PASSENGER SAFETY IN THE Light Railway Transit System
Vicente E. Montaño
This study aims to determine the riding public safety, through the safety incident count, in the Light Railway Transit station 1 and 2, from period 2004 to 2014. The Poisson distribution, a binomial distribution is used because of the event under study, the safety incident, is a rare occurrence, discrete event. The result reveals that in LRT 1 and 2 doors pining (P(X>=5) = 0.432502379) and (P(X>=5)=0.46020507); dragging, (P(X>=1) =0.191207865), and suicide intent, (P(X>=1) = 0.408167287), for LRT 1 are not rare events. However, the train collision incident, (P(X>=1) =0.0001973), is a rare event in LRT 1. Furthermore, the Chi-square result shows that a significant difference (c2 (1, n = 603) = 0, p< .01), exist between the door pinning incident in LRT 1 and 2. It appears that LRT 2 pinning incidents (n = 321) are disproportionately over-represented and LRT 1 pinning incidents (n = 258) are under-represented.
The Philippines faces a serious transportation problem. The country’s highly urbanized area and motorized dominated city, Metro Manila, has 15 million day-time populations. The city’s public transport system is inefficient and disorganized (Frialde, 2015). Largely, the deteriorating traffic problem troubled a growing economy. On top of these, public commuters are demanding a safe, efficient mass transport system (Romualdez, 2014).
The Light Railway Transit (LRT) is an accepted safe and efficient mass transport system. The LRT primarily operates in an exclusive right-of-way and run on several coupled cars (Thompson, 2014). The construction of the first LRT in the country is the recommendation of the study commissioned by the World Bank from 1976 to 1977. The Ministry of Transportation and Communication adopted the elevated version of the LRT due to numerous intersections along the routes. The LRT system was constructed to provide mass transportation and not as a revenue earning project of the government. The cost of construction was P2 billion. As expected, the government during the LRT first year operation loss P216 million (LRTA, 2015).
Three LRT lines started operations in 1984, 1999 and 2003. The Light Rail Transit Authority (LRTA), the government-owned and controlled corporation operated and controlled, LRT 1 and 2. The two LRTA operated lines on average carries, 579,000 passengers per day. Compared to the other similar mass transport system in the region, LRT fares are comparatively low. Annually, the government allocated a portion of its budget as a subsidy to the LRTA operation (ADB, 2012).
During peak periods, overcrowding is common and LRT lines overall load exceed 60 percent. To ease the congestion, the government has approved another LRT line through the public-private-partnership scheme (ADB, 2012).
The Japan international Agency (JICA) recommended several safety measures for the LRTS using their railway transport experience. JICA railway experts reveal that the poor maintenance, obsolete software technologies and depreciated facilities have made the LRT a dangerous mass transport system. In addition, the LRT coach designs are obsolete compared with other countries. These problems require improving and enhancing the LRT’s capacity to serve more passengers with less risk of an accident. Sparingly, financing is no longer the problem, but the plan implementation. In addition, the agency recommends accident prevention training, emphasizing in the system development which can analyze accidents similar to the current Japanese transport safety institution (JICA, 2014)
This study focuses on passenger safety in the LRT system. The study identified four (4) safety issues: train collision, suicide attempt, dragging and pinning (data.gov.ph., 2014). These safety incidents are the manifestation of safety-threatening situations which cause the clamor for the installation of safety measures (Lo, 2014). Although there are several issues on safety, LRT commuters, employees, students and workers continue to queue train stations demonstrate that they have very few choices to get to their destinations. Therefore an analysis of the accident probability provides insight on the improving LRT system (The Manila Times, 2014).
The ten (10) year data accumulation on safety incident in LRT 1 and 2 were obtained from data.gov.ph. The data used in this study is from the period 2004 to 2014 and incident was measured in frequency (data.gov.ph). The data analyzed were safety incidents which occurred in the ten-year period identified.
The study performed the Poisson distribution, a binomial distribution used where the event under study was a rare occurrence, discrete event. The Poisson distribution project a statistical property where the variance equals the mean (σ2 = µ). Observing the Poisson-distributed process probability for the event (x) with a mean = µ can be expressed as (Heldt, 1998).
illustration not visible in this excerpt
The x, represents the actual safety incident, resulting from the Poisson distributed process. The e value is approximately 2.71828, while, the µ, was the number of safety incident within the specified probability.
Poisson distribution has its weakness in accommodating heavy tails (De Jong & Heller, 2008). At the same time, its attractiveness is its common weakness in specifying the proportion of every possible count through a single parameter (VanRaden, 2008).
In determining if there were a significant difference in the door pinning safety incident between LRT 1 and 2, the study used the Chi-square test of independence or homogeneity. It tested whether safety incident frequency and the LRT line were dependent.
This study is based on the concept that a wider implication of an incident such as the risk present in station overcrowding and passenger trapped inside trains. For instance, during service disruptions, there is a risk to passenger safety and overcrowded stations. It also includes the train personnel ability to make contact during incidents and the safe access process. However, incident reports reveal that safety guidance is not being followed (Burr, Merrifield, Duffy, Griffiths, Wright & Barker, 2008).
Results and Discussions
Table 1 below shows that most safety incidents are the door pinning with 258 occurrences in LRT 1 and 321 in LRT 2. It is followed with 14 suicide intent, 8 dragging incident and two (2) train collisions all in the period from 2004 to 2014.
Table 1. LRT 1 and 2 Stations Safety Incidents Recorded 2004 to 2014
illustration not visible in this excerpt
Figure 1 demonstrates the four train safety incident cases and injury in Poisson distribution from 2004 to 2014 period: train collision, suicide intent, dragging and door pinning. At LRT 1 station, on average, 0.2 train collision per year occurred. In a ten-year period, two (2) train collisions were observed where the probability mass function left out values above 8.0, is already too small. The distribution is charted to the right; it is skewed to the right. Likewise, in LRT 1, on average, 1.4 suicide attempt per year occurred. In a ten-year period, 14 suicide attempts were observed where the probability mass function left out values above 8.0, is already too small. The distribution is chartered to the right; it is skewed to the right. In addition, at LRT 1 station, on average, the 0.8 dragging incident occurred. In a ten-year period eight (8) dragging incidents were observed where the probability mass functions left out values above 8.0 are already too small for a year. The distribution is chartered to the right; it is skewed to the right. Additionally, at LRT 1 station, on average 25.8 door pinning occurred. In a ten-year period, 258 door pinning incidents were observed where the probability mass functions left out values above 40 are already too small for a year. The distribution is chartered to the right; it is skewed to the right. Lastly, in LRT 2 station, on average, 32.1 door pinning occurred per year. In a ten-year period, 321 door pinning incidents were observed where the probability mass functions left out values above 40 are already too small for a year. The distribution is chartered to the right; it is skewed to the right.
illustration not visible in this excerpt
Figure 1. Four safety incident cases and injury Poisson distribution from LRT 1 and 2 stations from 2004 to 2014 period
Table 2 shows the probabilities that four safety incidents, train collision, suicide attempt, dragging and the door pinning recorded from 2004 to 2014 period are rare events or not. The standard is set at less than one percent probability (P(X>=1) =0.01). The cumulative Poisson, observing one (1) probable collision incident or more is a rare incident (P(X>=1) =0.0001973), according to the standard specification of 0.01 probabilities. The cumulative Poisson observing one (1) probable suicide incident or more is not a rare incident (P(X>=1) = 0.408167287), according to standard specification of .01 probabilities. The cumulative Poisson observing one (1) probable dragging incident or more is not a rare incident (P(X>=1) =0.191207865), according to standard specification of .01 probabilities. The cumulative Poisson observing 26 probable LRT 1 door pinning incident or more is not a rare incident (P(X>=5) = 0.432502379), according to standard specification of .01 probabilities. The cumulative Poisson observing 32 probable LRT 2 door pinning incident or more is not a rare incident (P(X>=5) =0.46020507), according to standard specification of .01 probabilities.
Table 2. Probabilities of four safety incident occurrence
illustration not visible in this excerpt
In the period of ten years (2004 to 2014), several incidents happened which reflects the unreliable safety measure in the LRT system. Due to unstable power two LRT trains collided in the Monumento station, although none were hurt, passengers were asked to walk along as railway back to the station, exposing them to risk (Poblete, 2015). Commuters are aghast to learn that they are riding unsafe trains, which the MRT Hong Kong Audit report confirmed. The audit reported that the railway systems in Manila are “poor.” This rating is interpreted as “major or extensive defects exist” and “customer requirements are consistently not being met.” The present “poor,” status is an indication that the present system is unsafe for regular operations (Mayuga, 2013).
Some of the light railway passenger injuries are caused by computers. In some circumstances, these injuries are “unauthorized test” performed prior to modifications (Neumann, 1994). Notably, most injuries are caused by both human error and system malfunction (Dempsey & Mathiassen, 2006).
Safety is very important in the LRT daily operation. The cost of safety and other operating cost are impossible to separate. At the same time, it is possible to estimate the cost and benefit of identified safety measure. Similar to the other mass transportation system the LRT has a different kind of safety issues and risk. For this reason, different measures are adopted such as signaling, fencing and other devices which may improve the passenger safety and protect the trains. LRT administrators recognized that some safety measures have greater benefits than cost and there are few with greater cost than benefits (Evans, 2010).
The operators are responsible for the safety system establishment and compliance, while the infrastructure managers and train operators coordinate. However, the safety regulators functions, according to MRT Hong Kong audit is only to oversee the process which to meet the operator’s safety system and compliance. The audits may also recommend compliance with the broad standard such as the signaling system. This is also described as co-regulations (Evans, 2010).
Unfortunately, in some cases when management decides to replace or upgrade their system, they are not certain that these changes do not lead to any safety violations or untoward consequences. In order to move closer to a better safety standard, the agency like the European Railway Agency established the Common Safety Methods (CSM) which coordinated safety certification across different member countries (Flammini, (Ed.) 2012).
The statistical results, c2 (1, n = 606) = 0, p < .01, indicate that the frequencies of pinning incidents by the station are not equally distributed; frequencies are statistically different from what would be expected by chance. It appears that LRT 2 pinning incidents (n = 321) are disproportionately over-represented and LRT 1 pinning incidents (n = 258) are under-represented.
Table 3. Frequency of pinning incidents by station
illustration not visible in this excerpt
It is up to LRT management to list the safety program and monitor its effective implementation. Whenever these safety incidents occurred, it requires a systematic evaluation of the actual circumstances causing the incident (Dhillon, 2012). Most importantly, communication plays a critical role to ensure that the suitable systems are in place for disseminating information to employees so that procedures are observed and errors avoided. Top among understanding human errors is the safety barrier concept. This is a safety mechanism which prevents undesired events from occurring or a safety net against incidents. Ideally, safety barriers must exist at each gateway stages. Its existence remains important in the railway safety review and development (Dadashi, N., Scott, A., Wilson, J. R., & Mills, A. (Eds.). 2013).
Through the Poisson probability technique, it is possible to measure the LRT line 1 and 2 safety performance. Safety gets high attention in a mass transport system. The present LRT system does not offer a guarantee of safety among its passengers. Most safety incidents recorded the door pinning, dragging and suicide intent to some degree are attributable to human factors, but it is difficult to associate human reliability to system safety. It is easy to calculate human error, however; the caveat is to look safety in a macro perspective view.
Philippines Transport Sector Assessment, Strategy, and Road Map. Asian Development Bank (ADB)
(2012). Retrieved on April 5, 2015 from https://www.adb.org/sites/default/files/institutional-document/33700/files/philippines-transport-assessment.pdf
Burr, T., Merrifield, S., Duffy, D., Griffiths, J., Wright, S., & Barker, G. (2008). Reducing passenger rail delays by better management of incidents. National Audit Office for the Office of Rail Regulation.
Dadashi, N., Scott, A., Wilson, J. R., & Mills, A. (Eds.). (2013). Rail Human Factors: Supporting reliability, safety and cost reduction. CRC Press.
LIGHT RAIL TRANSIT LINE 1 AND LINE 2 - SAFETY INCIDENTS 2004-2014. Data.gov.ph. Retrieved on March 15, 2015 from http://data.gov.ph/catalogue/dataset/light-rail-transit-line-1-and-line-2---safety-incidents-2004-2014/resource/5765086f-4809-439f-875b-6ab0d52d661a
De Jong, P., & Heller, G. Z. (2008). Generalized linear models for insurance data (Vol. 136). Cambridge: Cambridge University Press.
Dempsey, P. G., & Mathiassen, S. E. (2006). On the evolution of task-based analysis of manual materials handling, and its applicability in contemporary ergonomics. Applied Ergonomics, 37 (1), 33-43.
Dhillon, B. S. (2012). Safety and human error in engineering systems. CRC Press.
Doronila, A. (January 16, 2015). MRT ‘worst case scenarios.’ Retrieved on January 29, 2015 from http://opinion.inquirer.net/81483/mrt-worst-case-scenarios
Evans, A. (2010). Safety and Regulatory Reform of Railways.
Flammini, F. (Ed.). (2012). Railway Safety, Reliability, and Security: Technologies and Systems Engineering: Technologies and Systems Engineering. IGI Global.
Frialde, M. (January 2, 2015). YEARENDER: Traffic jams, tragedies mar 2014 for MMDA. The
Philippine Star. Retrieved on February 4, 2015 from http://www.philstar.com:8080/metro/2015/01/02/1408630/yearender-traffic-jams-tragedies-mar-2014-mmda
Hay, W. W. (1982). Railroad engineering (Vol. 1). John Wiley & Sons.
Heldt, J. J. (1998). Quality Sampling and Reliability: New Uses for the Poisson Distribution. CRC Press.
As Railway safety concerns in PH grow, Japan dispatches experts to boost safety of Filipino commuters. November 21, 2014. Japan international cooperation agency (JICA). Retrieved on March 12, 2015 from https://www.jica.go.jp/philippine/english/office/topics/news/141121.html
Lo, C. (January 13, 2014). Rescuing Manila's light rail network. Retrieved on January 28, 2015 from http://www.railway-technology.com/features/featurerescuing-manilas-light-rail-network-4147839/
LRTA History. DOTC. Light Railway Authority (LRTA). Retrieved on February 27, 2015 from http://www.lrta.gov.ph/index.php/2014-05-21-01-05-51/2014-05-22-02-12-18/2014-05-22-02-14-54
Convert the LRT/MRT into a real mass transport solution. The Manila Times. August 20, 2014. Retrieved on April 2, 2015 from http://www.manilatimes.net/convert-lrtmrt-real-mass-transport-solution/120423/
Mayuga, J. (September 27, 2013). Riding LRT, MRT now 'very risky', BusinessMirror. ABSCBN News. Retrieved on January 15, 2015 from http://news.abs-cbn.com/business/09/26/13/riding-lrt-mrt-now-very-risky
Neumann, P. G. (1994). Computer-related risks. Addison-Wesley Professional.
Poblete, A. (December 15, 2015). 8 Public Transport Issues that Made Commuters Cry in 2015.
Retrieved on January 10, 2015 from http://8list.ph/2015-transport-issues-metro-manila/ Probability formula. Poisson Distribution. Retrieved on February 8, 2015 from http://www.probabilityformula.org/Poisson-distribution.html#
Romualdez, B. (August 21, 2014). MRT’s plague of problems. SPYBITS (The Philippine Star) Retrieved on January 19, 2015 from http://www.philstar.com/business/2014/08/21/1359803/mrts-plague-problems
Thompson, J. (November 24, 2014). Light rail enters the West’s most sprawling metropolis. High Country News. Retrieved on January 17, 2015 from http://www.hcn.org/issues/46.20/two-cities-develop-public-transportation
VanRaden, M. J. (2008). Cumulative Logit-Poisson and Cumulative Logit-Negative Binomial
Compound Regression Models for Count Data. ProQuest.