In-tube Oxide Scale Plugging Warning of Supercritical Boiler Based on Data Stream Outlier Mining Technology

Zhigang Chen,Danjue Yuan,Hua Li,Zhending Zhang,Weiwen Ye,Dingping Liu,Kaihua Xu
DOI: https://doi.org/10.12783/dtcse/iceiti2016/6200
2017-01-01
DEStech Transactions on Computer Science and Engineering
Abstract:As supercritical boiler(SCB) works chronically under high temperature and high pressure, oxide scales easily generate and peel off inside heating surface tubes, which ends up in plugging. Aiming at massive amounts of real-time data produced in the operation of SCB, a novel algorithm, based on data stream outlier mining technology, was proposed for detecting wall temperature stream outliers in this paper to implement oxide scale plugging warning timely. By modeling and calculating, data stream labels were used to mark the abnormal frequent items. The warning would be triggered if they were marked three times continuously. The comparison of the experimental results and the real data streams proves that the algorithm has higher reliability and better accuracy.
What problem does this paper attempt to address?