Parameter dependency in CADET Match

CADET General Information
1

Hi everyone,
i have a question regarding the implementation of fixed parameter dependencies in CADET Match. Is it possible, to e.g. determine k_ads and k_des when k_eq is known? That would mean, that when a value for k_ads is tested, the corresponding value for k_des to be used in this simulation would be calculated as k_des=k_ads/k_eq and NOT be a further randomly varied value.

Best regards,
Thomas

2

What you can do is use the auto_keq transform and set your upper and lower bound on ka and then set the upper and lower bound of keq to the same value and it should work. There is a chance that setting the upper and lower bound to the same won’t work since I have not tested that specifically but if it does not you can set it to a VERY narrow range.

For example the code below should work if you don’t know ka but you think keq is 0.1. If you are not entirely sure on the exact value of keq you can actually just put in your confidence interval on the value of keq to give the system some wiggle room if needed.

{
			"transform": "auto_keq",
			"component": 0,
			"bound": 0,
			"location": [
				"/input/model/unit_001/adsorption/LIN_KA",
				"/input/model/unit_001/adsorption/LIN_KD"
			],
			"minKA": 1e-8,
			"maxKA": 1e8,
			"minKEQ": 0.099,
			"maxKEQ": 0.101
		}
3

Thank you very much! That seems like a very suitable approach. I have an additional question regarding the auto_keq transformation: If I give the parameter ka into CADET Match with this transformation and set the corresponding limits (for ka and keq), then i do not give the parameter kd as parameter into the algorithm, right? Instead, I just recieve final values for ka (and keq) and then can calculate the corresponding kd. That would also mean, if I do not know my keq beforehand, then I will recieve two fitted parameters (ka and keq), even though I just enter one fitting parameter (ka, but with auto_keq transformation). Is that correct?

4

No problem

As you can see in the transform it has bounds for both ka and keq. If you don’t know either parameter you can give them both wide ranges and it will fit both parameters. If you know one of them very well you can give that parameter very tight ranges and it will mostly just estimate the other one. In the reports that CADETMatch generates it will tell you ka, kd and keq so you don’t have to calculate it but it only actually estimate ka and keq.

Also if you look at location you can see you have to specify where ka and kd are located. The auto_keq transform is just a single transform that works on two variables.