|
Pages: [1]
|
 |
|
Author
|
Topic: Curve fitting air fans...or other devices? (Read 6956 times)
|
Vernier Forum Admin
Administrator
Full Member
Posts: 108
|
Posted on: 2006-06-06 05:01:02 by BriK
I have a need to develope calculations based upon tested data so that I can programatically predict the performance of a speicif device based upon manufacturer and/or tested data.
I am not a math wiz, so bear with me
As an example, I have a chart/graph containing tested data of a fan/blower. The chart has:
1) cubic feet per minute discharge (CFM) on the X-axis
2) inches of water static pressure ("H2O) on the Y-axis
3) and curves for 900, 1000, 1100, 1200, etc, up to 2400 revolutions per minute (RPMs) in 100 RPM increments on the graph
Each non-linera RPM curve defines specific RPM points for various CFM (x-axis) and "H2O (y-axis) values. The graph for a single fan/blower would have many RPM curves.
I assume I could use Grahical Analysis software to develope equations for each graphed RPM curve. Correct?
However, how would I find a CFM and "H2O point that would be between the graphed RPM curves? How could I develope a solution to determine the RPM vaules between the tested curves? Would I find the closest known RPM curve value, then find the next highest known curve RPM value, and then use ratios between the two known RPM curves to try and determine the unkown RPM curve value?
Or is there a better approach?
What is a good approach?
Is this not the tool to use?
Thanks for your time
|
|
|
|
|
Logged
|
|
|
|
Vernier Forum Admin
Administrator
Full Member
Posts: 108
|
Posted on: 2006-06-17 17:33:24 by Herman Trivilino
Wow! For someone who's not a "math wiz" you've got some pretty specific requests, there. Unfortunately, without seeing the data you've got, I don't know how to answer any of them. Why are you interested in this analysis of these graphs? What do you plan to do once you get the answers to your questions?
|
|
|
|
|
Logged
|
|
|
|
Vernier Forum Admin
Administrator
Full Member
Posts: 108
|
Posted on: 2006-06-20 05:53:24 by BriK
Here you can find an example from a fan manufacturer:
http://www.continentalfan.com/pdfs/tmkimpeller.pdf
On the second page is a graph similar to that which I described. The curves shown are a fixed RPM such that if you, for example, change the static pressure on the y-axis you can see how the flow (CFM) is effected on the x-axis...this is actually not a good example as the charts here only show one RPM curve for a perticular fan size, while the manufacturers I will be using provide charts with many RPM curves. Just to complicate things a bit, not all manufacturers catalog thier products the same way, annoying, but true, some will show constant flows (CFM) and varied RPMs and static pressure in tables rather than charts. But it amounts to the same things...
As to to your other questions...sorry to ramble here a bit...
My intent, as a development engineer tasked to develope some standard heating and cooling equipment, would ultimately be to develope formulas and tools (charts or programs) for the sales/marketing folks to quickly zero in on a standard unit size once my project is complete.
Many manufacturers of the components I will need to use in my project provide charts defining thier products operation under certain conditions. Other components can have charts developed through testing or simulation programs. Most of these charts contain multiple sets of curves generated/plotted on x/y axis. For example, fan charts will have many non-linear RPM curves plotted on an x/y axis. So I am thinking that it should be farily simple to take each of these RPM curves and develop a regression formula for each, and then with some known values programatically determine which fans best fit the application, and when/if a required point falls between these known RPM curves I could programatically select the two closest curves for a perticular fan and then use some form of extrapalation to determine the x/y values of a point between the two curves.
I have done similar things in my past, but back then I used linear regression DOS programs so I always had to keep my ranges within some tight contraints in order to minimize the error of using linear regression on non-linear curves. I now find myself needing to do this again, but many more tools seem to be available, and for more than just linear regression. Graphical Analysis seems to be a nice tool which I think could help me, and is nicely priced. Hence my curiosity.
Does that help?
|
|
|
|
|
Logged
|
|
|
|
Vernier Forum Admin
Administrator
Full Member
Posts: 108
|
Posted on: 2006-11-01 09:48:30 by Scott
Question: "How would I find a CFM and "H2O point that would be between the graphed RPM curves? How could I develope a solution to determine the RPM vaules between the tested curves? Would I find the closest known RPM curve value, then find the next highest known curve RPM value, and then use ratios between the two known RPM curves to try and determine the unkown RPM curve value?"
Answer: You might use the Affinity Laws or Fan Laws, which apply to Fans / Blowers and Pumps as well.
Flow (CFM) varies as RPM Varies 1:1
Pressure varies by the Square of the change in Speed or Flow:
(New RPM / Old RPM) ^2 * Old Pressure = New Pressure.
Brake Horsepower varies by the Cube of the change in Speed or Flow:
(New RPM / Old RPM)^3 * Old BHP = New BHP.
Fortunately with Fans / Blowers and Pumps it is easy if you do not change the
ductwork or piping. The 'system' duct / pipe follow these same laws.
If you are moderately familiar with Excel, you can easily Graph your own Flow /Pressure and BHP curves any any given RPM based on the original manufacturers curve test data. If the density changes, due to temperature, elevation or gas density, density corrections have to be made.
Regards,
Scott
FanAir Co.
|
|
|
|
|
Logged
|
|
|
|
|
Pages: [1]
|
|
|
|
|
|
