Improving performance for a 13C isotope separation plant using multivariable fractional order controllers

High purity ¹³C isotope is widely used in numerous areas of medicine, including tests to evaluate metabolism problems, ulcer diagnostics, studies of amino-acids kinetics and especially to highlight metabolic changes in the human brain that help diagnose earlier important diseases such as Alzheimer and other neuropsychiatric abnormalities. It is for this particular reason that efficient methods for separating and enriching the ¹³C isotope are widely sought. One of these methods is the cryogenic distillation of the two stable carbon isotopes, ¹²C and ¹³C. However, due to the low separation factor (1.007 at -194°C) a single distillation column is not enough to raise the natural concentration of ¹³C from its original 1.1% to 90% Therefore, cascaded columns are generally required. Nevertheless, such complex equipment involves the individual control of each column and also a supervisory control of the entire cascade. Prior to the design of a supervisory control system, the individual control of each column has to be developed. Due to the modeling errors, the variable time delays, the multivariable nature, strong couplings and interactions and also because any deviation of the process outputs from the prescribed set-points might lead to compromising the entire separation process, a robust control algorithm should be employed. Fractional order controllers have proven to be an interesting approach towards solving the problem of closed loop robustness. It is therefore one of the choices for controlling the ¹³C isotope separation column. In this paper, the authors design and compare two multivariable controllers, a fractional order controller and its integer counterpart. The simulation results show that the multivariable fractional order controller outperforms the classical one, and represents a simple and robust choice for raising the concentration of ¹³C.