is it possible to calculate the lateral Delta-V from the acceleration rates?

This forum is for the open and public discussion of Accident Reconstruction topics. This is open to any registered user of the CrashForum.info site.
Post Reply
User avatar
FCI
Posts: 5
Joined: Mon Aug 19, 2019 6:38 am
First Name: Damien
Last Name: Farrell

is it possible to calculate the lateral Delta-V from the acceleration rates?

Post by FCI »

Hi, is it possible to calculate the lateral Delta-V from the acceleration rates?

I have a long delta v. Target vehicle hit on the near side as it was turning right.

Thanks
Attachments
612C1CCA-F52D-4FF7-ABA2-DD1432E5EBD4.jpeg

User avatar
mchenrysoftware
Posts: 76
Joined: Mon Mar 05, 2012 7:49 pm
First Name: brian
Last Name: mchenry

Re: is it possible to calculate the lateral Delta-V from the acceleration rates?

Post by mchenrysoftware »

Folks weren't responding to this so after 3 days...sorry to be posting so much...

Answer is YES...however...it depends...
  • On the logic/trigger for the recording of lateral acceleration since when it starts and ends may affect what total lateral DeltaV is calculated.
For full scale testing and simulation[*] acceleration is used to calculate the DeltaV total and in X & Y directions
  • See link below for background information, equations, and references
    • The reference cited also discusses issues we found with full scale test analysis procedures:
    • Testing facilities need to consider the possibility of a changing sign of the lateral acceleration: it can cause issues.
    • We recommended calculating the total deltaV and PDOF and then determine the lateral acceleration (see paper referenced in link)
    • Of course in a t-bone this shouldn't be an issue (all lateral acceleration is in one direction)
Here is the link: Brian

FOOTNOTE:
[*]Simulations that use acceleration to calculate DeltaV are ONLY those that calculate crush forces and moments DURING the collision and so have acceleration time histories available (like SMAC)
Simulations programs that assume "Instantaneous exchange of momentum" during collisions DO NOT calculate any crush forces as they occur so the time history of accelerations during simulated collisions are not calculated or available.
Brian McHenry, McHenrySoftware.com
NEW!! LEASE pricing of msmac3D Software!! 3D Crash Reconstruction and Simulation Software.

User avatar
FCI
Posts: 5
Joined: Mon Aug 19, 2019 6:38 am
First Name: Damien
Last Name: Farrell

Re: is it possible to calculate the lateral Delta-V from the acceleration rates?

Post by FCI »

Thats great, thanks Brian.

User avatar
Rusty Haight
Posts: 624
Joined: Tue May 10, 2011 3:24 pm
First Name: W. R. Rusty
Last Name: Haight
Contact:

Re: is it possible to calculate the lateral Delta-V from the acceleration rates?

Post by Rusty Haight »

FCI wrote: Fri May 08, 2020 5:18 am Hi, is it possible to calculate the lateral Delta-V from the acceleration rates?

I have a long delta v. Target vehicle hit on the near side as it was turning right.
I'm glad Brian caught this, I missed it. First, I'm going to copy this to the CDR User Forum, it'll be useful there as well.
Next, delta-V IS acceleration and time so, yes, you can calculate delta-V from the acceleration rate in the Toyota download; however...

The main thing to keep in mind here is: what are you working with? If this were, for example, a set of Chrysler lateral acceleration data like this:
Example lateral data from a pre-"563" Chrysler system
Example lateral data from a pre-"563" Chrysler system
What you need to look for is (a) resolution and (2) clipping. Brian mentioned the changing polarity: this is a good example of that. Resolution is going to be in the reported value for g's and in terms of time. Here, the data is reported in 1ms intervals at a resolution of about +/-.5g.

If you're working with this kind of data, you get to pick where you start and stop the calculation based on the wave shape; however, as Brian pointed out, with a trigger or "wake up" threshold in the range of 2g, the Toyota data you have would necessarily clip off some of the "front end" of the curve. Having said that, look at the wave shape above, 2g, in terms of time, is about 2 ms into the crash pulse and would make less than 0.01mph difference in the ultimate calculated delta-V. The clipping (at the top, the "flat lines" at just over 30g) will have a greater impact on the calculated delta-V. In the end, the calculated delta-V from the data above is about 29mph whereas the instrumented delta-V (with a greater range for acceleration) comes in closer to 32mph.

In your Toyota data, I don't think, given the acceleration reported for the labeled "0" sample is about 0 m/s^2, the initial wakeup under reporting is a real concern. Having said that, let's talk about time. Notice the time is in 10ms intervals. As it turns out, the longitudinal delta-V in your data set is a cumulative average reported at each 10ms interval. The module "saw" the event over at about 80ms and then, per 49CFR563, reported the delta-V out another 30ms (then the grey area begins). Here, I'm not sure what you mean by "long delta-V," this is actually pretty short.

To use this Toyota data to calculate the lateral delta-V you could go with the acceleration times time (10 ms) and do the cumulative calculation (I'm not going to reenter the numbers, if I had the CDRx file it would be a quick export to a spreadsheet and we could also confirm polarity from the Data Limitations). That would give you a delta-V but I suspect it might be a little under reported given the averaging that goes on to get to the numbers you DO see in the table. Still, as Brian pointed out, you could use that at the corresponding longitudinal data point (I'd use 80ms here given that's where the longitudinal max occurs but I'd want to see what's in the calculated lateral component first) and find PDOF and, if that works, and compares with other calculations you've done, I'd feel pretty good about the result.
- Rusty Haight
Collision Safety Institute

Post Reply