WetBulb Temp Calculation

Hello all,

I am trying to calculate Wetbulb temperature with the known values being Drybulb Temp. and RH%.

I know it is not a "simple" formula but it seems that it could be done with only these two variables.The reason I say this is the fact that you can look up wetbulb on a pychrometric chart using only these two variables.
I have tried two different formula's but both seem to gain more error toward one end of my range or the other...

My range is 50-60° F to 165° F and 0% RH to 100% RH. Very broad - I know.

If anything, Could anyone point me in the direction of a site that might contain a calculation for this.I have searched the net with limited results.

I have also thought about implementing a sort of lookup table in the PLC and just entering a ton of data and coming up with some indirect addresses mess to look the values up.

Also, just as a side note- Don't worry about PLC, I just need the raw formula, I can get it into the PLC.

Have you tried plotting the points on a chart in Excel then fit a curve to it to obtain your formula?

I found THIS... Note the Excel formula download.



-Eric

Just note that on that site it says the range is 0 to 40F, and that accuracy diminishes after that. But maybe you could tweak it (didn't download myself, just wanted to point it out...).

Psychrometric Equations

I found the attached some time ago. This should have everything you need.

BurnerMan: That plotting the curve on a chart is already available on a psychrometric chart. The difficult part is figuring out the formaula.
Wanderer: Your right, once you start getting up in the range I need to be in, that 0-40F formula goes way out of range.
Tom: I had taken a look at that a few times in the past but all the formulas listed assume that I know at least 5-8 of the other variables, I really only have DryBulb and RH% to work with.
My current formula works pretty well from 165 F down to 135 F. I have set up an excel table to see the relationship of the (error of calculation) as the Temp goes Down as well as the range of RH% values. I think I should be able to set up a lookup table in PLC memory that stores both variables in 10 degree increments and 10 RH% increments. Then depending on my Temp and RH%, I could lookup the percent error and use some SCP functions to correct the WetBulb Result. That should be doable, right... I'll keep yall posted.. Thanks for the responses.

Strain your brain. More math of the advanced kind.

PLC Gremlin, it posible to find an equation that will do what you want as long a you have data for the range you are interested in. Basically we would find the default values for those parameters you don't know and assume they are constant. Is this a good assumption in your application?

Can you post a .csv file of the data? It should be arranged like this:

0 Dry Bulb Temp, RH%, and Wet Bulb Temp
1 Dry Bulb Temp, RH%, and Wet Bulb Temp
2 Dry Bulb Temp, RH%, and Wet Bulb Temp
3 Dry Bulb Temp, RH%, and Wet Bulb Temp
4 Dry Bulb Temp, RH%, and Wet Bulb Temp
...


We(I) can find an equation from this.

Now where was this dicussed before? I bet Tom knows the thread and where I am going with this. Let see if someone else remembers or can find the thread through all the NOISE on this forum. I do this sort of think all the time to find models for motion systems.

A formula would be much faster and smaller than a look up table. This is good stuff to know. I don't see how any engineer can get by without it.

Is this the thread you were thinking of:

http://www.plctalk.net/qanda/showthread.php?t=16474&page=2&highlight=linest

See page 2

Peter,

Attached is an .XLS and a .CSV of Data.
Notice that I am lacking some Higher RH% values at the Higher Temps. Can not find a Psychrometric Chart to Fit these To. Even without these, it looks as though there is a relationship as the Temps go up that could still fit my highest temps (165F)..
In the mean time, I will take a look at the other posts in the above mentioned thread.. Thanks.
Also Peter, If I went with lookup table route, Speed and space is not really an issue in this app. And yes constants for other variables should be fine.

I know you know. I was just wondering if anybody else remembered.

Tom Jenkins said:

Is this the thread you were thinking of:

http://www.plctalk.net/qanda/showthread.php?t=16474&page=2&highlight=linest

See page 2

Click to expand...

Yes, that was a good thread. Tom, have you played with Excel's solver? This is an add in and isn't installed by default. The solver seems to be much more flexible than the LINEST function. I haven't used it yet as I do most of my work on Mathcad. I think I still have my old LINEST.xls that we used to find the equation that went through the 3 points in the example in the link you just posted.

PLCGremlin, I will look at this data tonight.

As it turns out I am working a C code that does this sort of thing right now. I have a linear version that works with similar capabilities of the LINEST function, now I am working on a non-linear version with constraints. That way I can embed these capability into my products.

Peter Nachtwey said:

Tom, have you played with Excel's solver?

Click to expand...

I've tried it. Solver in Excell, as far as I can tell, is pretty lame. It only works with one cell at a time. It is NOT the same as the Backsolve function in the old Lotus 123. To accomp[lish that you need a macro - essentially a QuickBasic program or script.

This looks to easy. What am I doing wrong?

PLC Gremlin said:

Tom: I had taken a look at that a few times in the past but all the formulas listed assume that I know at least 5-8 of the other variables, I really only have DryBulb and RH% to work with.

Click to expand...

So where is the equation? I looked in the psychrometric.pdf and nothing jumped out and bit me.

All I see are two equations:
RH=Pv/Ps and Pswb-Pv=B1(Twb-T)
Multiply both sides of the RH equation by Ps
Pv=RH*Ps
If I substitute for RH*ps for Pv in the second equation I get:
Pwsb - RH*Ps = B1 * ( Twb - T )
Divide both sides by B1 which is assumped to be constant
Pwsb/B1 - RH*Ps/B1 = ( Twb - T )
Now add T to both sides
Twb = Pswb/B1 - (Ps/B1)*RH + T
If I simpify Pswb/B1 and Ps/B1 into constants then

Twb = A - B*RH + T

Now one just needs to find A and B!
Did I miss anything? That looks simple enough. It looks too easy, A and B are simple linear constants.

Here's another link to a variety of humidity/temperature formulae. The author has also indicated typical errors which may be expected because of approximations etc.

I suspect (I haven't checked) that there may be enough here to confirm/deny Peter's formula.

http://www.gorhamschaffler.com/humidity_formulas.htm

I've never heard of the company whose site this is, so I can't give any indication as to the validity of the information.

Regards

Ken

Peter,

I don't know if you looked at the data that I sent but it is definately not linear. I was epecting something closer to what I am currently using.

T(Kelvin) = (T-32)/1.8+273.15
DryStat = EXP(19.163-(4063.2+(184089/Tk))/Tk)
WetStat = DryStat*RH/100
Twb(Kelvin) = (4063.2 + (SQRT((-4063.2^2)-(4*(19.163-LN(C5))*(-184089)))))/(2*(19.163-LN(C5)))

Twb(F) = ((TwbK - 273.15)*9/5)+32

This equation gets me real close but produces alot of error below 15% RH at all temps and worse in the lower temps (below 110 F).
I was just thinking someone might have done this before. I beleive I will go the route of using my equation then just applying an error compensation based on a table of values in PLC.
Thanks

Yes, I looked at the data but didn't do much because I could see the data didn't fit. I need the basic form of the equation so only the constants need to be calculated.