Estimation of heat generation by using echo state network for diagnosing thermal characteristics

Estimation of heat generation by using echo state network for diagnosing thermal characteristics

Temperature measurement can be an effective method for diagnosing anomalies in equipment. The amount of internally generated heat may be due to load or an anomaly. Therefore, it is desirable to know how much heat was generated at what time. Echo State Network (ESN), which expresses physical heat tra...

Full description

Bibliographic Details
Main Authors: Yukio Hiranaka, Koichi Tsujino
Format: Article
Language:English
Published: Elsevier 2021-12-01
Series:Measurement: Sensors
Subjects:
Reservoir computing
Echo state network
Heat generation
Temperature measurement
Deconvolution
Inverse estimation
Online Access:http://www.sciencedirect.com/science/article/pii/S2665917421002014
Description
Summary:Temperature measurement can be an effective method for diagnosing anomalies in equipment. The amount of internally generated heat may be due to load or an anomaly. Therefore, it is desirable to know how much heat was generated at what time. Echo State Network (ESN), which expresses physical heat transfer in a form of neural network, is a promising method for estimating the amount of heat by performing inverse processing of temperature changes. The authors have developed a method to learn the inverse transfer characteristics by using ESN. This paper reports the detail of estimation of heat generation along with the results using synthesized and measured data.
ISSN:2665-9174

Similar Items