Nonlinear system modelling using output error estimation of a local model network

Gray, G.J., Murray-Smith, D.J., Li, Y. and Sharman, K.C. (1996) Nonlinear system modelling using output error estimation of a local model network. In: Society for Computer Simulation Summer Computer Simulation Conference (SCSC '96), Portland, Oregon, USA, July 1996, pp. 460-465.

Abstract

The local model network is a set of models, each describing the same dynamic system but at different operating points. The outputs of these local models are weighted according to the current operating point and summed to give the local model network output. A local model network can be constructed from nonlinear continuous local models. The parameters of the local models can be identified using output error or maximum likelihood estimation. The results of this parameter identification can then be used as part of a structural estimation procedure indicating how parameters change with system parameters. This network can be simulated using a variable step length algorithm in a format amenable to a standard simulation program. An example is given in which the flow in a coupled tank system is analysed.

• Item Type: Conference Proceedings
• Additional Information: Proceedings published by Society for Computer Simulation, San Diego, California, USA, 1996.
• Keywords: Parameter identification, nonlinear models, local model networks, model validation, fluid flow, coupled-tanks system
• Status: Published
• Refereed: Yes
• Glasgow Author(s) Enlighten ID: Murray-Smith, Professor David and Li, Professor Yun
• Authors: Gray, G.J., Murray-Smith, D.J., Li, Y., and Sharman, K.C.
• Subjects: Q Science > QA Mathematics > QA76 Computer software; T Technology > TJ Mechanical engineering and machinery; T Technology > TK Electrical engineering. Electronics Nuclear engineering
• College/School: College of Science and Engineering > School of Engineering > Systems Power and Energy