Stochastic state-variable model for a paper-machine headbox

A state-variable model for a pilot plant of paper-machine headbox has been developed using actual input-output data. The linearised physical equations have been considered and effects of plant disturbance and measurement errors have been included. For estimating the unknown parameters of the model from noisy measurements, two recursive estimation techniques, namely `extended Kalman filter¿ and `two-stage estimation algorithm¿, have been used. The model validity has been checked by carrying out residual error tests for both the techniques. Further, estimated responses in both the cases have been compared with the actual output.