Paper Conference

Proceedings of eSim 2018: 10th Conference of IBPSA-Canada


Principal Component Analysis for the ongoing commissioning of northern houses

Behrad Bezyan, Radu Zmeureanu

Abstract: This paper presents the application of the Principal Component Analysis (PCA) for defining the threshold of normal operation of the heating system in two houses, detecting the outliers in the PC-based space of the heating system operation, and identifying those system variables which are the source of outliers. This case study uses measurements collected in December 2014 in two houses of Inuvik, NWT, as the training data set, which is then applied to measurements of February 2015 as the application data set. The temperature of supply and return water temperature for heating one house are the major sources of outliers identified from data of February 2015. The identification by the PCA of variables with abnormal values is validated by using of a modified data set.
Keywords: Principal Component Analysis, outlier detection, heating, ongoing commissioning, northern houses
Pages: 1 - 8