摘要 : This article proposes a novel reduced fault detection method based on the data-driven approaches for process monitoring. An overview on machine learning methods took place to illustrate the efficiencies for Fault Detection (FD). T... 展开 This article proposes a novel reduced fault detection method based on the data-driven approaches for process monitoring. An overview on machine learning methods took place to illustrate the efficiencies for Fault Detection (FD). The concept, the original ideas, the reduced steps, data storage and implementation conditions are discussed. The performance from the application viewpoint of the proposed Reduced Rank KPLS technique and all discussed methods are validated using an industrial benchmark. 收起