Safety instrumented system for process operation based on real-time monitoring.

摘要

Many industrial processes pose many potential threats to life or environment, especially in case of failure. Hazardous, flammable and reactive materials are often processed at elevated temperatures and pressures. The hazards posed by those materials need to be controlled and managed in order to improve the process safety. Safety Instrumented System (SIS) is a widely recognized tool by a number of industry sectors to prevent those hazards and thus reach the required safety objective. Recently, the process industry has started to realize the significance of SIS.; Due to the increased process complexity and possible instability in operating conditions, the existing control systems have limited ability to provide practical assistance to both operators and engineers. Therefore, much attention has been focus on suitable designs of system control components. This thesis proposes a new methodology for fault diagnosis, safety function formulation, and to implement safety instrumented system based on real-time monitoring. The methodology is comprised of three stages. The first stage is to model and simulate the target process system according to the observed system behaviors. The second stage is to adopt knowledge-based fault diagnosis technique, which implements the valuable knowledge from the experts and operators as well as a vast databank of information from a variety of sensors, for making optimal decision regarding current state of the process operation. Fuzzy logic is also used in this stage to make inferences based on acquired information (real-time data) and the knowledge. This stage is a fundamental part of the proposed methodology.; A computer-aided tool, implementing previous two stages, is developed on the platform of G2 expert system platform using GDA (G2 Diagnostic Assistant) components in the third stage. This tool is subsequently used to verify the methodology performance through both industrial and simulated data.; The proposed methodology is straightforward, flexible and easy to understand. Moreover, the developed fault diagnosis safety function may be utilized in developing various safety instrumented systems.