Hazard And Operability Study On Unsteady Operation Process In Prtroleum And Chemical Industry

In petrochemical production process, besides continuous steady operation mode, there are also unsteady operation modes, such as batch operation, startup, and shutdown processes. Comparing with continuous steady operation, unsteady operation processes tend to abnormal situations, at the worst accidents, due to involved a large number of operations of operators and periodically dynamic changes of process state.HAZOP (Hazard and Operability) analysis is recognized as the preferred Process Hazard Analysis (PHA) approach. HAZOP is widely used to identify potential hazards and operability issues. By using HAZOP analysis method, every conceivable process deviation, possible abnormal causes of deviation, and possible adverse consequences led by deviation can be identified systematically.Conventional HAZOP analysis is performed manually by domain experts, and it is a time-and labor-intensive task. Comparing with manual HAZOP analysis for continuous steady operation, which only need identify causes and consequences when process variables are too high and too low, manual HAZOP analysis for unsteady operation is not only time-consuming but also more difficult:(1) Unsteady operation process is a periodically dynamic change process. Potential adverse effect of one wrong operation may occur during later operation. Experts need to remember previous operation deviation during HAZOP analysis; (2) when two or more operation deviations are involved, experts need to analyze compositive effects of sequential maloperations.In order to eliminate burden of manual analysis, some automatic methods to perform HAZOP analysis for unsteady operation have been proposed. However, current methods still have following defects, which can be improved.Modeling aspect:(1) Lack of describing dynamic changes of variables and dynamic effects among variables when one operation is being performed. (2) Lack of studying modeling method for control loops.Automatic HAZOP analysis aspect:(1) Few guidewords relating to operation can be used. (2) Lack of sysmatically studying maloperation patterns. (3) Lack of studying automatic algorithm of producing maloperations. (4) Outputted adverse consequence content when variables are too high or too low need be added in model manually. This task increases modeling workload, and added content completely depends on modeling person's understanding on production unit. (5) Lack of studying automatic risk indentification for maloperation.Integration of automatic and manual analysis aspect:(1) Lack of focusing on verification for automatic HAZOP analysis results. (2) Lack of systematically identifying reasons for maloperations.In order to address these defects mentioned above, firstly a qualitative modeling and simulation method for unsteady operation process based on fuzzy logic was proposed. Process principle is expressed by Petri net-Fuzzy Directed Graph model. In order to address the 1st and 2nd defects in modeling aspect, effects among variables during dynamic changes and effects of continuous steady controllers on production process are involved in Fuzzy Directed Graph model. In order to address the 1st defect in automatic HAZOP analysis aspect, operating procedure is expressed in formal way so as to generate maloperations easily according to guidewords. Then operating process including maloperations is simulated qualitatively according to qualitative simulation algorithm.Based on the qualitative modeling and simulation method, an automatic HAZOP analysis method for unsteady operation process was proposed. In automatic HAZOP analysis aspect, in order to address 2nd defect, types of maloperation pattern and relationships between maloperation patterns and guidewords were studied in this paper; in order to address to 3rd defect, producing algorithms of operating procedure sets for single maloperation pattern and multiple maloperation patterns were proposed; in order to address the 4th defect, when effects of every operating procedure in the sets on production process are simulated qualitatively, if change strength of variable is too high or too low, variable value is too high or too low, or leak event occurs, use "Hazard Patterns for Too High/Too Low of Process Variable" and "Material Hazard Digraph" to automatically analyze possible hazards; in order to address the 5th defect, use formula to judge probability rank of one operating procedure happening, use properties relating to safety and amounts of materials in malfunction equipment to judge severity rank of consequence, and use predefined Risk Matrix to judge risk of operation process, which is represented by this operating procedure.Then, for unsteady operation process, a method of combining automatic HAZOP analysis and manual HAZOP analysis was proposed. In integration of automatic and manual analysis aspect, in order to address the 1st defect, a manual verification method for automatic HAZOP analysis results was proposed; in order to address the 2nd defect, various levels of factors affecting operators'performance were summarized, reasons and measures relating to these factors were summarized, and a procedure of systematically identifying maloperation reasons and proposing measures was given.In this paper, a batch production process case was used to explain proposed method. Finally, a debutanizing column case was used to completely veriy proposed method. These two cases verified that proposed method can well address defects of current automatic HAZOP analysis methods for unsteady operation process.Finally, proposed method integrates advantages of automatic analysis and manual analysis:can quickly get more complete analysis results by using automatic analysis; can verify analysis results, identify maloperation reasons and propose measures according to experts' experience and real conditions of plant; then a well complete and valid analysis report will be got. By using these analysis results, some proper measures can be taken to effectively avoid maloperations or significantly reduce hazards led by maloperations. Therefore, the research content of this paper has theoretical and practical significance.