On the use of correlation analysis in the estimation of finite-time delay in Soft Sensors design

Finite-time delay characterizes many processes, including industrial ones. Soft Sensors can take into account the delay and give real-time estimations of process variables. The estimation of the delay is relevant in Soft Sensors design. In the paper, the estimation of the time delay by using input-output correlation and a suitable norm is investigated. The role of different correlation indexes is analysed by using two classes of cases study. These include various nonlinearities and process data taken from different sources. Results show the suitability of the proposed norm to detect the finite delay when used with the suitable correlation index.