I don't know anything about your system, but that won't stop me from making a suggestion. The simplest way to deal with the problem is to make sure that clotting is the only uncontrolled factor. Temperature and flow rate can be held constant by mechanical means. I'm not sure about viscosity, but it can at least be measured with a viscometer to see if it changes. You could also see what the effect is of varying viscosity independently by adding different amounts of a viscogen. This would allow you to apply a correction factor for viscosity if it does vary during the experiment, thus leaving only clotting as a variable.

Thank you very much for the suggestion. Each factor (temp, viscocity and flow) is different within certain ranges. What I want to do is to calculate the TMP from these factors so that when the TMP is increasing, I know that it is because of clotting. So I guess I should look at each factor independently whith respect to TMP and combine this into one formula.

I don't think a factor by factor study will be successful since you risk to have interactions between each factors that could account for the final TMP you analyse. So what I sugget you is to use a full design of experiment DOE (2^n) in which you study all the noise factors (n) you assume to have an impact on TMP. Are you familiar with DOE? How many factors do you want to study? If you have 3 factors, you have to run 8 experiments to evaluate the impact of all factors (Temperature, flow and viscosity) on TMP. And in this case, you will have a perfect analysis of your system, including potential interactions between factors.

Thank you! No i'm not familiar with DOE unfortunately, but i'll look into it!

I have only three factors: flow, viscosity and temperature from which I and to evaluate the impact on TMP. (like you said)

Then it is quite straightforward. You have to design a 2^3 full factorial experiment. If you need some further advices, do not hesitate.

Good luck!