Process operation improvement methodology based on statistical data analysis
통계적 분석기법을 이용한 공정 운전 향상의 방법
With disseminationof Distributed Control Systems(DCS), the huge amounts of process operation data could have been available and led to figure out process behaviors better on the statistical basis. Until now, the statistical modeling technology has been susally applied to process monitoring and fault diagnosis. however, it has been also thought that these process information, extracted from statistical analysis, might serve a great opportunity for process operation improvements and process improvements. This paper proposed a general methodolgy for process operation improvements including data analysis, backing up the result of analysis based on the methodology, and the mapping physical physical phenomena to the Principal Components(PC) which is the most distinguished feature in the methodology form traditional statistical analyses. The application of the proposed methodology to the Balst Furnace(BF) process has been presented for details. The BF process is one of the complicated processes, due to the highly nonlinear and correlated behaviors, and so the analysis for the process based on the mathematical modeling has been very difficult. So the statisitical analysis has come forward as a alternative way for the useful analysis. Using the proposed methodology, we could interpret the complicated process, the BF, better than any other mathematical methods and find the direction for process operation improvement. The direction of process operationimprovement, in the BF case, is to increase the fludization and the permeability, while decreasing the effect of tapping operation. These guide directions, with those physical meanings, could save fuel cost and process operator's pressure for proper actions, the better set point changes, in addition to the assistance with the better knowledge of the process. Open to set point change, the BF has a variety of steady state modes. In usual almost chemical processes are under the same situation with the BF in the point of multimode steady states. The proposed methodology focused on the application to the multimode steady state process such as the BF, consequently can be applied to any chemical processes set point changing whether operator intervened or not.