A systematic analysis has been developed to investigate the performance of reduced-order models for multistage separation units in transient simulations. In particular, we have analyzed the accuracy of these models in reproducing some characteristic features of the dynamic behavior of the original full-order model, i.e. the eigenvalue of smallest modulus lambda(min) and the singular values obtained through open-loop singular value decomposition. It is found that as the relative volatility increases the maximum reduction factor allowed by a prescribed accuracy decreases. The present analysis shows that, with the above limitations on the value of the reduction factor, reduced-order models are able to capture the true dynamics of the corresponding full-order models. Furthermore, the results indicate that the accuracy requirements in transient simulations are more stringent than in steady-state simulations, i.e. transient models require lower reduction factors for a given accuracy.
Performance of Collocation Models For Simulating Transient Multistage Separation Units
MORBIDELLI, MASSIMO
1995-01-01
Abstract
A systematic analysis has been developed to investigate the performance of reduced-order models for multistage separation units in transient simulations. In particular, we have analyzed the accuracy of these models in reproducing some characteristic features of the dynamic behavior of the original full-order model, i.e. the eigenvalue of smallest modulus lambda(min) and the singular values obtained through open-loop singular value decomposition. It is found that as the relative volatility increases the maximum reduction factor allowed by a prescribed accuracy decreases. The present analysis shows that, with the above limitations on the value of the reduction factor, reduced-order models are able to capture the true dynamics of the corresponding full-order models. Furthermore, the results indicate that the accuracy requirements in transient simulations are more stringent than in steady-state simulations, i.e. transient models require lower reduction factors for a given accuracy.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.