Airbags in modern automobiles have been around for many years and their record in improving occupant safety in crashes is generally well accepted. However, many questions still arise regarding their deployment and performance in crashes. Some of these questions and the underlying issues, as well as statistical analyses of accident data are presented in this note.
Recent Airbag-Related Reports:
- A report from the Insurance Institute for Highway Safety (Braver et al, "How Have Changes in Airbag Designs Affected Frontal Crash Mortality?" IIHS, February, 2010) found that the newest airbags provide 'suboptimal protection' for belted drivers. This study concluded that the mortality rate for belted drivers was higher in airbag-equipped vehicles that comply with the law for frontal crash protection (FMVSS208, Advanced Airbag Rule) than in vehicles not certified as meeting this latest version of FMVSS.
- According to a recent story in the New York Times, airbag-related claims dominate the cases of accidents with injuries or fatalities as reported in the Early Warning Records submissions.
- Two reports in the Kansas City Star (Casey & Montgomery, "Airbags raise new alarms - Some didn't deploy in fatal crashes, newspaper learns", October 21, 2007; "Front airbags don't inflate in hundreds of crashes", October 22, 2007) estimated 1400 fatalities during the years 2001 to 2006 in frontal crashes due to non-deployment of frontal airbags.
- The estimate from NHTSA for the same period was of 576 fatalities occurring in frontal crashes where the front airbags didn't deploy
- Advances in airbags continue and new possibilities continue to be explored by manufacturers ("Air bag arms race: Carmakers see who can add most safety advances" http://www.detnews.com/apps/pbcs.dll/article?AID=/20070503/AUTO01/705030385/1148#ixzz0lrISmleW).
Airbag Issues in Field Incidents
The concerns regarding airbag-equipped vehicles generally fall into three categories:
- airbag(s) did not deploy in the crash where it was expected (by the vehicle occupants) to deploy ('non-deployment' claims);
- airbag(s) deployed in the crash where they were not expected to deploy and caused injury to the occupants ('unwanted deployment' claims);
- airbag(s) deployed properly in the crash but did not provide the expected amount of occupant protection ('deficient performance' claims).
In order to consider these claims, it is necessary to understand the engineering principles governing airbag deployment and performance. A brief and highly generalized overview is presented below. Due to the multiple types and configurations of sensors, airbags, and control units in the field, specific statements will require detailed accident analysis.
How Do Airbags Work? It is not generally understood that a vehicle's structure is the most important factor in making deployment versus non-deployment decisions. In a crash, the vehicle's structure experiences high deceleration (or acceleration) as the vehicle's velocity changes rapidly. Sensors attached to structural components (such as floor, pillars, etc) measure these responses and transmit them as electrical signals to a central control unit. This unit compares the gathered data to pre-programmed information and decides whether deployment is required. If deployment criteria are met, an electrical signal is sent to the inflator unit which ignites the propellants. The gas thus generated inflates the airbag. For frontal or lateral impacts, the bags remain inflated for a fraction of a second. For rollover airbags, the duration can be several seconds.
Front crash sensors are usually accelerometers attached to the vehicle's floor underneath the front seat(s). In many vehicles, the sensor and the control unit are integrated into one box referred to as a 'sensing and diagnostic module'. Some automobiles also incorporate additional forward sensor(s) that are attached to structural members underneath the hood and closer to the front end of the vehicle. These additional sensors are intended to detect localized impacts. Sensors for side airbags also are generally accelerometers that measure the lateral response at the attachment location. The sensors for determining imminent rollovers are often angular velocity or gyroscopic sensors, usually mounted on the vehicle floor near the center of gravity.
Airbag Design In Automobiles
Airbags cannot be designed independently. They need to be integral part of the entire automobile's design. Retrofit or aftermarket airbags may not provide the safety benefits. If necessary, stolen or cosmetically (only) damaged airbag units must be replaced in their entirety by identical unit from the same manufacturer for the same vehicle only after engineering confirmation that the repaired vehicle-airbag system will perform properly.
Since an airbag deploys only once and stays inflated for a very short duration, it is important that its deployment take place only when significant benefits are likely and that these be maximized in airbag-to-occupant interactions. The governing principles for vehicle-airbag integrated design are the following:
- The overriding requirement is that the airbag deployment must meet all applicable laws regarding such performance.
- In addition, airbags should:
- deploy in crashes when such deployment is likely to significantly reduce the severity of injuries suffered by the occupants;
- not deploy when no additional protection of occupants is likely from such deployment;
- not deploy in cases where such deployment itself may cause injuries.
Deployment and non-deployment criteria are developed by each vehicle manufacturer based on their own tests and are designed to maximize the advantages of deployment while minimizing unnecessary deployments
Bag's shape, volume, gas pressure and the rate of inflation are determined to a large extent by Federal Motor Vehicle Safety Standards as well as by other high-speed crash test regimen and ratings systems, such as the frontal and lateral NCAP tests by NHTSA and the frontal offset crash test and the side impact test conducted by the IIHS. In general, the higher the test speed at which airbag-equipped cars have to meet high-ratings requirements, the faster the required inflation speed of the bag ('faster deployed') and the higher is the total inflation pressure ('stiffer' bag).
Airbag(s) may be visualized as filling up the space between an occupant's body segments and the vehicle parts, thus reducing the relative impact velocity experienced by the body segments in the crash. Also, an airbag interposed between a moving body segment and a part of the vehicle distributes the impact forces over a wider contact area, thus reducing the peak acceleration that the body segment would otherwise experience.
Curtain airbags are designed to improve containment of the occupants inside the vehicle, reducing the likelihood of contacting injury-producing surfaces outside the vehicle.
These principles pose many engineering challenges because automobile manufacturers must design their products for occupants of all ages, all sizes and myriad different biometric properties. They also have to take into account the fact that occupants of an automobile may be wearing seatbelts or be unbelted and at the moment of crash, may be seated properly or in some unusual position. Most of the occupant biometric data is not unavailable in airbag deployment calculations and these decisions have to be based only on the data from the sensors.
Performance versus Expectations
Some of the reasons for deployment-related concerns may be the expectation that airbags will be deployed in all 'severe crashes', combined with a belief that visible damage to a vehicle is an indicator of crash severity. Other reasons for deployment-related concerns may arise from the existing accident reconstruction techniques of reporting the estimated 'ΔV at the vehicle's center of gravity' (which is defined as the change in the vehicle's speed during the crash) as the measure of the severity of in-plane crashes (rollovers are not considered to be in-plane). Since some publications have tried to establish statistical relationships between ΔV and occupants' injuries, expectations may exist that 'high ΔV' means a 'high severity crash' and vice versa.
The questions of interest are the following - Is ΔV the proper measure of a vehicle's crash severity? Should it be used as the deciding factor in airbag deployment decisions?
Crash Severity Definition: A definition of 'average' accident severity requires that both the ΔV and the time during which this ΔV occurs be defined. As an example, in a crash of an automobile into a rigid barrier at 56 kilometers/hour, the maximum ΔV (the sum of the initial velocity and the maximum rebound velocity) occurs at 70 milliseconds in the figure shown here. But, if the same vehicle had impacted a softer impact surface (such as a sand bank), the time to maximum ΔV would have been longer and the injuries to occupants probably much less. Thus, ΔV by itself is not a unique or complete measure of a vehicle's crash severity.
Published Research on Crash Severity, ΔV and Deployments: Let us examine some cases where better measures of crash parameters were available. A detailed study of
Figure 1: Frontal ΔV in Crashes with Airbag Deployments, May 2005-May 2006
airbag deployments and the role of ΔV has been published (Verma et al, "A Study of US Crash Statistics from Automated Crash Notification Data", Paper 07-0058, ESV Conference, 2007) and some results from this are described below. Shown in Figure 1 are the maximum values of ΔV for frontal crashes where airbags were deployed. It is seen from this figure that airbag deployments occur over a wide range of maximum ΔV. Also reported in the above paper were ΔV values for crashes where the airbag did not deploy but the ΔV reached a specified threshold (Figure 2). It may be observed that the ΔV values were as high as 75 kilometers/hour in some non-deployment cases.
Figure 2: Airbag Non-deployment - Maximum ΔV in Frontal Crashes
Figure 3: ΔV -versus- Time Record for Front Airbag Deployments
In order to relate these data to an average measure of crash severity, the ΔV was combined with the time information as shown in Figure 3. With this, each crash may be defined by an 'average deceleration' which is the slope of the line connecting the data point in Figure 3 to the 'time zero' on the x-axis. When an analysis of each individual case (one point in the plot) was conducted, it was observed that they all met the 'must deploy' criteria that had been established from crash tests. A similar evaluation of the non-deployment cases of Figure 2 has shown that none of the 'must deploy' criteria were found to have been violated in these cases although ΔV itself may have been high.
Statistical Analysis Of Field Data
Results from analysis of the National Automotive Sampling System - Crashworthiness Data System ('NASS-CDS') for the year 2008 are shown in Figure 4. The present analysis includes only vehicles with airbags and only cases with known ΔV. Cases of vehicle rollover and rear impacts were excluded. For the first curve in Figure 4 (shown as a continuous line) the plot of airbag deployment ratio was obtained by dividing 'the estimated number of cases where the airbag was
Figure 4: NASS-CDS 2008 Data - Airbag deployment ratio-versus-ΔV
deployed' by 'the total number of crashes in that range of ΔV'. It is observed that the lower the ΔV of the vehicle in the crash, the smaller the percentage of airbags deployed. As an example, for crashes with estimated ΔV in the range of 11- 20 kilometers/hour, airbags were deployed in only 44% of the cases.
The second curve in Figure 4 is of the ratio 'the number of crashes in the delta-V range' to the 'total number of estimated crashes'. It is observed that more than 50% of the estimated crashes have ΔV in the range of 11 to 20 kilometers/hour.
Analysis Of Airbag Operation In A Crash
Two aspects must be considered in deciding whether an airbag performed properly during a crash - its reliability as a system and its performance in occupant protection.
1. Reliability: Airbags and other occupant protection systems must have the highest possible degree of reliability. Such reliability is also implicit in US laws for crashworthiness (FMVSS) in the form of requirements that all sold vehicles pass mandated performance criteria. The reliability of a system is the product of the reliability of each constituent. Some of the constituents of 'airbag as a system' are:
- vehicle structure → integrity of the load path from impact point to sensor location, dynamic response properties;
- sensors → electromechanical integrity, functionality under operating conditions;
- electrical circuit & wiring → integrity and functionality under operating conditions;
- control unit → electric/electronic integrity, functionality in operating environment;
- inflator module → functionality, combustion properties;
- airbags → bag integrity, functionality in operating environment;
- mounting surfaces (steering column, instrument panel, etc.) → structural integrity, load carrying properties, functionality of deployment mechanism;
2. Performance: This implies that the airbag will optimize occupant protection when deployed. Airbags are 'supplemental restraints' and provide optimum safety benefit when the vehicle occupants properly wear seatbelts. The enhanced occupant protection from airbags comes from several factors:
- Vehicle Structure → primary mechanism for dissipating crash energy, determining vehicle motion, maintaining occupant compartment integrity;
- Interior parts of vehicle → seats determine occupant position and motion during crash; steering column and instrument panel affect the available space for airbag and impact velocity of occupants in front crashes; doors' interior and trim affect the response in lateral and rollover crashes, structure of roof, pillars and doors determine the airbag and impact locations in many crashes.
- Seatbelts → primary mechanism for reducing occupants' impact velocities and for keeping occupants in place for proper interaction with airbags.
- Pretensioners → these further enhance the effectiveness of seatbelts.
Dr. Mukul Verma is a well-known expert in Automobile Safety and Crashworthiness, Vehicle Structures, Product Design, and Statistical Analyses of Traffic Trends and Regulations . He has worked in many engineering and management positions at a major automobile manufacturer including assignments in R&D, vehicle design, analysis and testing and engineering program management.
See Dr. Verma's Profile on Experts.com.
©Copyright - All Rights Reserved
DO NOT REPRODUCE WITHOUT WRITTEN PERMISSION BY AUTHOR.