Abstract: In recent years, significant research has been conducted on Fault Detection and Diagnosis (FDD) using various approaches. For large-scale and complicated industrial processes early FDD is very challenging task and strenuous. Principle Component Analysis (PCA) and Signed Directed graph (SDG) are monitoring and diagnosis techniques that have been widely implemented in the industry. This study aims to develop and validate a datadriven PCA-SDG tool for FDD. Multivariate PCA was applied to develop fault detection and isolation tool, in addition to the automation of the contribution plot using SDGs. The SDG graphical model has been utilized for capturing process topology and connectivity to highlight the causality between process variables. In this study, two approaches for building the SDG were proposed. The first technique is based on time delay cross-correlations analysis of process data to obtain bivariate correlation coefficients used to construct the fault free SDG. The second approach is based on Transfer Entropy used to validate the associated arcs in SDG by computing the information transfer from one process variable to another. The petrochemical Tennessee Eastman Process (TEP) was used as a FDD case study. The developed tool was able to successfully detect and diagnose the TEP faults and their root causes with a 77% fault detection rate and 87% fault diagnosis rate. The FDD performance shown by the proposed PCA-SDG tool will have a huge impact on process safety, enhance environmental compliance, and significantly reduces financial losses.