| With the development of modern industries in macro-scale and integration, industrial processes have become more and more complex and contain a large number of measured variables. Once fault occurs, there will be the huge amount loss between human resources and economics. Therefore, the processing monitoring fault diagnosis has been become one of the most active research areas in process control. Meanwhile, with the progress of the CIPS, a great amount of process data can be sampled and collected in time. This assignment based on the development of data-driven process monitoring simulator, which utilize testable processing data and accurately monitor or control the process of process industrial production in time. All contents in the paper provide important meanings for system reliability improvement.According to the data-driven process monitoring simulator, the paper gives following results:(1) Aiming at analysis, research, and example validates of each data- driven algorithm, it can provide following conclusions:1) The liner span of PCA is greater than FDA's, which is more suitable for data separation. But FDA considers about the relationships among classes, it can be used better for class separation.2) During the lower dimensional class separation, PCA is similar with PLS. However, PCA provide much better result during the higher dimensional class separation.3) In the nonlinear situation, CVA utilized weights for maximum pertinency in variable field. However, PCA, FDA, and PLS are all belong to the liner system.(2)The paper gives a new data-driven algorithm. Compare with other algorithms, the improved algorithm reaches higher effects for data separation and better monitoring or controlling ability.(3) The assignment designs the structure of process industrial process monitoring simulator, which is based on data-driven. It contents: structure design of software and hardware, opening design of simulating platform, and opening realization of simulating... |
PDF Full Text Request