Lumped-parameter modelling and identification of distributed-parameter systems. Part 2: Model-parameter estimation using p.r.b.s.

Methods for estimating the parameters of the lumped difference equation models postulated for distributed parameter systems are presented. These methods utilise pseudorandom binary sequences to yield spatial and temporal pulse responses which may then be processed using instrumental variable techniques to give parameter estimates. Alternatively, regression methods are described which process either the system output at various points in space or the corresponding pulse response to give similar estimates. Computer simulation results are used to confirm the theoretical concepts.