Impact Attenuator Report

[Matt K]

Hey Everyone,

Due to the new rule changes, this is the first year our team has done any dynamic testing with our impact attenuator. We ran our tests, collected the data given by an accelerometer, and since we observed spikes over 40g in our data, we used Matlab to generate a 3rd order butterworth lowpass filter with a 100Hz cutoff frequency, and then used the filter command to filter our data. My question is, have any other teams used this method or a similar method in the past to filter their data, and if so, have you had any difficulties resulting from this method, or is there anything that we need to be aware of that could cause problems with the report review committee? Thanks in advance!

Matt

[Matt K]

Hey Everyone,

Due to the new rule changes, this is the first year our team has done any dynamic testing with our impact attenuator. We ran our tests, collected the data given by an accelerometer, and since we observed spikes over 40g in our data, we used Matlab to generate a 3rd order butterworth lowpass filter with a 100Hz cutoff frequency, and then used the filter command to filter our data. My question is, have any other teams used this method or a similar method in the past to filter their data, and if so, have you had any difficulties resulting from this method, or is there anything that we need to be aware of that could cause problems with the report review committee? Thanks in advance!

Matt

[Crispy]

Thats how we did it. No problems, got a decent "grade" at comp. Depending on the type of accelerometer, you may have acceleration drift issues that you may or may not want to deal with.

I'm really not sure how strict they are with these. Other than a bad grade at comp, does anyone know of negative consequences of a crummy impact attenuator report?

[moose]

I'm pretty sure they got rid of the "grade" this year for just a Pass/Fail type of thing.

Only two places I could really see it hurting you:
1. If you crash and need it, and your attenuator is actually a POS
2. If you get into a heated discussion over it in design for some reason.

Neither is likely. But both are possible.

[Simon Dingle]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Matt K:
Due to the new rule changes </div></BLOCKQUOTE>

Can you point me to where this rule change is please? At Germany 2009 the organisers weren't too happy with our slow-crush test method and said that 2009 would be the last year that they would let us pass without a full-scale test. However I've compared the rules from 2009 to 2010 (Section 3.21) and its exactly the same, is there somewhere else that this is written?

I have been working towards organising for full-scale testing but it's becoming a real hassle and taking a lot of my time that I would rather spend elsewhere. Now I'm wondering if it is actually necessary.

EDIT: Sorry, I didn't actually comment on your question! We made a hardware Butterworth filter, but as long as the organisers don't think you've faked your filtered data then I don't see a problem with using a Matlab filter. Also, if they aren't happy with it they'll let you know and then you have 10 days to provide the information they require. Probably a question for the rules commity if you are really concerned.

[Adambomb]

To my knowledge the Matlab filters are very accurate, and my impression (based on experience working with test engineers at Mercury Marine) is that software filters are becoming the standard. At any rate though, would be a nice to see what the rules committee thinks.

@Simon-We're in the same boat. Last year it really did sound like they were pushing hard for full scale impact testing, but again, I haven't seen anything in the rules to suggest it being mandatory. We just submitted our report with slow crush data...if it doesn't get approved, then we'll worry about killing ourselves getting dynamic testing done!

[Matt K]

It's not explicitly stated in the rules that you have to do dynamic testing, however, from last year, they added the restriction that there can be no peaks over 40g in the acceleration data. Previously, it would have been easy to do a slow crush (in fact, our team did that last year), since the only thing mandated was average acceleration. Since the new max peak rule though, I can't see how one could get the necessary data in any way except for dynamic testing.

Matt

[Adambomb]

Not necessarily. The rule in question (B3.21.2) merely states When using acceleration data... Never states that acceleration data must be used.

And that rule only applies to how acceleration data can be filtered. The 40g peak rule (also stated in B3.21.1) has been around since at least last year.

[moose]

here ya go folks.

http://www.sae.org/servlets/pr...ease&RELEASE_ID=1143 (http://www.sae.org/servlets/pressRoom?OBJECT_TYPE=PressReleases&PAGE=showRelease&RELEASE_ID=1143)

[moose]

I just posted a link. Should show up soon. But on the subject of dynamic testing or no, I agree that everything I have heard over the past few years since the rule was first introduced has pushed towards physical testing.
And looking at the rules now, it says:
"The team must submit test data to show that their Impact Attenuator, when mounted on the front of a
vehicle with a total mass of 300 kgs (661 lbs) and run into a solid, non-yielding impact barrier with a
velocity of impact of 7.0 metres/second (23.0 ft/sec), would give an average deceleration of the
vehicle not to exceed 20 g’s, with a peak deceleration less than or equal to 40 g’s."

To me, that means you need test data. But it doesn't necessarily need to be acceleration data (e.g. could be done on a force tester).

Back in '05 we had a semi-legit one, compared to other teams, and it consisted of some sheet aluminum riveted together and filled with expandable foam. Took all of an afternoon for a couple of us to make. I'd say these things have come a long way since the random pieces of crap teams used to stick on the fronts of their car.

[Schev]

Our report the past few years has utilized a quasi-static test with an instron machine. We had all our raw data from a few material tests, a few failures, then obviously the raw data for our solution material, which satisfied the given requirements. We were given a poor grade

Anyone who was at competition last year and attended the IA meeting the was held by Mr. Royce made it sound like they would only be accepting dynamic tests, and anything less would be a failure. Knowing full well it was difficult and maybe even dangerous, he urged us to get creative and that our previous tests that didn't produce dynamic effects would fail.

That being said, I expected to see the rules for 2010 incorporate a lot more pertaining to a IA testing although very few changes were made. Mr. Royce has since left SAE. So maybe he was the one pushing strongly for the required dynamic test, or maybe a $2000 registration fee was forcing them to limit how many people they could turn away, who knows

Although last year at competition I was strongly under the impression that our quasi-static testing on an instron machine would no longer be acceptable and would be deemed a failure. So under that impression we went with a full scale test this year

[moose]

That sounds fair - I just monitor whats going on a bit from afar, considering I haven't participated as a student for a couple of years.

However, if they were to require "full-scale dynamic testing" then I think the rules would have to say that specifically.

[Mike Cook]

With regards to physical testing:

I have gone through quite a bit of effort to physically test our crush zone and i'm not sure the rules committee has thought this all the way out. In my opinion, the chances of someone getting hurt from the testing itself, is much more likely than the the benefits one learned from the physical testing helping prevent someone getting hurt in a crash.

I work on a daily basis with big heavy trucks (anywhere from 17k to 60k) where we rig and lift with cranes. Most fsae shops are not even remotely equipped to do a true drop test safely. This is not something to get creative with. And I know a lot of you are going to say, 'cook, your stupid, we got this figured out, its not that tough' but there are a lot of fsae teams out there, and I know some of the retarded stuff I did when I was younger.

What I would like to see, if FSAE really cares about this crush zone, is make a spec one that you have to buy. Therefore there is no question about its structural integrity. I know this is an engineering competition, but safety is safety and there are plenty of other places to spend our time.

Mike

[moose]

100% agree. I've seen plenty of reckless/stupid things done in a shop before, and just plenty of people who don't know better trying to do something they shouldn't - be in a machine shop, or just working around a formula car. Some team going and dropping weights on some home-made rig, inevitably at 4AM, is asking for trouble to happen eventually

[js10coastr]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Mike Cook:
With regards to physical testing:

I have gone through quite a bit of effort to physically test our crush zone and i'm not sure the rules committee has thought this all the way out. In my opinion, the chances of someone getting hurt from the testing itself, is much more likely than the the benefits one learned from the physical testing helping prevent someone getting hurt in a crash.

I work on a daily basis with big heavy trucks (anywhere from 17k to 60k) where we rig and lift with cranes. Most fsae shops are not even remotely equipped to do a true drop test safely. This is not something to get creative with. And I know a lot of you are going to say, 'cook, your stupid, we got this figured out, its not that tough' but there are a lot of fsae teams out there, and I know some of the retarded stuff I did when I was younger.

What I would like to see, if FSAE really cares about this crush zone, is make a spec one that you have to buy. Therefore there is no question about its structural integrity. I know this is an engineering competition, but safety is safety and there are plenty of other places to spend our time.

Mike </div></BLOCKQUOTE>

Looking for the "like" button. Perhaps a standard (honeycomb aluminium) crush structure that doesn't require testing, but if you wanted to go the composite route you'd have to do full scale testing with a sled or ram a weight into a stationary portion; all this being recorded.

[moose]

Yeah; They could provide some data maybe on some materials like honeycomb, showing the effect of speed on crush strength, etc. From there teams could go and do either slow-speed instron type testing, or some calcs.

If a team was to go a non-standard geometry, then maybe some guidelines on the testing just for safety wouldn't be a bad idea.

[on energy drinks]

+1, safety first
http://i261.photobucket.com/albums/ii41/badeffectsengineering/IMG_7852.jpg

[Donna]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by moose:

However, if they were to require "full-scale dynamic testing" then I think the rules would have to say that specifically. </div></BLOCKQUOTE>

I wish they would say that specifically, being ambiguous and left open to interpretation is a complete pain. We did scaled quasi-static and impact testing for our crash structure last year, and only just got it passed for FSG after a phone call to the person checking our Impact Attenuator report and arguing the point that the rules don't call for full-scale testing. However, they made it pretty clear they'd only be accepting full scale testing for this year.

I think I agree about having a spec crash structure available to teams who don't have the facilities, or budget, for full scale testing. The quote we got from MIRA last year was £900 - and that was at half price. I did half-jokingly suggest getting one of our drivers to drive at the right speed into a wall, but I don't think our uni would approve...

[Howy76]

Donna,

Another place in the UK to conduct crash testing is at Millbrook. We used their facilities 2 years ago for almost half of what you were quoted by Mira and I have to say I thought it was money well spent.

We paid for half a days use and tested over half a dozen crash structures. Whilst the outlay was realtively high it was an investment and we continue to utilise the data now.

Kind regards,

Howard

[Michael Royce]

Schev,
You must have misheard. I have never said that only dynamic testing would be accepted. For the very simple reason that not all teams have access to a dynamic test facility, the Rules Committee all agreed from the very start that quasi-static test (crush) on something like an Instron is completely acceptable, as long as it is accompanied by the right calculations to prove the average and maximum decels meet the requirements. The Committee also took advise from people in the industry and said that testing a small sample was NOT acceptable due to scale-up problems. All this is reflected in the FAQ's which are still appropriate.

The only thing that has changed since I left the Rules Committee is that the IA Data Reports for the Michigan and California competitions will not be graded, A, B, C, etc. They will just be "Approved" or "Not Approved", although SAE has not indicated in their posting as to what evaluation criteria will be used.

However, I understand that at Formula Student UK, the IA Reports WILL be graded, and poor reports will have points deducted. What will happen in Germany, Italy, Brazil and Australia, I do not know. Teams need to read the event Supplementary Regulations.

A reminder - teams entered in the Michigan and California events need to have their reports in by 11.59 pm EST tomorrow, Monday, March 1st, or they get penalized 10 points per day.

[Bazanaius]

I couldn't agree more about the scaling up argument - some tests we did in an instron comparing different thicknesses and even x-section areas of the same material showed that it was very hard to scale the results (we ended up with an overly large IA for that very reason, but couldn't afford to go back and re-test a smaller one due to the price of honeycomb).

However, I think the arguments about strain rate on aluminium honeycomb are not entirely well founded. A book by Michael Ashby (I believe it's called 'engineering materials') shows data and a rule of thumb (I shall look it up when I get home) that certainly within the strain rates we're looking at the honeycomb will perform almost identically to that in an instron.

I can only second the argument that full scale testing is dangerous unless properly organised, supervised and measured. A £900 test at MIRA last year would have been ~10% of our budget for the year, and so I can say 100% that we would have gone with the 'push a pallet full of concrete off a bridge' technique. We would also probably have a hilarious story of how someone 'almost' lost their leg in the process.

Given the quite token nature of the IA's on these cars (I wouldn't take one on a hillclimb if you paid me), I think the idea of a stock spec. IA (with the option to do something different for bonus design points) which teams can just build to get through scrute is a good idea. It would probably also lead to large discounts from material suppliers on that particular part and bring down the cost for smaller teams.

I guess it depends on whether the IA is there for safety, or for design points.

[D Collins Jr]

I think its funny that everyone has forgotten that this is about the only application forum where there aren't really any spec'd parts. In the world. We're getting away from it a little, with the discussion of a standard fuel filler. I would vote that if we make the IA a spec'd part for purchase we've eliminated some of the core elements of what FSAE is supposed to teach (specifically testing and report writing). I agree that there are other areas to do this in, but without creating those opportunities, we can't get rid of the IA design process. FSAE is ONLY feasible due to its educational aspect. If we take too much of that away and make into too much of a fabrication contest, we won't get to keep playing racecar at school.