Skip to main content

Combustion Condition Monitoring Through Deep Learning Networks

Hossain, MD Moinul (2019) Combustion Condition Monitoring Through Deep Learning Networks. In: 11th International Conference on Applied Energy. . pp. 1-5. Elsevier, USA

PDF - Author's Accepted Manuscript
Download (1MB) Preview
[img]
Preview
Official URL
https://www.journals.elsevier.com/applied-energy/c...

Abstract

Combustion condition monitoring is essential in a power plant for maintaining stable operations and operational safety. Therefore it is crucial to develop an intelligent combustion monitoring system. Existing traditional methods not only need a large quantity of labeled data but also require rebuilding monitoring model for new conditions. Aiming these problems, the present study proposes a novel approach combining denoising auto-encoder (DAE) and generative adversarial network (GAN) to monitor combustion condition. By using the learning mechanism of the GAN, the robust feature extraction ability of DAE as a generator is improved. These features are then fed into the Gaussian process classifier (GPC) for condition identification. Especially, newly occurring conditions can be correctly classified by simply training the GPC, rather than training from scratch. Experiments performed on a gaseous combustor indicate that the proposed approach can extract representative features accurately and achieve high performance in combustion condition monitoring with the accuracy of 98.5% for original conditions and 97.8% for the new conditions.

Item Type: Conference or workshop item (Proceeding)
Uncontrolled keywords: Combustion condition monitoring, Generative adversarial network, Gaussian process classifier
Subjects: Q Science
Divisions: Faculties > Sciences > School of Engineering and Digital Arts > Instrumentation, Control and Embedded Systems
Depositing User: Md Moinul Hossain
Date Deposited: 21 Aug 2019 10:38 UTC
Last Modified: 18 Dec 2019 14:25 UTC
Resource URI: https://kar.kent.ac.uk/id/eprint/75918 (The current URI for this page, for reference purposes)
Hossain, MD Moinul: https://orcid.org/0000-0003-4184-2397
  • Depositors only (login required):

Downloads

Downloads per month over past year