percent or more and set the defect (failure) rate at 
less than 5 percent.\8\ To perform their analyses, railroads must 
review the test results at the end of a sampling period (at a minimum) 
and take further action if the testing reveals that 5 percent or more 
of the windows in the sample are defective. When assessing the probable 
cause(s) of any window test failures, railroads should consider whether 
the failures are a result of design issues, useful life issues, or 
other systemic issues common to a particular window design or windows 
in service of a similar age. If the test failure appears to be due to a 
systemic issue, then the potential exists for the failure to repeatedly 
present itself. In such cases, FRA strongly urges that the railroad 
consider replacing all the emergency windows or window components of 
like design or similar service age, as applicable.
---------------------------------------------------------------------------

    \7\ Although the goal is to have no defective units remaining in 
a railroad's emergency window population, FRA recognizes that 
because the railroad is performing a statistical sample that 
achieves a 95-percent confidence level, there will always be a 
possibility that some defective units remain.
    \8\ These numbers are not intended always to add up to 100.
---------------------------------------------------------------------------

    As stated in the E-Prep final rule, a railroad must repair any 
window found to be broken, disabled, or otherwise incapable of 
performing its intended safety function before the railroad may return 
the car to passenger service. See 63 FR 24669. This remains true even 
when the number of windows that failed is below the 5-percent defect 
rate threshold. Railroads should also document the remedial action(s) 
planned or taken to address the window test failures, and create a 
timetable for window inspection and replacement for the window type or 
car series to remedy the problem in the most expedient manner.

III. Maintenance of Emergency Window Exits

    As noted above, FRA expects railroads to periodically perform 
visual inspections to verify no emergency window exit has a broken 
release mechanism or other overt indication that would render it unable 
to function in an emergency. Ideally, railroads would incorporate these 
visual inspections as part of the interior calendar day mechanical 
inspections of passenger cars, since they already need to inspect the 
window markings daily to ensure that the safety-related signage is in 
place and legible. See 49 CFR 238.305(c)(7). As demonstrated by the 
1996 accident that led to EO 20 (in which some of the window gaskets 
could not readily be pulled out due to lack of lubrication and 
maintenance), it is important that maintenance, including lubrication 
or scheduled replacement of degraded parts or mechanisms, be performed 
using standard industry practice and/or manufacturer recommendations to 
ensure that window exits will operate as intended during an emergency. 
This will also help to prevent a situation where a passenger in an 
emergency would panic or be delayed by trying to determine how to 
remove a window after the pull handle breaks off or a piece of the 
gasket tears off, for example.
    Finally, FRA discovered in its investigations that some employees 
were installing the window gaskets with a sharp tool (such as a 
screwdriver), which may have damaged the gaskets and may explain why, 
when pulled, the gaskets were not coming out in one piece as designed. 
Therefore, to ensure that railroads perform proper maintenance, the 
railroads should ensure that employees have and use proper tools when 
installing emergency windows to avoid damaging the window gaskets.
    As noted previously, FRA is issuing this document to provide basic 
information and guidance to railroads operating passenger train service 
to ensure that they understand the existing regulatory requirements 
regarding the ITM of emergency window exits. FRA believes that 
compliance with the existing emergency window exit regulatory 
requirements will help ensure the safety of the Nation's railroad 
employees, passengers, and the general public. FRA may take other 
appropriate actions it deems necessary to ensure the highest level of 
safety, including pursuing other corrective measures under its rail 
safety authority.

    Issued in Washington, DC, on November 17, 2015.
Robert C. Lauby,
Associate Administrator for Railroad Safety Chief Safety Officer.
[FR Doc. 2015-29641 Filed 11-19-15; 8:45 am]
BILLING CODE 4910-06-P


