From road profile PSD to road profile history

[pietro82]

Hi all,

I need to perform a simulation of a vehicle modelled as lumped masses. So I have in mind to excite the model with a road elavation and to use the ISO 8608, that it lists road profiles of different kind of roads. how could I get the road profile history?

thanks

Pietro

 

[GregLocock]

1) measure it

2) synthesise a profile history that gives the same psd as the one you want. Obviously there are inifinitely many histories that would satisfy this, and some will cause different responses to others. Even if you have an amplitude probabilty distribution for the road then you still can have substantially different profile histories, and I'm guessing you don't have that.



Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[pietro82]

Hi Greg,

thanks for your reply. Of course that it's not possible to derive just one time history from a PSD. I tried to assume that the phase is null and to apply the inverse fourier transform, but I get a very odd time history, so I was wondering if there is another way to get the signal.

Thanks

Pietro

 

[IRstuff]

"I tried to assume that the phase is null"

That's precisely what Greg was referring to.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[pietro82]

Hi IRstuff,
thanks for your reply, the problem is the signal generated is not correct, there is a huge spike, so somehow I should assume a phase for the signal, but I don't know how to do that. Do you have an idea?

 

[IRstuff]

Of course there's a huge spike; "I tried to assume that the phase is null"

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[GregLocock]

what IRstuff is getting at is that your IFFT will start with every sine wave in phase, leading to the mother of all spikes.

So, assign a random phase to each frequency bin and IFFT the resulting complex spectrum. This will give you a set of random numbers with a gaussian amplitude distribution function (that is a good thing) but you need to check that the crest factor (peak/RMS) suits your road type. I do not have a good source for values there, experience says 3-5 for uniform sections of road without potholes. If you end up with an unreasonable crest factor, try a different set of random phases.

Here's Tom Irvine's script for the same job


Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[pietro82]

Hi Greg,

thanks a lot for your reply. I take a look to Irvine's matlab script.

Cheers

Pietro