Modern industrial processes are often accompanied with high temperature, high pressure, flammable and explosive environment, which makes the safety and reliability very important. Based on the dimension decrease and de-noise of the data by principal component analysis, a support vector data description approach is applied to process monitoring, which is independent of the normal distribution of the data support vector data description and its application to Tennessee Eastman process. The application to Tennessee Eastman process has proved the sensitivity of the method, which is helpful to taking measures on time and stablizing the product quality.
展开▼
