Airport Veer-off Risk Assessment: an Italian Case Study

The objective of this paper is to assess the veer-off risk of an Italian airport that is characterized for having near 12,000 annual movements. The name of this airport is not disclosed for security purposes. The methodology used followed the principles of probabilistic risk analysis in order to characterize the events and assess the corresponding damages. The study used statistical data about accident reports and local conditions that were collected following the standards of the International Civil Aviation Organization (ICAO). The methodology used in this work complies with the guidelines for the adjustment of lateral runway strips, edited by the Italian Civil Aviation Authority (ENAC). Besides, data available in worldwide databases of airplane accidents were also gathered and included as part of the analysis. The method used to assess the veer-off risk of the airport is consistent with probability and damage quantification methods published in the literature. The main variables considered in the analysis were traffic information, wind conditions, the existence of landside buildings adjacent to the runway, and the geotechnical conditions of the subgrade underneath the strip zones. For the assessment of the veer-off risk, the authors used primary data provided by the airport management body within the period 2013-2015 and secondary data available in the literature. The risk of veer-off was calculated in more than 1,500 points around the runway. Besides, the authors proposed maximum allowable risks in different locations, and these values were compared to the actual risk levels previously computed. The results of this comparison suggested that improvements in the soil capacity and/or in the airport management activities might contribute to achieve the proposed allowable risk. The results from this assessment showed that the two critical variables determining the risk of veer-off accidents in the airport under evaluation were wind conditions and the bearing capacity of the soil underneath the strip areas. Also, it was found that the highest veer-off risk level obtained within the Cleared and Graded Area (CGA, part of the runway strip cleared of all obstacles and graded) was 2 10-7, while the lowest level was 3 10-8, which are considered typical risk ranges in airport operations. In general, the results demonstrate that the adopted methodology is a useful tool to evaluate the veer-off risk of a specific airport. Besides, the method allows comparing the actual levels of risk with proposed target levels of safety. Consequently, the quantification of the veer-off risk levels offers the airport management body the possibility of implementing appropriate measurements in those cases where minimum safety requirements are not achieved. Keyword: airport, risk assessment, veer-off, strip, probability analysis. INTRODUCTION Safety plays a central role in the design and operation of infrastructure systems. Thus, any system should comply with some pre-specified minimum requirements, which are usually defined by codes of practice or equipment specifications, in order to guarantee an acceptable level of safety. In transportation engineering, safety has to be addressed to control the potential impact of an accident on the physical elements or components of the system and on the users, staff members and other people that interact with those components (e.g. Cantisani et al., 2012). Among the different modes of transportation, aviation is considered one of the safest (European Aviation Safety Agency, 2016). In fact, existing data demonstrate that on-board fatalities in aircrafts with a take-off mass above 5.7 Mg have been declining in the last decades (Boeing, 2016). Safety in flight operations aims at achieving the maximum protection of the people involved and at preserving the integrity of the resources involved in airport activities. In this context, Safety Management Systems are the tools defined by the International Civil Aviation Organization, ICAO, to improve safety standards in global aviation (ICAO, 2014). For these systems to be effective, they should be developed on the basis of riskbased management principles. Distefano and Leonardi (2013), for example, have proposed a risk assessment methodology after considering aircraft accidents between 1980 and 2010. Besides, several other studies reported in the literature have presented risk assessment for various accident types (e.g., Attaccalite et al., 2012; Ayres et al., 2013; Feng & Chung, 2013; Loprencipe & Di Mascio, 2016). The results from these procedures are expected to provide accurate data on the actual level of safety of some airport-related operations that, if needed, could be used to make decisions in order to achieve minimum safety requirements. In the specific case of veer-off accidents, there exist some specific methods to assess the probability and risk of these events (Airport Cooperative Research Program, 2014; Norwegian Civil Aviation Authority, 2001). A veer-off is an overrun accident in which the aircraft leaves the side of runway, stopping in the lateral areas; i.e., Cleared and Graded Area (CGA), runway strips or farther (Cardi et al., 2012). Depending on the length of the drift and on the structures adjacent to the runway, these accidents can have significant consequences. In some cases, they compromise both the airport system and the nearby airport areas. Veer-offs are produced during takeoff or landing movements due to the misalignment in the VOL. 12, NO. 3, FEBRUARY 2017 ISSN 1819-6608 ARPN Journal of Engineering and Applied Sciences ©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved. www.arpnjournals.com 901 header landing or the loss of direction while running on the track. The five main factors causing these events are: weather conditions, pilot’s behaviour, mechanical condition of the aircraft, functional and structural characteristics of the airport management system, and overall condition of the runway (Di Mascio et al., 2008). The type of movement (landing or take-off), the type of runway (instrument or non-instrument runway), and the type of flights (passenger aviation, cargo, training, military or agricultural flights) affect the probability of a veer-off event (Moretti et al., in press a). However, existing statistical data show that the accident rate of veer-offs for general aviation is 10% higher than that for commercial aviation. In several cases, veer-off accidents cause serious mechanical damages to airplanes and significant health consequences in the people involved (i.e., passengers, crew members, operation staff, etc.). Existing data have demonstrated that in around 5% of veer-offs the aircraft stopped over the airport boundaries; therefore, the role of the strips located next to the runway are critical to control these lateral overruns. According to the ICAO Annex 14 (2013a), the strip aims to “reduce the risk of damage to aircraft running off a runway” by means of specific requirements of subgrade bearing capacity and longitudinal and transversal slope of the CGA and the strips. Also, the ICAO regulations state that the CGA should be “(...) constructed as to minimize hazards arising from differences in load bearing capacity to aeroplanes which the runway is intended to serve in the event of an aeroplane running off the runway” (ICAO, 2013a). Different national aeronautical authorities have implemented this requirement by defining a minimum value of a bearing capacity index of the soil underneath the CGA. For example, ENAC (2008) requires a minimum CBR (Californian Bearing Ratio) of 15% for the soil in these areas and a subsidence of the airplane gear less than 15 cm. If these values are not achieved, the agency should evaluate the necessity of conducting specific veer-off risk assessments (ENAC, 2008). The strength of the soils in the CGA and strips areas should be engineered to minimize the hazards that could arise from airplanes running off the runway but also to allow the access of emergency vehicles. Consequently, the bearing capacity of the CGA should be designed based on two main criteria: 1) the load transmitted by the critical project airplane, which is defined as that having the higher Maximum Take-off Weight (MTOW), and 2) the heavier aid that could be supplied to the fire brigade operating in the airport. Given the relevance of this factor, the improvement of the geotechnical conditions in these areas constitutes the instrument proposed by the Civil Aviation Authority to mitigate potential damage and consequences in case of veer-off events. This paper presents a case study of veer-off risk assessment for an Italian airport, which name is not disclosed due to security reasons. The study was carried out using the methodology proposed by Moretti et al. (in press a, in press b) in which available data on the probability of veer-offs and average damage levels were analysed against specific data of the examined airport. The value of applying this technique in the veer-off risk assessment of a specific airport, like the one analysed in this work, is the possibility of identifying the critical factors related to potential accidents. This, in turn, provides the information required by the airport management body to design a strategy to guarantee a minimum safety condition. The methodology adopted by the authors allows for an objective assessing of the risk level at the airport of interest and the results of the analysis are values comparable to a target level of risk to be defined during the analysis. METHODOLOGY According to the guidelines published by ENAC (2008), the risk analysis of veer-off accidents should include the frequency and effects of the events. The consequences of veer-offs should consider both mechanical damages generated on the aircraft and health effects caused on the people involved (i.e., on board or outside the aircraft). Because these are rare events, the risk analysis cannot be limited to the data available for the considered airport. Instead, the study requires the combination of global data, and specific data related to the geographical and morphological features of the runway (RWY). The general approach to this risk analysis is empirical, and it is based on the analysis and interpretation of events described in existing reports. The authors considered over 3,500 veer-offs, which occurred between 1946 and 2015. The data were collected from the following international sources: the National Transport Safety Board (NTSB, 2014), the Australian Transport Safety Bureau (ATSB, 2015), and the Boeing Company (2013, 2014, 2015, 2016). The analysis of this information has already permitted to obtain basic information regarding this type of accidents. Moretti et al., for example, found that for commercial flights with a MTOW over 30 Mg (ton) the frequency of occurrence of a veer-off accident within the period 1980-2015 was 1.44 10-7 (Moretti et al., in press a). Besides, it has been already stated that this type of accidents is more frequent during landing than during take-off movements. Indeed, 75 percent of the worldwide events occurred during landing (Norwegian Civil Aviation Authority, 2001). Finally, it has been also reported that the accident rate for general aviation is 10% higher than for commercial aviation (National Transportation Safety Board, 2016). Another important aspect to consider when conducting veer-offs risk assessments is the path and final location of the aircraft. The final point of the drift path leads to determine the lateral probability distribution, which is the probability that at the end of a veer-off the aircraft travels beyond a certain distance from the centreline of the runway. The Norwegian Civil Aviation Authority (2001) has found that this probability could be defined through an exponential function, such as (Equation 1): VOL. 12, NO. 3, FEBRUARY 2017 ISSN 1819-6608 ARPN Journal of Engineering and Applied Sciences ©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved. www.arpnjournals.com 902 ) ( ) ( n ax e x p   (1) where p is the probability that the aircraft at the end of the veer-off is x meters away from the RWY centreline, while a and n are coefficients that depend on the conditions at the examined contour. The distance x is calculated with respect to centre of gravity of the aircraft. For landings on instrument runways, Moretti et al. (in press b) showed that the specific equation is (Equation 2): ) 0219 . 0 ( x landings e p   (2) Similarly, for take-offs on instrument runways, the corresponding equation is (Moretti et al., in press b) (Equation 3): ) 0143 . 0 ( x offs take e p    (3) Equations 2 and 3 will be used in this work to describe the lateral probability distribution in a veer-off accident in the analysis. This curves allow assessing the probability of veer-off according to a statistical approach which is not specific to this airport. These probability results will be combined with wind conditions and traffic information (composition and runway use) to compute the actual probability of a veer-off accident in the examined airport. Regarding the damage quantification method, the guidance provided by the Aviation Authority of the United Kingdom (Civil Aviation Organization, 2014) was adopted to assess the damage level related to this type of accidents. Thus, the consequences of a plane crash are associated with a numerical value that ranges between 1 and 5, where 1 corresponds to ‘low severity’ and 5 to ‘high severity’. The existence of land-side buildings adjacent to the runway and geotechnical conditions of the soil underneath the strip areas was considered to quantify the expected damage after a veer-off accident around the runway. The combination of the veer off probability and the corresponding damage in more than 1,500 points around the runway was then used to assess the actual risk levels of these events and to obtain iso-risk curves. INPUT INFORMATION FOR THE RISK ASSESSMENT The airport under study has one instrument runway for air operations. This runway (RWY) 36/18 is 3,000 m by 45 m asphalt paved, and it has one full length parallel Taxiway (TWY), which is 24 m wide. The TWY provides access to the south and east general aviation areas. Two fast exits are available halfway across the runway. The strip of the RWY extends laterally to a distance of 150 m on each side of the RWY centreline, and the CGA extends laterally to a distance of 75 m on each side of the RWY centreline. Also, the airport passenger terminal is located on the east side of the airfield. Figure-1 illustrates the general plan of the airport. The specific characteristics of the airport, in terms of the wind and traffic conditions, the buildings located nearby the runway and the characteristics of the runway usage are presented next. Figure-1. Airport plan. Wind The Meteorological Aerodrome Reports (METAR) for the period 2013-2015 of the airport were obtained from the database available on the website www.wunderground.com (METAR, 2015). When available, hourly information was considered. Overall, a total of 26,088 records were collected, each one containing the following information: date, time, wind speed and wind direction. Since the date of the accidents is known, it is possible to calculate the frequency of the wind direction in the examined years, as observed in Table-1. In the Table, letters N, E, S and W correspond to north, east, south and west coordinates, respectively, while n/a means that the information was not available. Each combination of letters N, E, S and W corresponds to one of the 16 cardinal directions of the wind rose (e.g., NNE is northnortheast). VOL. 12, NO. 3, FEBRUARY 2017 ISSN 1819-6608 ARPN Journal of Engineering and Applied Sciences ©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved. www.arpnjournals.com 903 Table-1. Statistical data of wind direction (2013-2015 period). Wind direction Heading (°) Number of hourly registrations Frequency (%) variable 5,002 19