Mao Qianjun, Liang Zhiyuan, Liu Donghua, et al. Temperature Sensor Fault Detection in Chiller Based on One-class Support Vector Machine Algorithm[J]. Journal of refrigeration, 2019, 40(5).
Mao Qianjun, Liang Zhiyuan, Liu Donghua, et al. Temperature Sensor Fault Detection in Chiller Based on One-class Support Vector Machine Algorithm[J]. Journal of refrigeration, 2019, 40(5). DOI: 10.3969/j.issn.0253-4339.2019.05.130.
Temperature sensor faults may lead to abnormal system operations that can damage the chiller system and reduce its life span. Herein
a fault detection method based on the one-class support vector machine (OC-SVM) algorithm has been proposed. Fault-free data were used to train the OC-SVM model for detection of temperature sensor biases. The optimized model parameters were obtained by the 10-fold cross validation method. Four chiller datasets
including in-site and laboratory data
were used to validate the proposed method. Results showed that the OC-SVM showed good fault detection performance on the four chiller datasets
with the effect of fault detection being especially obvious for the screw chiller (dataset I). The detection efficiency reached 100%
when the absolute value of temperature sensor fault biases at chilled-water side was greater than 1 ℃.