Accident Data & Statistics U.S. Rotorcraft Accident Data & Statistics Lee Roskop, FAA Rotorcraft Standards Staff International Helicopter Safety Team Scott Tyrrell, FAA Rotorcraft Standards Staff U.S. JHIMDAT Government Co-Chair January 23, 2013
Overview Accident Data and Trends Industry Sector Analysis Analysis by NTSB Finding Code Recent Concerns IHST Review SPSs, IRs, Occurrences U.S. JHIMDAT Current Actions Questions and Answers
U.S. Rotorcraft Accident and Fatal Accident Counts 16% decrease 152 accidents in 2009 to 129 accidents in 2011 but lost nearly all of that improvement with a 15% increase from 129 accidents in 2011 to 148 in 2012
U.S. Rotorcraft Accident & Fatal Accident Rates Estimated 11% increase in accident rate, 25% in fatal accident rate from 2011 to 2012 *Rates estimated using flight hours from the FAA’s GA & Part 135 Activity Survey
Current Accident Rates vs. IHST Goal IHST’s U.S. Goal by 2016: 1.8 accidents per 100,000 flight hours
Measuring Significance of Improvements 5 Year Snapshots by Month Plotted average accidents by month 2001-2005 Superimposed lines for 2 x standard deviation Observed progress based on 5 year increments Plotted 2006-2010 Plotted 2011-2015
Average Accidents/Month: 2001-2005
Average Accidents/Month: 2001-2005 with 2 x Standard Deviation The dashed blue lines represent 2 standard deviations above and below the red line.
Average Accidents/Month: Comparing 2001-2005 with 2006-2010 The lower the green line plots below the red, the better. Ideally we’d like to see the green line plot closer to the bottom dashed blue line.
Average Accidents/Month: Comparing 2001-2005 with 2011-2015 The 2011-2015 time period currently is only populated by 2 years, 2011 and 2012. Some nice reduction during the summer months, but stagnancy in many Spring and Fall months.
Industry Sector Analysis
IHST Industry Analysis: 2000, 2001, 2006: 523 Accidents 46.4% attributable to Personal/Private, Instructional/Training, and Aerial Application
FAA Follow Up On Industry Analysis: 2010-2012: 411 Accidents 61.3% attributable to Personal/Private, Instructional/Training, and Aerial Application
IHST Industry Analysis 2000, 2001, 2006: 84 Fatal Accidents Note that while Instructional/Training is still 3rd highest in the chart, the fatal accident percentage is much lower than the accident percentage from the previous chart. Not so for personal/private. Note also the location of Aerial Application in this pie chart for comparison on the next slide.
FAA Follow Up On Industry Analysis: 2010-2012: 69 Fatal Accidents
Accident Rate Estimates (per 100,000 flight hours) 2000, 2001, 2006 2010-2012 Personal/Private 31.99 Instructional/Training 7.84 Aerial Application 11.35 Helicopter EMS 3.70 Personal/Private 25.71 Instructional/Training 4.96 Aerial Application 16.99 Helicopter EMS 1.87 1. Flight hours from CYs 2000, 2001, 2006, 2010 extracted from FAA’s GA & Part 135 Activity Survey. 2. Flight hours from CYs 2011 and 2012 estimated using historical flight hour increase of 3% per year over CY 2010 hours. High number of accidents compared to low amount of exposure (flight hours) means high accident rates for personal/private, instructional/training, and aerial/application. Note the increase in aerial application rate in 2010-2012 estimates.
Fatal Accident Rate Estimates (per 100,000 flight hours) 2000, 2001, 2006 2010-2012 Personal/Private 6.27 Instructional/Training 0.60 Aerial Application 0.84 Helicopter EMS 0.93 Personal/Private 5.60 Instructional/Training 0.31 Aerial Application 1.46 Helicopter EMS 0.55 1. Flight hours from CYs 2000, 2001, 2006, 2010 extracted from FAA’s GA & Part 135 Activity Survey. 2. Flight hours from CYs 2011 and 2012 estimated using historical flight hour increase of 3% per year over CY 2010 hours.
NTSB Findings from 2008-2012 NTSB Methodology 0303403582 Assigns multiple “Findings” to each accident Since 2008, each “Finding” uses a 10 digit “Finding Code” Each 2 digits in the code drills down to a more specific factor For example: Analyzed most frequently used “Finding Codes” Codes in place for 479 of the rotorcraft accidents from 2008-2012 May show best opportunities for improving safety 0303403582 03 03 40 35 82 Environmental issues Physical environment Terrain Snowy/icy Effect on operation Only about half the 2008 accidents were assigned a finding code. For the other years, particularly 2012, there are accidents were the investigation is still open so a code has not yet been assigned.
Top 2 digit “Finding Codes” Personnel Issues: Cited in 77% of accidents Aircraft: Cited in 71% of accidents Environmental Issues: Cited in 40% of accidents
Top 10: 4 digit “Finding Codes”
Top 10: 6 digit “Finding Codes”
Recent Trends: 2012 closes poorly Oct-Dec 2012 Accidents 60% higher than same period in 2011 Highest Oct – Dec total in 5 years Tied, 2nd highest Oct – Dec total last 10 years Fatal Accidents 57% higher than same period in 2011 Tied, highest Oct – Dec total last 10 years Mitigation efforts by FAA, HAI, AOPA, and IHST Mass media effort in mid October after 4 fatal accidents in 8 days
Recent Trends: 2013 starts poorly January 2013 Record setting accident pace during first 9 days 7 accidents, 3 fatal accidents, 11 fatalities If sustained, pace was for 24 accidents, 9 fatal accidents, 38 fatalities Projection would have far exceeded highest total for any January the last 30 years Accident pace slowed during mid month Still projected to have a high total, but not record setting
What has changed? 5 years of steady progress Consistent improvement, despite lagging the IHST’s desired 80% accident rate reduction 4 consecutive months of regression Especially concerning since we are in the traditionally “lower accident” months
Recent FAA Safety Alert for Operators (SAFO) Increased manual handling errors for Air Carrier pilots Effects of modern aircraft technology Pro: Enables precision, improves workload management Con: Continuous use can lead to degraded pilot ability to quickly recover the a/c from an undesired state Recommendations: More training on when to use automated systems More exercise of manual flying skills Lesson: Technology and Training must compliment each other Cockpit advances should facilitate (not replace) basic pilot skills/decisions
IHST Review
US-JHSAT: Past Efforts Data set – NTSB Accident Dockets U.S. civil registered helicopters Type certificated (amateur built helicopters excluded) Includes ‘Public Use’ and restricted category Total of 523 helicopter accidents analyzed CY2000 - 197 accidents (Report issued IHSS 2007) CY2001 - 174 accidents (Report issued IHSS 2009) CY2006 - 152 accidents (Report issued IHSS 2010) Compendium Report – 3 year roll up/summary (Report issued 2011) We all know about the Jim Reason’s Swiss Cheese model. The JHSAT uses a similar concept. They analyze the NTSB accident docket and any other information that may be available for an accident. Google Earth is an excellent source to see how terrain may have factored in to an accident. The JHSAT does not second guess the NTSB’s Probable Cause, instead they identify where an action, or inaction, is documented that eventually led to the accident. These actions are then tabulated as a Standard Problem Statement (SPS). The thought being that each SPS is like a hole in the Swiss cheese. When the holes line up, the accident occurs. This theory is such that by covering that hole the accident may have been avoided. So what is the ‘best’ way of covering the hole? The team decides on the best Intervention Recommendation (IR) for that particular circumstance. When all accidents have been analyzed, the SPSs and IRs are totaled and prioritized by Frequency. Again, what are the most common problems that resulted in an accident and what are the most common Interventions that will have the greatest impact on the Accident Rate.
US Accidents by SPS Lvl 1 - Compendium
US Accidents by IR Lvl 1 - Compendium
US Accidents by Occurrence Types - Compendium
Loss of Control Details - Compendium
Intervention Recommendations - IR3 Top 10 from “Compendium” Report Install cockpit recording devices Improve quality and depth of NTSB investigation and reporting Autorotation Training Program Follow ICA procedures with confirmation of compliance Simulator Training - Advanced Maneuvers Personal Risk Management Program (IMSAFE) Training emphasis for maintaining awareness of cues critical to safe flight Better Mx QA oversight to ensure adherence to the ICA/Manual Mission Specific Risk Management Program Enhanced Aircraft Performance & Limitations Training
Current JHSIT Toolkits “Addresses” 8 of the Top 10 IRs Helicopter Flight Data Monitoring Toolkit Install Cockpit Recording Devices Training Toolkit Training Emphasis for Maintaining Awareness of Cues Critical to Safe Flight Autorotation Training Program In-flight Power/Energy Management Training ADM Training SMS Toolkit Mission Specific Risk Management Program Establish mission Specific SOP and Flight OPS Oversight Program Maintenance Toolkit Follow ICA Procedures with Confirmation of Compliance Fred Brisbois will talk in more detail about the JHSIT’s toolkits, but this chart shows how their toolkits address the JHSAT’s recommendations. Of the Top 10 we just looked at, their Toolkits cover 8 of them already.
From JHSAT to JHIMDAT Goals of analysis evolved in 2011 JHSAT complete: Identified problems and recommended interventions JHIMDAT initiated: Measure progress of implementation actions Economic realities = more resource constrained effort Economic downturn resulted in reduced industry participation Original meetings of once/month reduced to 3-4 times annually Detailed analysis on every accident not performed Time consuming process not feasible with more limited meeting times More strategic effort toward specific targeted areas Industry may need to lead in areas where deeper study is desired
Current US JHIMDAT Analysis Efforts Assess Progress by Analyzing Recent Years Data set from CY2009 – 2011 High level analysis has categorized: Industry where a/c is used Activity being performed when accident occurred Occurrence/Sub-Occurrence Code defined “What Happened?” Compare CY09-11 results to data in Compendium So where does the JHSAT go from here. First, the 3 years of data is being compliled into a single dataset. As the analysis methodology changed slightly over the years, this is not a trivial task, but is nearing completion. A compendium report will be available for next year’s IHSS in Fort Worth TX. The compendium will provide a more statistically relevant baseline of how helicopters were crashing pre-IHST. It will also provide the baseline that post-IHST safety enhancements have affected helicopter safety. The JHSAT will also use the Compendium Report to generate Pocket reports for specific missions and common occurrence types. We need to get the message out to all operators.
U.S. JHIMDAT Future Efforts Targeted Industry Outreach ALEA Ag Ops NTSB Finding Codes Compare findings from recent accidents to Compendium Improved tracking of System Component Failures Classify system failures using established taxonomies Air Transport (ATA)/Joint Aircraft System/Component (JASC) Update to Compendium, compare recent accidents
IHST Upcoming Action Items Launch Regional Teams North America Helicopter Safety Team (NAHEST)
Questions?
Back-up IHST Slides
US Accidents by Operation Most accidents occur in Part 91 operations. Not surprising, this is the FAR part most helicopters operate under.
US Accidents by Mission Most accidents occur in Personal or Training flights. This is where operating hours would help us refine the picture. Rather than the raw number of accidents, where is the highest accident rate. For instance, look at aerial application. In CY2000, Aerial Application represented 14% of all helicopter accidents in the US. By CY2006, it was only 4% of the total. Were they suddenly that much safer? Or were they flying less helicopters in that mission. Helicopter EMS on the other hand got progressively worse as the time went on. Or did it? The HEMS industry was growing during this period. Without flight-hour data, by mission, it is difficult to tell the whole story.
US Accidents by SPS Lvl 1 The US JHSAT started analyzing the CY2000 accident data. These NTSB had completed the investigations and had closed the dockets on most of these accidents. In addition, litigation on the accidents had been completed. This allowed the manufactures and operators the freedom to discuss the accidents in an open forum. After CY2000 was complete the JHSAT started on CY2001 for the same reasons and then began the CY2002 data set. The story, as you can see by this chart, was relatively constant. Rather than complete CY2002, the JHSAT, with the approval of the IHST ExComm, went to the last year of accidents prior to IHST beginning; CY2006. If CY2006 looked the same as 2000 and 2001, the intervening years could be assumed to follow the same pattern. For the 523 accidents analyzed, Pilot Judgment and Actions, Data issues, and Safety Management Systems were the three most frequently cited issues.
US Accidents by IR Lvl 1 Likewise, Training, Data Information (including HFDM), and SMS were the 3 most frequently cited Intervention recommendations.
US Accidents by Occurrence Types When the JHSAT first started analyzing accidents, we assumed that widely varying missions, would have their own unique sets of Problems and corresponding recommendations. The JHSAT was surprised at the end of the analysis to see that 80% of the recommendations were NOT specific to any mission. This presented a problem to the Implementation team. The former Co-Chair of the US-JHSIT asked the JHSAT to tell him HOW the helicopters were crashing. This led to a new form of analysis. Instead of analyzing by mission, the JHSAT began analyzing by Occurrence Type. The JHSAT used a modified CAST-ICAO Common Taxonomy Term (CICTT) to identify what type of accident occurred. In this type of analysis, an accident can have more than one occurrence type associated. For instance, if the T/R driveshaft fails, the pilot may attempt an autorotation. By virtue of being in the accident database, the autorotation was probably not successful. If the pilot lost control and crashed the helicopter, there would be three occurrence type associated. Coincidently, the chart shows Loss of Control, Autorotation, and System Component Failure as the top three occurrence types in the US analysis.
Loss of Control Details To assist the JHSIT in coming up with effective Safety Enhancements, the JHSAT subdivided many of the occurrence types. There are many ways to ‘Lose Control’ and each may have a different way to prevent it in the future. Performance Management, allowing the rotor RPM to fall off for example, was the leading cause of Loss of Control. But look at number 4 – Interference with Controls. These are vastly different.