Time based data reconciliation and decision support for a CFB boiler

This paper covers a method for operator decision support, where physical simulation models are used to connect different physical variables to each other. By comparing energy and material balances for a larger process area inconsistencies in single process parts and sensor measurements can be detected, by following the development between single measurements and values predicted from the simulations. This information then can be used as input to e.g. a BN, Bayesian Network, for decision support. The application has been for a CFB boiler at Mälarenergi AB. The simulators have been made in Modelica respectively a more advanced model in APROS.