On the role of LFT model reduction methods in robust controller synthesis for a pressurized water reactor

We consider model reduction and control design for a pressurized water reactor (PWR) of a nuclear power plant. In order to obtain less conservative design and analysis results, a recent model reduction technique for uncertain systems modeled as linear fractional transformations (LFT) on repeated scalar uncertainty sets is used to determine the smallest dimension uncertainty description required to adequately cover the system behavior. Controllers for the PWR are then designed using μ-synthesis techniques. The change in both the operating power and reactivity of the fuel are modeled as slowly varying structured uncertainty, and the objective is to maintain performance over all parameter variations. A parameter-dependent model is constructed from identified models obtained using a nonlinear simulator. The models are simplified using a systematic model reduction method for uncertain systems; this proves to be critical in the design process of μ-controllers with large structured uncertainty, The resulting closed-loop performance of the PWR is compared to previous design results