MULTILEVEL APPROACH TO NUCLEAR FUEL BURNUP OPTIMIZATION

The aim of the paper is to show how the complex, overall burnup optimization problem, subject to the constraint of the desired power distribution, can be solved by decomposition into less complex coordinated subproblems. The solution has been obtained by use of the multilevel approach. The advantage of this approach is that it makes the computer solution of the problem of optimization practical. Two decomposition structures are considered, one for discrete and one for continuous reactor refueling. In the second case the optimization problem is subject to constraint in a form of an inequality containing a differentiable operator. To solve this problem the generalized Kuhn-Tucker theorem is used. To determine the optimum control of the desired power distribution, the Kulikowski approach is applied. As a result, the cyclic optimization process for both structures is obtained in which the information is exchanged between suitable level controllers.