2013 Volume 14 Pages 11-22
In industrial plants, soft sensors have been widely used to estimate difficult-to-measure process variables online. The predictive accuracy of soft sensors decreases due to changes in the state of chemical plants, and soft sensor models must adapt to the process changes by using new measured data. However when a model is reconstructed with data that have low variation, the model cannot predict abrupt changes of process characteristics. The predictive performance of adaptive models depends on databases. We therefore propose an index to monitor database, i.e. database monitoring index (DMI), and a database monitoring method using the DMI. The DMI is based on similarity between two data and is defined as a rate of absolute difference of an objective variable and similarity of explanatory variables. The more similar two data are, the smaller value the DMI has. When new data is obtained, DMI values are calculated between new data and all data in a database. If the minimum value of the DMI values is large, the new data is added to the database. By using the DMI and selecting new data, the amount of information of a database can enlarge while curbing the rise in the number of data in the database. Through the analysis of simulation data, we confirmed that the appropriate monitoring of a database could be achieved and the preidctive accuracy of adaptive soft sensor models could increase by using the proposed DMI.