Parameter Estimation for the Wiener Dynamic System With Unmeasured Continuous-Time Correlated Stochastic Disturbances

Unmeasured disturbances in chemical processes are often not stationary and frequently have a time-correlated, continuous-time (CT), stochastic behavior. The common way of addressing this type of disturbance is to treat it discretely as a noise component of the output. Not only can this treatment lead to inaccuracy caused by a discrete approximation of CT noise, but it requires a constant sampling time of the outputs to use past estimates of the noise in on-line prediction. The CT method in this work is not restricted by the discrete approximation and proposes a method for estimating dynamic and noise parameters for first-order, single-input, single-output (SISO) Wiener systems with a CT stochastic disturbance. Though a simulation study, the accuracy of the proposed method is demonstrated and compared to other methods in the literature.