Mathematical modelling of SO2 flow control system to an isotopic separation column using artificial neuronal logic and fuzzy logic

Full text: After we have done experimental measurements in a 15N isotopic plant we have determined the chemical reaction temperature curve. Its profile is similar with a Gauss type one and has a maximum point on T axe and two quasi symmetrical branches on d axe, around the maximum point, namely: upside branch for d < 43 cm and downside branch for d >43. The middle of the glass refluxor where take places chemical reactions is d=43 cm ; here is the point of the maximum of curve temperature where the control system is properly tuned. At same time we have determined another two lateral temperature profiles which have the maximum points in d=28 cm, respectively in d=58 cm. The principle on which our modelling is based consists of the surface calculation which ranks between T curve and x axe. During the regulation process, the temperature profile goes up and down around d=43 cm central position. As the temperature curve goes up from the central position, the curve surface under the upside branch increase, and surface under down side branch decrease, and vice versa, when the temperature curve is going down. The two branch surfaces and the relationship between their changes can be a matter of principle of the neural mechanism of excitation and inhibition which is mathematical well assisted by neural artificial logic theory, by MATLAB special tools and by LabVIEW mathematical functions. After the neural logic algorithm decided the temperature curve movement direction and movement speed, outputs a set of data, that are used to be processed by a fuzzy logic algorithm. As response, the fuzzy logic outputs execution signals and duration-times to control SO2 flow. (author)