2016 ARC- - CSI CONF.
ollision C The International Compendium for Crash Research
Volume 11 Issue 2
Motorcycle Crash Testing
Evaluation of Slam Stick for Data Acquisition Rollover Scene Evidence Low-speed, Rear-end Crash Analysis collisionmagazine.com Cover.indd 1
1/23/2017 2:11:42 PM
Volume 11 Issue 2.indd 2
1/23/2017 2:01:05 PM
Contents
Volume 11 Issue 2
inside 6
Letter From The Editor
38
features 22
The Accelerations Present During The Trip Phase Of A Soil-Tripped Rollover Crash – An Update
by: Nathan A. Rose, Neal Carter, and Gray Beauchamp 30
My Turn At The Wheel: Principal Direction Of Force Real Or Abstract?
by: Erik Carlsson 38
50
Impact Speed Determination In A Head-On Crash Using Bayesian Networks
by: 2106 ARC-CSI Crash Team Boot Camp: Gary Davis, Andre Doria, Michael Morris, and Randy Eldridge 50
Evaluation Of The MIDE Slam Stick X As A LowCost Accelerometer And Data Acquisition System For Vehicle Crash Testing
by: David M. Hallman, Robert D. Anderson, Wesley Grimes, Michael DiTallo, Kenneth Salisbury and Kevin Vosburgh 64
Low-Speed Rear-End Crash Analysis MARC1 Use In Test Data Analysis And Crash Reconstruction
86
by: Rudy Limpert, Ph.D. 72
Piston-Type Bumper Isolator Compression Versus Delta-V In Bumper-To-Barrier And Bumper-tTo-Bumper Impacts
by: Robert D. Anderson, Michelle R. Hoffman Russell L. Anderson and Michael Rosenfield 86
The Longevity Of Scene Evidence From A Rollover - A Case Study
by: Nathan Rose and Neal Carter 96
Motorycycle Crash Testing: Advanced Boot Camp Was Born
by: Michael DiTallo, Thomas Green, Wes Grimes, Ken Salisbury, Brent Munyon, Tom Lawson, James Whelan, Eric Moody, Kevin Vosburgh, and Eric Paul
96
case study 8 110
Evaluating Crash Data From Late Model GM Vehicles by: W. R. Rusty Haight Case Problem Solution www.collisionpublishing.com
Volume 11 Issue 2.indd 3
Collision Magazine - Volume 11 Issue 2 3
1/23/2017 2:01:09 PM
dy
tu e S
Cas
Evaluating Crash Data from late model GM vehicles w. R. Rusty Haight Collision Safety Institute
T
he case problems for this issue of Collision are composed of two parts, based on data sets from crashes involving two recent model GM vehicles. The theme for the case problem set is: “this is a reading intensive technology.” The following pages include select data tables and graphs from the larger reports and, as the reader can see, some pages and data tables one might normally find have been left out to keep the space in Collision manageable but still provide the reader sufficient information from which to evaluate and work the case problem as presented. Undeniably, other analysis might be undertaken; however, for the limited scope of these case problems, there is sufficient information provided. We should start the analysis with the idea that to be able to effectively use CDR Tool retrieved data, one needs to first answer the question: is the retrieved data in this report from the crash under investigation? In other words, can you associate the data you have, as it is presented, to the crash you’re working on? The Verano Case Problem The first case problem involves a set of data from a 2013 Buick Verano. For this case problem, you may assume that you’ve seen the vehicle and it has some
88 Collision Collision Magazine Magazine - Volume - Volume 11 Issue 10 Issue 2 2
Volume 11 Issue 2.indd 8
side damage, specifically damage to the right side of the vehicle just to the rear of the B pillar. You have examined the vehicle and find no airbags deployed; however, the driver and passenger pretensioners have deployed and when you turn the key “on” during your vehicle examination and imaging, you see the SIR warning lamp is “ON.” One can see that the data set for the Verano reflects a single event recovered and it is labeled (on the first page of the report) a “Non-Deployment.” In the “System Status at Event” data table, it reads: “Event Record Type = Non-Deployment.” However, we also see that there is a B0052 “DTC present at Time of Event” and we find this passage associated with that DTC data element in the Data Limitations: “…- For Deployment Events, DTC B0052 (Deployment commanded) shall be recorded with the remainder of the data for this event even though it occurred after Event Enable. …” Then, in the Event Data (Record 1) data table we find the data element “Pretensioner Deployment, Time to Fire, Driver (Driver Pretensioner Time From Time Zero to Deployment Loop #1 or Loop #2 Com-
www.collisionpublishing.com www.collisionpublishing.com
1/23/2017 2:01:11 PM
mand Criteria Met) (msec) = 5ms” (and the same for the passenger pretensioner). So, all this would seem to beg the question: 1. Is this data set, the crash record retrieved using the CDR Tool, really representative of a “Deployment” or a “Non-Deployment” Event? (You might imagine yourself on the stand having to explain this so a simple “yes” or”no” wouldn’t be sufficient here.) Answering this question leads us toward an evaluation of whether or not the data is from the crash you’re working on. The Corvette Case Problem The previous case problem (for the Buick Verano) in this issue sets the stage for this more complex analysis of a crash involving a 2014 Corvette from which the second data set presented here was retrieved. For this case problem, in addition to the CDR Tool retrieved data, we are provided with a video of the crash from a nearby home security camera (select screen captures are provided in this issue of Collision) as well as a rough diagram of the motion of the Corvette into the crash (prepared for Collision magazine by Tom Potter). From an examination of the vehicle and scene, we find that damage to the right front wheel/tire of the Corvette, damage to its right rear wheel and tire, some damage to the rear of the vehicle at the rear bumper and damage to the left side of the car, at about the A pillar from contact with a light pole (as sen in the video). We find that there were no airbags deployed although the driver and front passenger seat belt pretensioners have been deployed. From an evaluation of the scene and vehicle information available (setting apart from the CDR Tool report for the moment) together with a review of the security video footage, we would conclude that the Corvette was driving through a left hand curve in the road and began to yaw counter clockwise from its lane of travel across the oncoming lanes of travel toward the far curb. It struck the curb on the far side of the opposite lanes of travel first with its right front wheel/tire. Then, continuing to rotate counterclockwise, it struck an embankment beyond the curb with the back end of the car. The yaw largely stopped or at least slowed and the Corvette moved more laterally, right side lifting more than left, with the left side
presented until it struck a light pole with the left side of the car at about the A pillar. It came to rest at the base of the pole. The CDR Tool report shows three recorded events, two Non-Deployments and a Deployment. They are recorded in three records (Event Record 1, 2 and 3). While there may be a number of methods one might use to evaluate this data in terms of the order of steps in that evaluation, these questions are posed for this case problem not so much to guide the analysis in a specific order but ultimate to guide the solution as presented later in this issue of Collision. 1. What is the chronological order of the events recorded? Meaning, the data is presented as “Event Record 1 (then 2 and 3) but is that the chronological order in which they occurred? 2. In the data table “System Status At Time of Retrieval” we see these data elements: “Deployment Event Counter,” “Multi-Event, Number of Events (Event Counter)” and “OnStar Notification Event Counter.” How do these data elements relate to the events under investigation? Meaning, it would appear that there have been 4 lifetime events observed by the airbag control module (Multi-Event, Number of Events (Event Counter) = 4) although there are only three recorded. You should be ready to have to explain which events are recorded and, since only three are recorded, what happened to the 4th event? 3. Taken in light of the previous case problem involving the Verano, the data in Event Record 3 for the Corvette is described as a “Deployment;” however, like the data in the Verano case, the only devices commanded for a deployment appears to be the driver and passenger belt pretensioner. One should expect to have to address that apparent conflict or at least “difference” in terms of the naming convention. Like the Verano data set: is this data set, the crash record retrieved using the CDR Tool, really representative of a "Deployment" or a "Non-Deployment Event?”
Editor's Note: The following pages are excerpts from the related CDR reports. The case study solution can be found at the end of this issue.
www.collisionpublishing.com
Volume 11 Issue 2.indd 9
Collision Magazine - Volume 11 Issue 2 9
1/23/2017 2:01:12 PM
ed w e i v e R Peer
The Accelerations Present during the Trip Phase of a Soil-Tripped Rollover Crash – An Update
I
Nathan A. Rose, Neal Carter, and Gray Beauchamp Kineticorp, LLC
ntroduction In Volume 11, Issue 1 of Collision, we examined longitudinal accelerations present during the trip phase of a fullscale rollover crash test reported by Asay and Wooley in 2010 (Test #2 in SAE Paper 2010- 01-0521). The trip is the phase during which lateral forces applied to the tires or wheels on the leading side of the vehicle cause the trailing side tires to lose contact with the ground. Eventually, these forces cause the center of gravity to roll past the leading side tires and the vehicle begins to roll over. Historically, analysis of the trip phase has focused on the lateral forces that generate the rollover (see Cooperrider, 1990 and 1998, for instance). However, the vehicle in Asay’s test experienced both longitudinal and lateral deceleration during the trip phase. The brakes were not applied on the test vehicles during these tests until well after the rollover had commenced, so the longitudinal accelerations had to have been generated from the tire interaction with the soil.
22
Collision Magazine - Volume 11 Issue 2
Volume 11 Issue 2.indd 22
www.collisionpublishing.com
1/23/2017 2:01:24 PM
In our first article on this topic, we reported multiple sets of speed calculations using the principle of conservation of energy. Some of these calculations neglected the longitudinal deceleration and others incorporated them. These calculations demonstrated that, at least for this test, it was important to include the longitudinal deceleration during the trip phase in order to calculate an accurate vehicle speed at the beginning of the loss of control. Of course, a conservative speed estimate could be obtained without including the longitudinal deceleration. In this particular test, which utilized a four-wheel drive 1991 Isuzu Rodeo, the test vehicle was shifted into neutral and towed up to and released from a speed of 73.5 mph (118.3 kph). After release, the vehicle was steered, using an automated steering controller, with a sharp left steering input of approximately ¼ turn that caused the vehicle to travel to the left across the roadway and to yaw counterclockwise. One second later, the vehicle was
steered back to the right until the steering stop for the vehicle was reached. This steering input was maintained for the remainder of the test. The vehicle then continued off the left side of the road into the dirt, but reversed its yaw direction, developing a significant slip angle as it yawed in a clockwise manner. The vehicle deposited furrows in the dirt and then began rolling over. The vehicle rolled 7 times over 181 feet (55.3 meters). During a course on rollover crash reconstruction in September 2016, Wes Grimes pointed out to us that the longitudinal forces present during the trip phase in this test were likely due to the fact that the front tires of the test vehicle were steered significantly to the right during the trip.1 Because of the steer input, a lateral force applied to these front tires would apply both a lateral and longitudinal force to the vehicle. Figure 1 contains nine frames from the video of this test that depict the trip phase. The rightward steer of the front wheels is evident in these frames of video.
www.collisionpublishing.com
Volume 11 Issue 2.indd 23
Collision Magazine - Volume 11 Issue 2 23
1/23/2017 2:01:25 PM
rn u T My
eel h W he T t a
Principal Direction of Force Real or Abstract? Erik Carlsson
30
Collision Magazine - Volume 11 Issue 2
Volume 11 Issue 2.indd 30
www.collisionpublishing.com
1/23/2017 2:01:26 PM
Excerpt from a transcript of a deposition: Attorney: What was the PDOF at the collision? Deponent: It hadn’t been determined yet when the vehicles collided. Attorney: What do you mean, hadn’t been determined yet? Deponent: PDOF is a specific angle an accident reconstructionist assumes or determines when analyzing a traffic accident. Attorney: Why would he assume an angle? Deponent: To simplify the reconstruction of accidents where the colliding vehicles change direction during the impact, and the collision force therefore covers a range of angles instead of having a specific direction.
I
believe many accident reconstructionists agree with the explanation given by the deponent quoted above, though I doubt that many would be as straight forward as a defense expert witness was when asked at his deposition how he had established the PDOF he used in his computer simulation: You pick a number to start with, then change, and change, and change, until you get it right. Naturally, by selecting a single direction of force rather than to use a range of directions, we greatly simplify the reconstruction of traffic accidents where the trajectories of the vehicles play an important role in the analysis. But if we presume that a single PDOF can truly enable an accident reconstructionist to determine an accident vehicle’s movements at a collision, can the same PDOF also be used to determine the movements of occupants inside the vehicle during the impact, if they are assumed to be unrestrained? This article is about a fatal traffic accident case where a defense expert witness postulated that the PDOF he claimed applied to the collision force also applied to the inside of the accident vehicle, a heavy truck. He could therefore determine that the truck driver, who he claims was not wearing the seatbelt, ended up on the passenger side of the cab when the truck hit a concrete wall with the its right front corner. [The driver was burned to death when he jumped out of the truck after it was engulfed in flames following impact with a concrete road divider.] www.collisionpublishing.com
Volume 11 Issue 2.indd 31
Collision Magazine - Volume 11 Issue 2 31
1/23/2017 2:01:27 PM
n o i t a n i m r e t e D g d n i e e s p u S h t s c a a r p C m n I O s d k a e r in a H ian Netwo s e y a B
ed w e i v e R Peer
38
Collision Magazine - Volume 11 Issue 2
Volume 11 Issue 2.indd 38
www.collisionpublishing.com
1/23/2017 2:01:30 PM
A
Gary Davis, Andre Doria, Michael Morris, and Randy Eldridge
s in years past the 2016 ARC-CSI conference offered a Crash Test Boot Camp experience, where a select group participated in the conference crash testing. This article analyzes the head-on crash test between a 2011 Ford Crown Victoria and a 2016 Volkswagen that was assigned to the Boot Camp team (2016 ARC-CSI Crash Test #2). The target impact configuration was a collinear head-on impact with impact speeds for both vehicles at 48.5 kph (i.e. 97 kph closing speed). Both vehicles were remotely operated by Phantom Driver systems and instrumented with IST accelerometers, V-Box video/GPS units, and data collection systems to record crash pulses and impact speeds. Vehicle specifications and test instrumentation data results are summarized in Table 1. www.collisionpublishing.com
Volume 11 Issue 2.indd 39
Collision Magazine - Volume 11 Issue 2 39
1/23/2017 2:01:31 PM
ed w e i v e R Peer
Evaluation of the MIDE Slam Stick X as a Low-Cost Accelerometer and Data Acquisition System for Vehicle Crash Testing. David M. Hallman, Robert D. Anderson, Wesley Grimes, Michael DiTallo, Kenneth Salisbury and Kevin Vosburgh
50
Collision Magazine - Volume 11 Issue 2
Volume 11 Issue 2.indd 50
www.collisionpublishing.com
1/23/2017 2:01:37 PM
bstract The purpose of this study was to evaluate the viability of the Slam Stick X (Figure 1), manufactured by MIDÉ as a self-contained data collection system and accelerometer for crash testing. Two Slam Sticks were purchased and installed in crash tested vehicles alongside the “typical” higher cost equipment used by the 2015 Southwestern Association of Technical Accident Investigators (SATAI) and 2016 ARC-CSI crash teams. The collected acceleration data sets from both systems were processed as usual and the filtered acceleration data and Delta V data from the various instrumentation was compared to the Slam Stick data.
A
The initial testing and data comparisons from the 2015 SATAI conference revealed significant differences between the collected and processed data from the Slam Stick and the reference equipment. The Slam Stick acceleration values were significantly lower than data collected utilizing the reference equipment. The Slam Stick Delta V graphs were also different than the normal or expected Delta V graphs. This difference was found to be an artifact of the type of accelerometers utilized in the Slam Stick. The stock Slam Stick contains a piezo-electric accelerometer while the typical systems for vehicle crash testing use piezo-resistive or MEMS (Micromachined ElectroMechanical System) accelerometers. The important difference between the two is that piezo-electric accel-
erometers experience low frequency roll-off and piezoresistive accelerometers do not. The stock Slam Sticks were upgraded with a capacitive MEMS accelerometer by MIDÉ and retested alongside IST 3C and IST 3D data collection equipment at the 2016 ARC-CSI Crash Conference. While there are still some challenges with the Slam Stick’s use, good agreement was found between the two data sets. ntroduction Crash testing and collecting acceleration data from crash tests is necessary to have a good understanding of the forces generated during vehicle impacts. This in turn provides crash reconstructionists and engineers with empirical correlations between accelerations, Delta V, PDOF and occupant injuries. High frequency data acquisition equipment that is reliable and robust is required to collect accurate acceleration data. The cost of the typical data acquisition equipment can be a significant barrier to entry for those who would benefit from being able to conduct crash testing and data collection for training and research purposes. Traditional data acquisition systems can cost $1,500 to $2,500 per channel with an initial expense of $7,500 to $15,000 depending on the system. This significant expense for equipment that will likely see limited use exceeds the resources of many individuals, companies and public agencies. The expense of the typical system led to the search for a viable, lower cost acceleration data measurement and collection option.
I
Figure 1: Slam Stick X with aluminum and plastic enclosures
www.collisionpublishing.com
Volume 11 Issue 2.indd 51
Collision Magazine - Volume 11 Issue 2 51
1/23/2017 2:01:38 PM
ed w e i ev R l ria o t i Ed
Low-Speed Rear-End Crash Analysis MARC1 Use in Test Data Analysis and Crash Reconstruction Rudy Limpert, Ph.D.
I
ntroduction As indicated in Short Paper PCB1-2015, I am publishing another paper of low speed rear-end crash test analysis. I hope a detailed MARC1 analysis of specific crash tests may be of help to those who are new to MARC1 and those experts who may need a refresher. Crush depth measurements are not easily obtained from photographs. In many tests crush depth values are not published, and worse, often not even measured. I suspect that most reconstruction experts do not analyze low-speed crashes, and hence, experts are not familiar what input data are required to calculate impact speed. Conducting MARC1 crash tests- on-paper will in many cases shed helpful light on the facts involved. Photographs and measurements of crush depth are always important for an accurate reconstruction in low-speed cases.
Low-Speed Rear-End Crash Test: Nissan Micra K11 v. VW New Beetle
The crash damage of the Nissan and VW are shown in Figures 2 and 3 below. The crush damage of the Nissan seems to be more extensive than that of the VW. No crush measurements were published.
Figure 2: Nissan Frontal Damage
A Nissan Micra K11 rear-ended a VW New Beetle in an inline low speed test. The impact configuration is shown in Figure 1. The impact speed was 13.91 mph. The Nissan was braking at 0.59g deceleration during impact. The VW was stationary with its transmission in neutral.
Figure 3: VW Rear-end Damage. Figure 1: Nissan-VW Rear-end Pre-Crash Impact Configuration 64
Collision Magazine - Volume 11 Issue 2
Volume 11 Issue 2.indd 64
www.collisionpublishing.com
1/23/2017 2:01:46 PM
Published Crash Data
Crash Test-on-Pape Using MARC1 – W2
The AGU Crash Datenbank (www.agu.ch) data were published on the internet as shown below. When I first saw the data table, my curiosity was peaked. The data were exactly what a MARC1-2015 – W4 low-speed rear-end collision analysis requires as input data such as crush energy, energy equivalent speed and stiffness values (lb/ft). Crush depth data were not given. How accurate the data ranges were could only be answered through MARC1-W4 analysis,
We first run a simple crash-test-on-paper with MARC1W2., which does not include any braking during impact. MARC1-W2 analyzes an inline collision with the use of conservation of energy across the impact, conservation of momentum across the impact (no external forces are considered such as tire-to-ground braking forces), and the coefficient of restitution.
The German-language data are summarized in Table 1 with imperial units.
The results show an acceptable data comparison when braking is excluded (W2) as shown below in Table 2.
The acceleration and velocity time histories are shown below for the reader to gain additional understanding of the low-speed inline crash test dynamics.
www.collisionpublishing.com
Volume 11 Issue 2.indd 65
Collision Magazine - Volume 11 Issue 2 65
1/23/2017 2:01:46 PM
ed w e i v e R Peer
Piston-Type Bumper Isolator Compression versus Delta-V in Bumper-to-Barrier and Bumper-to-Bumper Impacts Robert D. Anderson, Michelle R. Hoffman Russell L. Anderson, Michael Rosenfield
72
Collision Magazine - Volume 11 Issue 2
Volume 11 Issue 2.indd 72
www.collisionpublishing.com
1/23/2017 2:01:51 PM
I
ntroduction In low speed collisions involving vehicles’ bumpers, the evidence of the amount of piston-type bumper isolator compression has been correlated with the severity of the collision. The extent of isolator compression is in effect a record of the magnitude of the collision force. For vehicles with isolators that have been involved in a low-speed collision, comparing the extent of isolator movement to that produced in collision tests of known impact severity is a reliable and accepted method for assessing impact severity. Since the mid-1990’s the use of relationships developed for each automobile manufacturer has been common.
Although many manufacturers have moved away from piston-type bumper isolators, BMW and Audi/ Volkswagen equipped some of their automobiles with isolators up to the 2006 and 2009 model years, respectively. Ford had isolators on some of their vehicles as recently as 2011, and Porsche had isolators on some of their vehicles as recently as 2013. Starting in the 1990’s, low speed impact analyses commonly included evidence of the extent that isolators compressed during the impact. Isolator-equipped bumpers were common on 1990’s automobiles and much of the published data relating to isolator compression versus ∆V is from that era (Szabo and Welcher, 1992; King, et al., 1993; West, et al., 1993; Malmsbury, 1994; Siegmund, et al., 1994; Bailey, et al., 1995; Siegmund, et al., 1996; King, et al., 1997; Anderson, et al., 1998; and King, et al., 1999). King, et al., (1993) concluded from static and dynamic bumper compression tests that isolator compression generally increased with speed change and that testing for each vehicle was the most accurate method to arrive at a ∆V for a particular amount of isolator compression. The largest and most comprehensive data set of testing on isolator-equipped vehicles was published by Siegmund, et al., (1994). Specific average isolator compression-∆V trends for vehicle manufacturers were identified, which reduced the need for thorough testing of each specific vehicle. It was observed that struck vehicle braking, and/or additional weight from passengers and cargo produced over-estimates of the struck vehicle’s ∆V.
Both Siegmund et al., (1994) and Bailey et al., (1995) found that using isolator compression data from barrier impacts may over-estimate the ∆V in car-to-car impacts for some vehicles and that care must be exercised to ensure that the isolators are not bottomed out. Siegmund, et al., (1996) found that the direct barrier and the momentum, energy and restitution methods provided reasonable predictions of the actual ∆V sustained by vehicles in car-to-car collisions, but that the quasi-static isolator force method did not. While investigating more affordable methods to obtain isolator compression-∆V data, King, et al., (1997) found that vehicle characteristics, such as suspension properties, center of mass location, as well as the bumper mounting conditions including the bumper cover attachment may affect the isolator compression∆V relationship. King, et al., (1999) concluded that single isolator testing could be used to develop the relationship between isolator compression and vehicle speed change for some, but not all, types of isolators. In practice, the manufacturer trends identified by Siegmund et al. (1994) have been commonly used in low speed impact analysis since the mid-1990’s. The majority of all the data cited above was obtained from 1980’s vintage automobiles. Recent crash test results have shown an apparent shift in the isolator compression versus ∆V trend for Ford automobiles (Ruberte, et al., 2014), suggesting that using the data from Siegmund et al., (1994) resulted in an underestimate of the vehicle ∆V. The purpose of this paper is to present data to help bridge the gap between the historic and newer data for Ford, BMW and Volkswagen, and to present a method for measuring real time isolator compression.
M
ethodology A series of low speed impacts were conducted on vehicles with piston-type bumper isolators. Both front and rear bumper impacts were conducted on 4 different vehicles (identified in Table 1) to determine the isolator stroke as a function of vehicle velocity change for vehicles manufactured during the 1990s.
Before testing, each vehicle was weighed with CAS Wheel Weigher scales, which have a 10,000 pound per wheel capacity in 5 pound increments. The test weights are shown in Table 1. It is noteworthy, that
www.collisionpublishing.com
Volume 11 Issue 2.indd 73
Collision Magazine - Volume 11 Issue 2 73
1/23/2017 2:01:53 PM
ed w e i v e R Peer
The Longevity of Scene Evidence from a Rollover A Case Study
I
ntroduction
As with most any type of crash, reconstructing a rollover crash often begins with the development of a scaled diagram depicting physical evidence from the crash. This evidence will often include tire marks on the roadway, tire furrows in an off-road surface, gouges and scrapes, wheel landing marks on asphalt or in the dirt, vehicle-to- ground impact marks on an off-road surface, vehicle window glass deposits, and other debris from the vehicle. This evidence enables the reconstructionist to determine the location, speed, and timing of the loss of control, the trip, and the rollover. Analysis of the motion of occupants who were ejected from the vehicle can also be carried out. Locating evidence on a scaled diagram, of course, depends on the evidence being documented in some form (photographs or measurements) before that evidence disappears. This evidence documentation may be completed either by investigators on scene immediately after the crash or by later investigators who may come to the site long after the crash occurred. Of course, some types of evidence deteriorate more quickly than others and, at some point, many pieces of evidence will become unavailable for direct documentation by later investigators.
86
Collision Magazine - Volume 11 Issue 2
Volume 11 Issue 2.indd 86
www.collisionpublishing.com
1/23/2017 2:01:58 PM
The longevity of tire marks, off-road furrows, and impact marks on an off-road surface depends on many factors, including the initial prominence of the marks, the type of asphalt or off-road surface, the traffic volume through the area, the weather, and the frequency of roadway and off-roadway maintenance and modification. Gouges and scrapes in an asphalt surface can be long-lasting as long as a road is not repaved or modified in a way that removes them. Vehicle window glass deposits can be a long lasting piece of evidence from a rollover. Przybyla [2015] studied the longevity of glass debris fields over a period of two years. He placed glass debris fields at three locations in the United States where they would be subjected to the influence of slope, weather, mowing operations, and other factors that could result in movement of the debris field. Each area of glass debris was periodically mapped to determine the degree to which the glass debris had moved or changed in size.
Nathan Rose and Neal Carter Kineticorp,LLC www.collisionpublishing.com
Volume 11 Issue 2.indd 87
Collision Magazine - Volume 11 Issue 2 87
1/23/2017 2:01:59 PM
dy
tu e S
Cas
Case Study Solution:
Evaluating Crash Data from late model GM vehicles w. R. Rusty Haight
Collision Safety Institute
T
here may be different ways to look at a given case where different people would reasonably come to the same ultimate conclusions and the solutions to the case problems offered here should not be taken as the only way to address the use of CDR Tool retrieved crash data in a more general sense. The solutions offered here are specific to the data presented in these cases keeping in mind that the theme for the case problem set is: “this is a reading intensive technology.”
that is needed to record a Non-Deployment Event, is five MPH [8 km/h]. …”
The Verano Case Problem
-Battery Cut-Off Deployment …”
The question posed for the Verano data set was: is this data set, the crash record retrieved using the CDR Tool (from the Verano), really representative of a “Deployment” or a “Non-Deployment” Event? While this focuses on a seemingly narrow topic, in a larger sense, one should see the question leading toward evaluating whether or not the data is or can be associated with the crash under investigation. But more specifically to this first case problem, it pushes one toward really reading the CDR Tool report in totality together with a background of related references and influential documents rather than picking and choosing one or two elements and “running with it.” A large part of really understanding what’s going on here is, of course, to look at the Data Limitations text for this report as a starting place. The Data Limitations text for the Verano data set includes, among others, these two passages: “… There are two types of recorded crash events for Front, Side, and Rear (FSR) Events. The first is the Non-Deployment Event. A Non-Deployment Event records data but does not deploy the air bag(s). The minimum SDM Recorded Vehicle Velocity Change,
110 110 Collision Collision Magazine Magazine - Volume - Volume 11 Issue 10 Issue 2 2
Volume 11 Issue 2.indd 110
“…Also, a Non-Deployment event can be recorded if one of the following occurs without the Deployment of any of the frontal air bags, side air bags, or roll bars: -Pretensioner(s) only Deployment -Head Rest Deployment
In the first passage, perhaps the part which should stand out to the CDR Tool data user in this example is: “... A Non-Deployment Event records data but does not deploy the air bag(s). ...” In the context of the idea that “this is a reading intensive technology,” another axiom to apply here might be: “what it doesn’t say is often just as important as what it does say” (with respect to this technology and in particular in the Data Limitations or in the body of a report). The passage tells us, for certain types of events, what a NonDeployment Event is and specifies that it’s where the airbags are not deployed but it does not include or reference seat belt pretensioners. In the second paragraph - in what might be seen as amplification or maybe something of a clarification of the first entry - we find that an event where there were there was a “Pretensioner(s) (only) Deployment” without the deployment of any airbags, that would be recorded and labeled a “Non-Deployment.” It would seem to be clear that the Data Limitations text is telling us (1) a Non-Deployment Event record is a set of data where there were no airbags deployed and, at the same time, (2) a Non-Deployment Event is one without airbags but with a pretensioner deployment. With the use of the phrase “pretensioner (only) deployment,” the apparent disconnect seems to ask the
www.collisionpublishing.com www.collisionpublishing.com
1/23/2017 2:02:14 PM
question: “so when is a ‘deployment’ a “Non-Deployment’?” (One clue here might be found in the select capitalization of the words.) One might think it is almost like a riddle but if we start the analysis by looking to virtually any common dictionary definition of deployment we find the word used to mean “to use or employ something in an effective way.” Next we might look to SAE J1538 the “Glossary of Automotive Inflatable Restraint Systems” where we find this definition: “...3.78 DEPLOYMENT ... The activation of an inflatable restraint device.” Notably there is no use of the word “Event” as part of this definition. There is no definition in that document for a “non-deployment;” however, the definition for a deployment uses the phrase “inflatable restraint device” and it would seem, given other references found in J1538, that a seat belt (safety belt) pretensioner is not considered “inflatable.” That is, of course, consistent with the way we think of how a belt pretensioner normally works; it’s not something we think of as being “inflated.”
identifier “02011_RCM-AB10P”) includes this passage: “... Some RCM may also categorize Non airbag deployment event. This type is an event in which non airbag devices such as pretentioners, knee bolster etc… have deployed. Note that such event can be overwritten given a subsequent '”deployment” event. ...” (sic) We’re back to the idea that maybe pretensioner deployments aren’t, well, “Deployments...” In that context then, we might next consider 49CFR563. There we see (among others) these definitions which include the word “deployment:” “...Deployment time ... means ... the elapsed time from crash time zero to the deployment command, or for multi-staged air bag systems, the deployment command for the first stage. ...” “... Time to deploy, pretensioner means the elapsed time from crash time zero to the deployment command for the safety belt pretensioner (for both driver and right front passenger). ...”
In the Bosch CDR Tool Help file glossary we find these definitions:
“... Event means a crash or other physical occurrence that causes the trigger threshold to be met or exceeded, or any non-reversible deployable restraint to be deployed, whichever occurs first. ...”
“... Deployment (Event) ... Acceleration observed along one of the car’s axes sufficient to cause the control module’s crash sensing algorithm to “enable” or “wake up” and which is sufficient to warrant a commanded deployment. ...” “... Non Deployment (Event) ... Acceleration observed along one of the car’s axes sufficient to cause the module’s crash sensing algorithm to “enable” or “wake up” but which does not warrant a commanded deployment. ...” If we look next at CDR Tool retrieved crash data for a historical reference we find, for example, in Ford data as far back as model year 2001 that an event where there was an airbag commanded for deployment with or without belt pretensioners or where there was a belt pretensioner but no airbag commanded both types of events were referred to as “Deployment Events.” Ford reports change; however, corresponding to about the time of the implementation of 49CFR563, adopting a naming convention where a recorded event without an airbag deployment - but where only the belt pretensioners were “deployed” - would be identified as a “Non-Deployment Event.” For example, a report from a 2013 Ford Escape (Data Limitations text
“... Time zero means whichever of the following occurs first: ... (3) Deployment of a non-reversible deployable restraint. ...” All of these from 49CFR563.5 would seem to give us clues but no real specific definition of what a “deployment” or “non-deployment” is (or isn’t). The first associates the word “deployment” with a function of an airbag system but makes no reference to belt pretensioner. The second definition, like the first regarding a time from “time zero,” includes the word deployment and references belt pretensioners. The last two bring a new term - “non-reversible deployable restraint” into play but provide no direct insight into what that means (or doesn’t mean). We find at least what might be considered an interpretation of this concept in Chrysler report Data Limitations where we see these passages:
www.collisionpublishing.com
Volume 11 Issue 2.indd 111
“... Activation of only the (Active Head Restraints), if stored, will be a nondeployment event. ...”
Collision Magazine - Volume 11 Issue 2 111
1/23/2017 2:02:15 PM