| T O P I C R E V I E W |
| mluca001 |
Posted - 08/02/2018 : 05:50:54 AM
Origin Ver.: 8.1
Operating System: Windows 10
I am trying to set up a Fitting Function for the modified Arrhenius equation, but I can't get it to converge which leads to large errors, so I don't trust the values enough to actually use. I have no idea how to set up the parameter initialization and think it is part of the problem. The modified Arrhenius equation is:
k = A*T^n*exp(-E/(R*T)) which I simplified to y = A*x^n*exp(-C/x)
I did set up a function for the Arrhenius equation and it works. I copied the parameter initialization from Exp1P2. I don't have initial parameters when I used this function, but it did converge.
|
| 10 L A T E S T R E P L I E S (Newest First) |
| Purva Mathur |
Posted - 01/27/2023 : 3:41:24 PM
Hi mluca001, I am also trying to fit the modified arrhenius function but since I am very new to the origin I am having some difficulties in setting up the function. I would be grateful if you could share the function files. My email id is purva.mathur99@gmail.com. Thanks in advance :) |
| Shirley_GZ |
Posted - 08/17/2018 : 05:09:55 AM
Hi Mluca,
Our expert has looked into your opj, he found there are too little data points for this fitting, especially only two points in the rise segment. Then, the initial values of parameters are not good enough, and the parameters "A", "E" and "n" are interdependent.
You can estimate the good initial values for the parameters and then perform 1 iteration at each time till getting a ideal parameter value, then fix this parameter and do the subsequent iterations till converge, just like below:
Thanks,
Shirley
Originlab Technical Service Team |
| mluca001 |
Posted - 08/17/2018 : 04:17:17 AM
I tried with the latest version and I am still getting the same issues where it won't converge even after multiple maximum iterations. I will email the file plus the function you sent to the email you provide previously. |
| Shirley_GZ |
Posted - 08/17/2018 : 02:06:01 AM
Hi Mluca,
I am afraid you have to use this function in our latest version, you can download our trail to have a try,
https://www.originlab.com/demodownload.aspx
Thanks,
Shirley
Originlab Technical Service Team |
| mluca001 |
Posted - 08/17/2018 : 12:01:41 AM
Hi Shirley,
Thanks for the files.
That is the correct email, but I have not seen any emails including the Spam folder. What is the email address that it was sent from.
Do have I have to use the latest version for the fit or can it be done with 8.1? |
| Shirley_GZ |
Posted - 08/15/2018 : 03:34:44 AM
Hi Mluca,
I have already replied you with files via:mluca001@gmail.com.
Anyway, please download the file from our ftp, ftp://ftp.originlab.com/USERS/Support/ArrheniusMod.zip
About how to install the files, please follow the instaction below:
quote:
Please download our trail, https://www.originlab.com/demodownload.aspx , to install our latest version.
Then, drag-and-drop the FDF file to Origin workspace to install it to the category "Exponential" .
Copy and paste other three files to <User Files Folder>\fitfunc. You can open user files folder by selecting "Help: Open Folder: User Files Folder" menu.
Then, you can try to do the fitting with this function.
Thanks,
Shirley
Originlab Technical Service Team |
| mluca001 |
Posted - 08/14/2018 : 4:30:20 PM
Hi Shirley,
I've check my email and I have not seen anything. I also sent an email about it to tech@originlab.com and haven't received a reply. |
| Shirley_GZ |
Posted - 08/10/2018 : 06:53:12 AM
Hi Mluca,
We have built the function for you and sent the function files by mail.
Would you please check the registered mail box to have a try? Please feel free to contact us with any problem with it.
Thanks,
Shirley
OriginLab
Originlab Technical Service Team |
| mluca001 |
Posted - 08/08/2018 : 10:02:36 PM
I have tried this using my values for A and C from my ln(Arrhenius) fit, but the fit is still not converging. After entering my initial parameters and use "Fit until converged" and it went through the maximum iterations (400) and even do this 4 times (1600 iterations) and it still did not converge and the results have the errors larger than the values. |
| Shirley_GZ |
Posted - 08/03/2018 : 05:40:58 AM
Hi Mluca001,
I will suggest you first get the initial values of parameter "A" and "C" from the linear fit of ln("Arrhenius function"):
ln(y)=ln(A)-C/x
On this, you can perform linear fit on your data with the apparent fit (https://www.originlab.com/doc/Quick-Help/Apparent-Fit)checked, and then use the slope and intercept values to calculate the value of "A" and "C ", according to the formula above.
Finally, you can use the value of "A" and "C ", and n=0 to initialize the parameters of the modified Arrhenius equation to perform the fitting you desired.
If you still have problem on this, please send us the FDF file and project files with source data(via:tech@originlab.com) to let us do the fitting for you.
Thanks,
Shirley
Originlab Technical Service Team |
