Improved Models for Risk Assessment of Runway Safety Areas (2011)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

23 Risk is the likelihood of the worst credible consequence for a hazard. Many overruns, veer-offs, and undershoots have resulted in aircraft hull loss and multiple fatalities, and there- fore, the worst credible level of consequences may be assumed to be catastrophic, according to the severity classification defined by the FAA and presented in Appendix F. In some situations, a pilot may lose control of the aircraft, resulting in the destruction of the equipment with possible fatalities, even when the aircraft accident takes place inside the RSA or the runway; however, in the majority of accidents, the RSA will offer some protection to mitigate consequences. Consequences will depend on the type of structures and the level of energy during the aircraft collision. Possible obstacles may include buildings, ditches, highways, fences, pronounced drops in terrain, unprepared rough terrain, trees, and even navigational aids (NAVAID) structures, like approach lighting system (ALS) towers and Localizer antennas, particularly if mounted on sturdy structures. The energy of the aircraft during the collision is related to its speed when it strikes the obstacle, i.e., the greater speeds are expected to result in more severe consequences. Also, the consequences will depend on the type of obstacle. An aircraft striking a brick building at 40 mph may be destroyed whereas if the obstacle is a perimeter fence less severe consequences are expected to occur. The variables assumed to have an impact on consequences resulting from overruns, veer-offs, and undershoots are: • Obstacle type, size, and location; • Aircraft size (wingspan) and speed; and • Number of obstacles and location distribution (shadowing). The basic approach is that presented in ACRP Report 3, as summarized in the ensuing sections. Additional details on how it was incorporated in the analysis are provided. The approach described in ACRP Report 3 was intended to model accident and incident consequences so that they could be combined with the probability of aircraft overruns and undershoots for an assessment of risk. The approach is rational because it is based on physical and mathematical principles. Modeling Approach for Risk The basic idea was to assess the effect of different obstacles at various locations in the vicinity or inside the RSA. The approach integrates the probability distributions defined by the location models with the location, size, and characteristics of existing obstacles in the RSA and its vicinity. The implementation of the approach required some simpli- fying assumptions so that it could be integrated with the fre- quency and location models. The following are the assumptions used: 1. Aircraft overrunning, undershooting, or veering off the runway will strike the obstacle in paths parallel to the run- way direction. This assumption is necessary to define the area of influence of the obstacle. 2. Four categories of obstacles are defined as functions of the maximum speed that an aircraft may collide with an ob- stacle, with small chances of causing hull loss and injuries to its occupants: a. Category 1: Maximum speed is nil (e.g., cliff at the RSA border, concrete wall). b. Category 2: Maximum speed is 5 knots (e.g., brick buildings). c. Category 3: Maximum speed is 20 knots (e.g., ditches, fences). d. Category 4: Maximum speed is 40 knots (e.g., frangible structures, ALS). 3. Severe damage and injuries are expected only if the aircraft collides within the central third of the wingspan and with a speed higher than the maximum for that obstacle category. The concept is explained in the ensuing section. C H A P T E R 4 Consequence Approach

4. The lateral distribution is random and does not depend on the presence of obstacles. This is a conservative assumption because there are events when the pilot will avoid the ob- stacles if he has some directional control of the aircraft. The accident/incident database contains a number of cases when the pilot avoided ILS and ALS structures in the RSA. The main purpose of modeling consequences of aircraft ac- cidents is to obtain an assessment of risk based on the likelihood for the worst credible consequence. It was not deemed neces- sary to develop a consequence model for each type of accident, as was done to model frequency and location. The approach can be used to address any of the five types of incidents included in the analysis. The basic idea is to use the location models to estimate the incident occurrences for which the aircraft will have high en- ergy when striking an obstacle, thus resulting in serious con- sequences. It should be noted that neither of the models used in the approach provides an estimate of the aircraft speed; however, using the location model and the average aircraft deceleration during a runway excursion, it is possible to infer the probability that the speed is above a certain level when reaching the obstacle. Figure 30 is used to illustrate the case for overruns and help understand the principle. This approach was introduced in ACRP Report 3. The x-axis represents the longitudinal location of the wreck- age relative to the runway departure end. The y-axis is the prob- ability that the wreckage location exceeds a given distance “x.” In this example, an obstacle is located at a distance D0 from the departure end, and the example scenario being analyzed is an aircraft landing overrun incident. Figure 30 shows an ex- ponential decay model developed for the specific accident scenario, in this case, landing overruns. There are three distinct regions in this plot. The first re- gion (green) represents overruns that the aircraft departed the runway but the exit speed was relatively low and the air- craft came to a stop before reaching the existing obstacle. The consequences for such incidents associated with that specific obstacle are expected to be none if the x-location is smaller than D0. The rest of the curve represents events that the aircraft ex- ited the runway at speeds high enough for the wreckage path to extend beyond the obstacle location. However, a portion of these accidents will have relatively higher energy and should result in more severe consequences, while for some cases the aircraft will be relatively slow when hitting the obstacle so that catastrophic consequences are less likely to happen. Using the location model, if x-location is between D0 and D0+Δ, it may be assumed that no major consequences are expected if the obstacle is present. The value of Δ is estimated based on aircraft deceleration over different types of terrain (paved, unpaved, or EMAS) and crashworthiness speed criteria for aircraft. It should be noted that Δ depends on the type of terrain, type and size of aircraft, and type of obstacle. Frangible objects in the RSA are less prone to causing severe consequences. It also should be noted that lighter aircraft may stop faster and the landing gear configuration also may have an effect on the aircraft deceler- ation in soft terrain, but these factors are not accounted for in this approach. Using this approach, it is possible to assign three scenarios: the probability that the aircraft will not hit the obstacle (green region—resulting in none or minor consequences); the prob- ability that the aircraft will hit the obstacle with low speed and energy (yellow region—with substantial damage to aircraft but minor injuries); and the probability that the aircraft will hit the obstacle with high energy (orange region—with substantial damage and injuries). For those events with low energy when impacting the ob- stacle, it is possible to assume that, if no obstacle was present, the aircraft would stop within a distance Δ from the location of the obstacle. The problem is then to evaluate the rate of these accidents having low speeds at the obstacle location, and this is possible based on the same location model. This probability can be estimated by excluding the cases when the speed is high and the final wreckage location is significantly beyond the obstacle location. To complement the approach it is necessary to combine the longitudinal and transverse location distribution with the presence, type, and dimensions of existing obstacles. The basic approach is represented in Figure 31 for a single and simple obstacle. Laterally, if part of the obstacle is within the yellow zone, as shown in Figure 32a, medium consequences are expected; how- ever, if any part of the obstacle is within the orange zone, as 24 D o o P{d < D o } = 10 0% - P {d > D o } o P low = P{ d > D o } - P {d > D o + } o + P high = P{ d > D o + } D o D o + o and D o + Distance x from runway end Pr ob ab ilit y d Ex ce ed s x is distance to Obstacle, d is distance the aircraft came to stop Prob of Incident with d < D Prob of Incident with d > D low energy Prob of Incident with d > D high energy Area (yellow) between D represents % occurrences at low speed (energy) when hitting obstacle (low consequences) P{ d > D o } P{d > D o + } Figure 30. Approach to model consequences of over- run accidents.

shown in Figure 32b, and the speed is high, severe consequences are expected. If the obstacle is off the orange and yellow zones, no consequences related to that obstacle are expected. In Figure 33, Obstacle 1 is located at a distance x1, y1 from the threshold and has dimensions W1 × L1. When evaluating the possibility of severe consequences, it is possible to assume this will be the case if the aircraft fuselage or a section of the wing close to the fuselage strikes the obstacle at high speed. Thus, it is possible to assume the accident will have severe conse- quences if the y location is between yc and yf, as shown in the figure. Based on the location models for lateral distance, the probability the aircraft axis is within this range can be calcu- lated as follows: where Psc = the probability of high consequences; b, m = regression coefficients for the y-location model; yc = the critical aircraft location, relative to the obstacle, closest to the extended runway axis; and yf = the critical aircraft location, relative to the obstacle, farther from the extended runway axis. P e e sc byc m by f m = − − − 2 Combining this approach with the longitudinal distribution approach and the possibility of multiple obstacles, the risk for accidents with severe consequences can be estimated using the following model: where N = the number of existing obstacles; a, n = regression coefficients for the x-model; and Δι = the location parameter for obstacle i. Multiple obstacles may be evaluated using the same prin- ciple. A shadowing effect also is taken into account when part of obstacle i+1 is behind obstacle i. Because it is assumed that aircraft will travel in paths parallel to the runway centerline, any portion of the obstacle located behind at a distance greater than Δi is disregarded from the analysis. A quantitative assessment of risk likelihood will be obtained as a function of operating conditions (aircraft, weather, runway distances available) and RSA dimensions and conditions (pres- ence of EMAS, presence, location, size, and type of obstacles, etc.). For the analysis, the user may select the alternative to evaluate the probability that an aircraft will go off the RSA or that severe consequences will take place. Implementation of Approach The implementation of the proposed approach is best ex- plained using one example. Figure 34 depicts an area adjacent to the runway end with two obstacles. The area isn’t necessar- ily the official airport RSA but any available area that can be P e e esc byci m by fi m i N a xi i n = − ( )− − = − +( )∑ 21 Δ 25 x y RSA y1 y2 x (dist. to obstacle) Figure 31. Modeling consequences. Obstacle 1/3 WS Wingspan (WS) a) Obstacle 1/3 WS Wingspan (WS) b) Figure 32. Lateral location versus consequences.

26 Lateral Location Probability Distribution Obstacle yc yf Psc w1 y sc Figure 33. Modeling likelihood of striking an obstacle. x Obstacle 2 - Tree Category 4 x1 x2 Obstacle 1 - Building Category 2 Cliff y Figure 34. RSA scenario with obstacles. used by an aircraft overrunning the runway end. The example shows the safety area surrounded by a cliff limiting its bound- aries. Obstacle 1 is not frangible and is classified as a Category 2 obstacle (e.g., building), maximum collision speed of 5 knots, located at distance x1 from the runway end. For this obstacle, the maximum speed without severe consequences is estimated to be 5 knots. A second obstacle is a small size tree classified as Category 4, maximum speed of 40 knots, and located at dis- tance x2 from the runway end. The remaining safety area is de- fined by the cliff surrounding the RSA and such boundary is classified as Category 1, maximum speed of 0 knots. The typical aircraft deceleration in unpaved surfaces is 0.22g, where g is the acceleration due to gravity (32.2 ft/s2). Using the relationship between acceleration, velocity, and distance, Δ can be calculated as shown in Table 6. The Δ values presented will be used to reduce the safety area so that only the effective portion where the aircraft may stop without severe damage is considered in the analysis. To perform the analysis, the frequency and location models are combined in a manner similar to that for the analysis without obstacles; however, the safety area is transformed to account for the presence of the obstacles, as shown in Figure 35. The area used to calculate the probability as a function of the aircraft stopping location is shown in green. It should be noted that the safety area in the shadow of Obstacle 1 is much larger than that for Obstacle 2 for three reasons: 1. Obstacle 1 is wider than Obstacle 2. 2. The maximum speed for striking Obstacle 1 (Category 2) is lower than that for Obstacle 2 (Category 4). Obstacle Category Max Speed (knots) (ft) (See Figure 30) 1 0 0 2 5 20 3 20 80 4 40 320 Table 6. Obstacle categories.

3. Obstacle 1 is located closer to the runway end, and aircraft speed is higher at this point than that at the location of Obstacle 2. The analysis will provide the probability that the aircraft will overrun the runway and the incident will have severe conse- quences, thus providing an estimate of risk. Additional Simplifications Additional simplifications were necessary to implement the approach. One such simplification was the use of maximum aircraft design group (ADG) wingspan instead of the actual air- craft wingspan. Without this simplification, a different safety area configuration would be required for each aircraft, greatly increasing the time to do the analysis. Using the ADG wingspan reduced the process to six steps, one for each ADG. A second simplification was also necessary to reduce the time to perform the analysis. Although the obstacles are cat- egorized according to the maximum speed to cause severe consequences, each type of aircraft will have a different max- imum speed. However, it would be very time-consuming to apply these differences in the calculations. Therefore, the maximum speed in the proposed approach depends only on the type of obstacle rather than the interaction between the obstacle and the aircraft. 27 x 2=20ft 4=320ft Cliff y ft 4= t Figure 35. Effective RSA for analysis.