Hao Xia

Eventuality synthesis for controlled linear automata

The synthesis of hybrid controllers that satisfy eventuality specifications is studied. Linear automata that accept continuous inputs are considered. The feedback control derivation is based on reachability analysis. A computational procedure for the reachable set, based on Fourier elimination and convexity analysis techniques, is derived. A feedback controller that guarantees the system will visit a given convex target set without violating state constraints is obtained.

Controller benchmarking based on economic benefits

Benchmarking is a tool for improvement and should be considered as a benefit-driven activity. The ultimate goal of benchmarking is performance improvement. One of major difficulties in the benchmarking of MIMO systems comes from the shear number of control loops contained in a plant. By directly focusing on the improvement of plants economic performance, we present a controller benchmarking framework based on the existing multi-level control hierarchy. At each level of the hierarchy, a constrained optimization which reflects the economic objective is performed. The active constraints recorded in the optimization process are treated as the goal of improvement in the lower levels. Consequently the constraints in each layer which limit the performance of the plant are isolated. The retuning decision is made based on the result of optimization of an economic metrics. In this manner, the benchmarking problem becomes not only manageable but also meaningful.

Controller benchmarking algorithms: some technical issues

(1) Strategy and Planning Coordinator (West Africa), BG Group, Reading, UK, (2) Professor of Automotive Engineering, Faculty of Engineering, Kingston University, London, UK, (3) Research Fellow, Department of Electronic and Electrical Engineering, University of Strathclyde, Glasgow, UK, (4) Siemens AG Power Generation, Muelheim an der Ruhr, Germany, (5) Siemens AG Power Generation, Muelheim an der Ruhr, Germany, (6) Manager, CESI SpA, B.U. Transmission & Distribution, Milan, Italy.

Controller benchmarking of a heat-exchange network

Benchmarking is a tool for improvement and should be considered as a benefit-driven activity. The ultimate goal of benchmarking is performance improvement. By directly focusing on the improvement of plants economic performance, we present a controller benchmarking framework based on the existing multi-level control hierarchy. At each level of the hierarchy, a constrained optimization which reflects the economic objective is performed. The active constraints recorded in the optimization process are treated as the goal of improvement in the lower levels. Consequently the constraints in each layer which limit the performance of the plant are isolated. The retuning decision is made based on the result of optimization of an economic metrics. The benchmarking practice for a heat-exchange network is used to illustrate the idea.