A Novel Safety Assessment Approach Based on Evolutionary Clustering Learning

The safety risk assessment is a structured and systematic methodology aiming at enhancing the complex engineering system safety. It has been gradually and broadly used in the industrial process control system nowadays around the world. In this paper, a novel safety assessment approach based on evolutionary dictionary learning and fault tree analysis for the complex engineering system is proposed. First, historical signals are utilized to conduct the clustering learning dictionaries by norm similarity matching model and patch-based evolutionary dictionary learning algorithm. Second, the support vector machine method is employed to identify and reflect the normal and fault operating states. Third, an improved safety risks method is proposed to reflect the probable hazards of different faults on the basis of the fault tree analysis. Finally, this processing on online signals is to offer an effective safety assessment index and update the evolutionary dictionaries and safety routing metrics. The related experiments are constructed to demonstrate that our proposed approach can achieve high performance.