In modern chemical processes

disturbances often occur at a local equipment or stream.It will affect the performance of the whole process due to connectivity of streams and complex control system.To identify the root cause of the fault and the path in which the disturbance propagates is of great significance for the safety or economics of chemical production.The algorithm of time-delay analysis is regarded as a practical method for fault identification with some good aspects

such as model-free causal analysis

fully data driven methods

simple interpretation.An improved time-delay approach

which combined wavelet denoising and node-divide techniques was proposed.The new method could overcome the strong noise characteristics of the data from chemical production.And it could reduce the interference of redundancy delay which was inherent of time-delay algorithm used for plant-wide analysis.The approach was applied to the Tennessee Eastman(TE)model and the result proved its good quality for industrial application.