Time-delay identification in dynamic processes with disturbance via correlation analysis

A time-delay identification algorithm in closed-loop dynamic processes with disturbance based on maximum correlation analysis was proposed in this paper. To denoise the output data, the response of process disturbance was first estimated on routine data and then eliminated from the output. Next the incremental denoising process data were proposed to be used against the influence of system dynamic on the time-delay estimation. The correlations between the incremental input and different delayed output were computed and the delayed time corresponding to the maximum correlation was identified as the time-delay. The application in the Wood-Berry process and an industrial process showed the validity of the proposed algorithm.