(32m) Root Cause Diagnosis and Analysis Method for Abnormal Conditions in Chemical Processes Based on Mechanism and Real-Time Data Correlation Analysis

Abnormal conditions are precursors of chemical process safety accidents. Timely diagnosis and analysis of the root causes of abnormal conditions are of great significance for preventing chemical process safety accidents. However, the current root cause diagnosis method is mainly based on expert knowledge, which is subjective and time-consuming, limiting its applicability in large-scale and complex processes. In addition, the causal relationship between variables may change under certain process conditions. Therefore, it is difficult for methods based purely on expert knowledge to diagnose the root causes of abnormal conditions efficiently and accurately. In response to the above problems, this paper proposes a root cause diagnosis and analysis method for abnormal conditions in chemical processes that integrates mechanism and real-time data correlation analysis. First, a symbolic directed graph model is constructed based on process mechanism analysis to qualitatively analyze the propagation path of abnormal conditions; then, based on the time series data between abnormal condition-related variables determined by qualitative analysis, the Granger causality test method is used to quantitatively analyze the causal relationship between variables and determine the propagation path and root cause of abnormal conditions; finally, the proposed method is applied to the abnormal condition of crude oil with water in a constant pressure reduction unit as an example. The results show that the root cause diagnosis method that integrates knowledge and real-time data correlation analysis can more accurately identify the root cause of the abnormal condition of crude oil with water as the high flow rate of crude oil into the unit.