A Causality Capturing Method for Diagnosis Based on Transfer Entropy by Analyzing Trends of Time Series

Abstract

Since modern industrial processes become much larger and more complex, efficient and effective causality detection methods are needed to capture the process topology, diagnose root causes of widespread or even plant-wide process malfunction, and further ensure the safety of processes. A modified transfer entropy method, named trend transfer entropy, is proposed in this paper, which focuses on analyzing trends of time series rather than the original series themselves and thus, compared to the traditional transfer entropy, proves to be more robust in conditions of data drifting and noise disturbance. Moreover, the new method can reduce computational load effectively, saving valuable time before the occurrence of an accident. Simulation studies are presented to illustrate the procedure and features of the proposed method.

Sensor self-diagnosis method based on a graph neural network

“…The data between different sensors in the system are not isolated, but interdependent. Due to the unique advantages of redundancy in terms of the analysis and processing of system related Most current studies have used the idea of analytical redundancy to evaluate the reliability of the system, or have applied redundancy to the analysis and processing of variables, and have not quantitatively evaluated the redundancy levels of sensors [20]. Following recent developments in information technology, the measurement data from different sensors can now be jointly analysed, and relative entropy theory can be introduced to analyse the redundancy between sensor measurement data, which provides a basis for the self-diagnosis of sensors.…”

Detection of Cause-Effect Relations Based on Information Granulation and Transfer Entropy

“…Reference [ 19 ] proposed the transfer zero-entropy for causality analysis based on the zero-entropy and zero-information without assuming a probability space. Additionally, symbolic transfer entropy [ 20 ] and trend transfer entropy [ 21 ] extended the TE to symbols or trends of time series instead of original continuous values. The multiple-unit symbolic dynamics and transfer entropy were used to analyze the dynamic causal relationships in longitudinal data [ 22 ].…”