| Issue |
MATEC Web Conf.
Volume 272, 2019
2018 2nd International Conference on Functional Materials and Chemical Engineering (ICFMCE 2018)
|
|
|---|---|---|
| Article Number | 01003 | |
| Number of page(s) | 7 | |
| DOI | https://doi.org/10.1051/matecconf/201927201003 | |
| Published online | 13 March 2019 | |
Data-driven state monitoring of air preheater using density peaks clustering and evidential Knearest neighbour classifier
1 Key Laboratory of Energy Thermal Conversion and Control, Ministry of Education, School of Energy and Environment, Southeast University, NanJing, 210096, P.R. China
2 Department of Chemical and Biological Engineering, University of Sheffield, Sheffield S1 3JD, UK
* Corresponding author: shenj@seu.edu.cn
Data-driven state monitoring requiring a little priori knowledge plays a key role for timely fault detection and is therefore of great importance for the safe and economical operation of the thermal power plant. The main drawback for most of the existing data-driven methods is the complex procedure of data preprocessing and model training especially when unlabelled operating data is used. To overcome this issue, this paper proposes a new framework of data-driven state monitoring approach for the thermal power plant devices. The approach is composed of two steps. In the first step, density peaks clustering(DPC) is performed on the historical data to generate labels for the data. Then in the second step, evidential K-nearest neighbour(EKNN) method is used to monitor the current state based on the labelled historical data and operating data. Verifications on operating data of an air preheater system of a 1000MW thermal power plant show that the proposed method can identify various air leakage states accurately and efficiently.
© The Authors, published by EDP Sciences, 2019
This is an Open Access article distributed under the terms of the Creative Commons Attribution License 4.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.