State-Space Modeling using Operating Regime Decomposition and Local Models

摘要: Abstract Building non-linear state-space models of complex processes using first principles is a resource demanding task. Here decomposition the problem by constructing simple local that are valid only within certain operating regimes proposed. A smaller number interacting physical phenomena dominant each regime. Hence, tend to have considerable less complexity than global models. Using estimates validity models, given current point, it shown how may be interpolated model. The resulting model should not considered as model, since into and interpolation inherently empirical. Neither looked upon an empirical amounts process knowledge used both for regime choice We believe present approach contributes modeling in cases with limited step towards framework integrating modeling.