Design for Process Security

Chemical process security has been an area of interest due to the potential of process industries as a possible terrorism target. As a result, model-based process security analysis methods that try to detect process vulnerabilities in a fundamental way have been proposed. The design aspect of the problem, however, has not been addressed yet. In this work, an adverse-MPC approach, which enables estimation of security vulnerability for a given process is proposed, and is coupled by a secure-design problem that retrofits the design to minimize the vulnerability. The proposed method is realized using a bi-level formulation, where the security evaluation is posed as a constraint in the outer design problem. The method described in this work enables design of processes that are inherently less vulnerable to security threats. Further, the method can be extended to combine economics into the security problem, hence enabling design of optimal designs in terms of both security and cost efficiency. The algorithm can also be utilized for design of robust security systems. The application potential of the approach is demonstrated by a runaway reactor example.