Hazard and operability (Hazop) analysis is the study of systematically identifying every conceivable deviation, all the possible abnormal causes for such deviation, and the adverse hazardous consequences of that deviation in a chemical plant. Modern chemical plants are extremely complex and hence are difficult to analyze and assess from this Hazop perspective, thus raising serious occupational safety related concerns. Because of this complexity, they are also more vulnerable to equipment failures, as witnessed by the recent chemical plant accidents. The major goal of this proposal is to develop a knowledge-based system framework for automating Hazop analysis of process plants. Such automated systems are extremely important for improving the occupational safety of chemical plants. Hazop analysis is often carried out by a group of experts poring over the process flowsheets for weeks or months. Thus, Hazop analysis is a very difficult, labor-intensive, and time-consuming process. In our proposed knowledge-based framework, we recognize and exploit two important features of the Hazop analysis: (i) even though each Hazop analysis is unique-to a process, it is systematic and logical; and (ii) many aspects of the analysis are the same for different process flow sheets. Thus, Hazop analysis can be automated through the use of knowledge-based systems. In this project, we propose a knowledge-based systems approach which introduces several novel techniques in a model-based framework to attack this problem. These are: (i) decomposing the knowledge-base into process specific and process general knowledge, (ii) developing generic cause-and-effect models of various processes and process equipments, and (iii) an object-oriented implementation framework. We have successfully tested the proposed approach on a small prototypical case study. Encouraged by the preliminary results, we propose to investigate this approach further by developing more comprehensive generic models library of complex process equipments, process interactions, process materials, and by testing on complex industrial scale Hazop case studies. We also propose to develop efficient search techniques for managing the complexity, handle recycle and feedback loops, and develop an intelligent object-oriented graphical interface for this system.