Alarm Optimization Research And Application Of Petrochemical Plant Based On Mechanism Model

Alarm system constitutes an important part of the user interface of the large modern industrial facilities, so it plays an important role in the prevention, control and mitigate of abnormal situations. Therefore, the optimization of the alarm and alarm processing strategy is significance for the normal operation of the chemical plant. Distillation column is a mass transfer equipment widely used in chemical industry production, having a very important position in the chemical industry. The safety and smooth operation of distillation column has become an important part of chemical production, also directly affects the economic benefits of the chemical industry.Firstly, we makes a statistic analysis on the key performance indicators of the alarm system to the alarm system makes a statistic analysis on the key performance indicators. On the basis of key performance indicators, the alarm system is evaluated for how to optimize the alarm system. Then on the basis of dead area, we introduces the alarm limit optimization design based on data filtering to solve repeating alarm, and use data filtering effectively to verify the process data with noise and enhance the robustness. The combination of dead area and data filtering prevent the produce of repeating alarm effectively. But, alarm optimization based on the measured signal is very limited, finding alarm root can thoroughly solve the alarm problem. Fault diagnosis strategy also is a kind of effective method to deal with the alarm, because the alarm will also disappear as long as the fault is eliminated, the alarm will also disappear. Secondly, we carried out the fault diagnosis and detection methods based on the two-tier model. It employs the nonlinear model developed earlier to monitor distillation process and a corresponding linear model to identify abnormal source when the large deviations of measured values occur. The inner distillation fault parameters are estimated through linear least square method based on the linear model. Third, we use Himmelblau algorithm to decompose a large system of chemical process which is high-dimensional and difficult to solve, into a number of low-dimensional sub-problems; then we use median filter and lifting wavelet analysis method to remove noise from the small systems and to enhance the effect of robust.This method is applied to the stripping tower of Tennessee Eastman Process (TEP), and pressure loss coefficient change in bottom feed is analyzed. Case study shows that this method can not only give the exact cause of the malfunction and shorten the time of diagnosis, but also improve the diagnostic accuracy.