Improper process control can lead to significant increase in the process operation cost. Industrial processes control incorporates various types of control systems (e.g., supervisory control and data acquisition (SCADA) systems and distributed control systems (DCS) in petrochemical plants). Troubleshooting of industrial processes is critical for maintaining product quality, safe operation, and efficiency. The complication of petrochemical plants troubleshooting increases with intensifying plant problem. When there is a delay to make an appropriate decision to deal with the process trouble, outstanding detriments and adverse influences might take place for both plant operations and human considerations. The establishment of a self-regulating and routine control to operate a complex plant safely and economically is required to accurately and effectively identify the defects of the plant. Fuzzy logic is a convenient approach to manage the ill-defined information quality and the vagueness of human decision making. This research focuses on the development of a fuzzy logic model to improve the hydropurification reaction systems operation. The main objective of the study is to determine the leading operational parameters of the system and give solutions and systematic procedures to cope with the raised problem through fuzzy inference approach. The developed fuzzy rule-based expert system model is applied to the hydropurification reaction process/system of purified terephthalic acid (PTA) production plant. Hydropurification reaction system of PTA production is the most influential unit of crude terephthalic acid purification process, which crucially determines the final product quality. Its proper control and operational optimization can achieve significant economic benefits since the system deals with slurry phase at high pressure and temperature. This system consists of many pieces of equipment including a fixed-bed reactor, heat exchangers, hydrogen compressors, a high-pressure steam production boiler, and crystallizers. The model is established by selecting the vital controlling parameters which affect the product quality, palladium supported on carbon (Pd/C) catalyst deactivation, the equipment lifetime, and the process safety. The model is able to give proper actions relevant to each operating parameter fluctuation. The program outputs give information on the effects of any controlling parameter on the system and the final decision in cases of different types of alarms. The troubleshooting model is incorporated in a computer program coded into MATLAB software environment with detailed explanation. The model is evaluated and validated using the data of an industrial PTA hydropurification plant. The developed computer program can be used in parallel with the control system (e.g., DCS) of the process. Therefore, the model can be considered as a valuable training tool for the field and DCS room operators and engineers. The implementation of the model into the actual process can reduce the number of shutdowns, decrease the consumption of the chemicals (e.g., caustic), decline the maintenance time, increase the lifetime of equipment, lengthen the expensive catalyst lifetime, lessen the operation cost, and rectify the safety concerns of the process. The developed model can be effectively employed/used for similar processes dealing with the costly hydropurification processes.
