Harmonic Coefficient Calculation Software

[nerd189]

I have been working through Fourier Analysis for customer supplied pressure vs. crank angle data for a few viscous TVD design projects and it can be quite a cumbersome process. My question to you guys is are there programs out there that will take input cylinder pressure vs. crank angle data and compute harmonic coefficients? We use the harmonic coefficients for our torsional analysis program input sheet. What programs would you suggest?

Our company is looking to save some engineering time and still be able to obtain accurate input data.

We are also looking for an updated crankshaft torsional vibration simulation software, if you have any suggestions. We work with gasoline and diesel applications, both 2 and 4 stroke.

Thanks in advance!

 

[ivymike]

I wrote a DFT routine (macro) in excel for that purpose.

You might see if there is software shown on Ricardo.com that meets your needs - Valdyn is probably their right tool for crank TVs these days. We use in-house code with outsourced labor at my present employer.

 

[SomptingGuy]

If you intend to do a lot of investigative time-series analysis, you may find something like Matlab more useful than the ubiquitous Excel. Or if you don't feel like paying for it, Octave is almost as good for most things.

Both of these have built in functions for Discrete Fourier Transforms (I think both use FFTW these days). There's not much of an entry barrier, once you've got your head around the basics.

Of course you'll need to experiment with their FFT functions until you are confident that you understand what they are actually giving you. It takes time, but it's (in my sad, nerdy view) really exciting when you finally "get it".

- Steve

 

[nerd189]

Thank you guys, I do understand how it is when you "get it."

We have MathCad here so matlab is out of possibility but I could look into Octave.

So there are no specific programs available to do this process without writing one? I guess that should have been my question from the start.

 

[SomptingGuy]

A specific program?

It depends how competent you are as a programmer. Most (if not all) of our programs use the FFT algorithm described in Numerical Recipes. The code itself is simple. Then again, a brute force DFT is also pretty simple. Getting data in and out is the major issue and only you know how you'd like the I/O to work.


- Steve

 

[GregLocock]

It'd be easy to do in Mathcad, although not as a production piece of code.

20 years ago i could have pointed you at hardware that would do what you ask, but I'm not up on software.

We'd call it a gated analysis, using synchronous sampling, but Steve may have a better terminology.

Cheers

Greg Locock


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

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

 

[nerd189]

Thanks again for the great responses.

Looks like I have enough to convince our Engineering Manager that it will be worth it to take the time to write a program. Now I just need to brush up on writing them.


Thank You