INGAR   05399
INSTITUTO DE DESARROLLO Y DISEÑO
Unidad Ejecutora - UE
congresos y reuniones científicas
Título:
Control system performance monitoring based on optimal action selection
Autor/es:
AVILA, LUIS OMAR; MARTINEZ, CARLOS ERNESTO
Lugar:
Londres
Reunión:
Simposio; 22nd European Symposium on Computer Aided Process Engineering - ESCAPE 22; 2012
Institución organizadora:
University College London, UK
Resumen:
Optimal action selection chooses the controller design that makes the closed-loop behavior as close as possible to the desired one. A fully probabilistic design framework is proposed to describe both the closed loop and its desired behavior in probabilistic terms and uses the Kullback?Leibler divergence as their proximity measure. For performance monitoring, the relative entropy is chosen due to its sensitivity to pattern frequencies of characteristic templates in a time series. The relative entropy is increased by any deviation from optimal operation in a process system which makes it ideal for supervising Real-time Optimization (RTO) and dynamic optimization schemes. Frequencies for eight basic patterns or templates observed in the stochastic process of the controlled variable are used to estimate the proposed performance index. Blood glucose control in the human body is used to exemplify the proposed methodology.