No abstract

A plant alarm system provides critical information to operators as the third layer of independent protection layers when a chemical plant is undergoing an abnormal situation. Therefore, methods for designing plant alarm systems are crucial for plant safety. Because plants are maintained as part of the plant lifecycle, plant alarm systems should be properly managed and adapted to changes that occur throughout the lifecycle. To manage changes, the design rationale of the plant alarm system should be explained explicitly. In the present paper, we explain a design problem for plant alarm systems and propose a method for designing alarm systems by using di erential and algebraic equations, cause-e ect matrices, and preference indices.

Recently, cyber security becomes a serious problem for not only OA (Office Automation) systems but also PA (Process Automation) and FA (Factory Automation) systems. Even the controllers, which are not connected to Internet directly, have been attacked with malwares, such as Stuxnet and Quantum. When control system fails, it may lead to serious accidents such as explosion or leakage of poisonous and deleterious substances. For process control, cyber-attack is one of the causes to threaten safety. The authors of this paper had proposed zone division of process control networks to ensure cyber security and safety. To apply the method, it is necessary to build CE (Cause Effect) matrices which express the qualitative information of the plant and controllers. It is very troublesome for large-scale plants. CAD (Computer Aided Design) tool for zone dividing is proposed in this paper. CE matrices are generated by using DAE (Differential and Algebraic Equation) registered in equipment modules of plant CAD such as ASPEN or Pro II. The candidates of zone division of process control networks, which can assure the safety against concealment and remote operation by cyber attackers, can be proposed.