Grant: Discovery Grants Program - Individual

In industry, distributed parameter systems are widespread. Many high performance control methods have been proposed for distributed parameter systems but most have not been able to be applied in practice. However, high fidelity models are necessary to promote advanced control methods to industrial applications. This research program proposes to develop identification techniques using empirical data so that distributed parameter models for industrial systems can be obtained with high accuracy. State space models have been one of the most preferred model forms in system engineering. 2-dimensional state space models that can be used to represent distributed parameter systems were proposed in the 1970's and have interested many researchers since. Identification of 2-dimensional state space models has been considered a very challenging topic and only limited results are available. In this research program, it is proposed to explore 2-D discrete state space models and the identification techniques of the model forms. Methods for extending the subspace identification methods available for lumped parameter systems to distributed parameter systems will be investigated. It is expected that the identification methods from this research program will generate high fidelity models for distributed parameter systems for process industries, and will help to promote application of distributed parameter control techniques.