depending on what kind of problem is that, do you need to classify something or you need to do regression? seems like the problem is non linear, try to apply logistic regression algorithms. first create a math analysis of your problem, then apply ML
The choice of an AI depends on the # of variables influencing the phenomenon. Major variables influencing EOR depend on the nature of the reservoir and the hydrocarbon.
In your case you can define quantitatively the variables ( e.g WAT , flow rate , viscosity , composition ,pressure , density etc of the crude oil) that affect wax deposition and inhibition. And then based on this data you can use as inputs for traning or fitting a model via regression or machine learning techniques to predict wax formation . ANN can work for this case
One of the approaches is to combine the AI based modeling techniques such as ANN, SVR, RBFNN with nature-inspired algorithms such as GA and PSO after formulating your optimization problem (which is related to flow assurance phenomenon) .
The best alternative is to combine the physics and chemistry formulations of the phenomena with the ML modeling techniques and the nature-inspired optimization approaches.