Non-stationary separation processes of multicomponent isotope mixtures seem to be efficient in enriching components of small abundance. Using the separation of 36SF6 from its multicomponent isotopic mixture SF6 as an example, a numerical simulation method is applied to predict a nonstationary separation process occurring in a short gas centrifuge cascade, and also a corresponding experiment is carried out. The concentrations of the components in the SF6 isotope mixture are measured at different times during the separation process. The simulation results are in good agreement with the experimental ones. As an example of application of the simulation method to practice, a cascade is designed to separate the 36SF6 component to a concentration of over 99% from the isotopic mixture of natural abundance.

It is difficult to enrich isotope components of small abundance to a high concentration by means of gas centrifuge cascades. A nonsteady-state approach is studied for the separation of end components and middle components. Unlike ordinary conventional separation cascades, which have two withdrawals at the two ends of the cascades and one feed in between, the cascade used in the nonsteady-state approach has only one withdrawal at the either end and one feed. The nonsteady-state separation approach is analyzed from two aspects: the material recovery and the operation time efficiency, for producing one unit mass of material, and is compared with conventional cascades. The results show that the approach is advantageous over the conventional cascades for obtaining the same amount of a desired component, especially for the separation of end components.

The transient process is studied in gas centrifuge cascades for separation of multicomponent isotope mixtures. Two important practical factors, the holdups in connecting pipes among separation stages and material losses, are taken into account in the partial differential-difference equations that describe the concentration distribution of components. The equations are solved at each time step through the q-iteration method. The effects on transient processes are investigated in terms of the total holdup in cascades, the cascade cut, the separation factor, and material losses and may be very significant.

An implicit method is presented to study the transient process of concentration distribution of multicomponent mixtures in separation cascades for gases. The problems that the first-order method (the so-called "transient approach") encounters have been overcome. The new method is second-order accurate in time and unconditionally stable. The nonlinear difference equations at each time step are solved by the q-iteration method. A simple cascade is studied as an example, and comparisons are made with the first-order "transient approach." The results show that the new method is advantageous over the "transient approach" in both accuracy and computational efficiency.

Determining the concentration distribution of mixture components in a separation cascade plays an important role in analyzing and optimizing separation performance. A new iterative method, called the "q-iteration method," is presented to solve the nonlinear difference equations describing the concentration distribution of a multicomponent mixture in separation cascades at steady state. Iteration takes place over quantity q, the ratio of the head concentration to tail concentration of certain components, other than the concentration itself in commonly used methods. Numerical tests are carried out for both short and long square cascades. Comparisons are made with the transient approach. The results show that the new method is both robust and efficient.

Separating a certain middle component of isotopes from a multicomponent isotopic mixture is much more difficult than separating the two end isotope components. Numerical investigation revealed that by controlling the cut of a separation cascade, defined as the ratio of the product rate to the feed rate, it is always possible to separate a multicomponent mixture into two specified groups of components, a light group and a heavy group, in just one separation run. The cut is equal to the concentration sum in the feed of the components in the light group. The effects of the cascade length, shape, and feed location were studied. The results suggest that without using complicated cascades such as the M-cascade or the Q-cascade the separation of the middle component can be easily achieved.