Use Of Multidimensional Scaling For Fault Detection Or Monitoring Support In A Continuous Commissioning

Abstract

It is well known that the building energy userepresents a significant part of the total energy use,ca. 40% in USA according to the Building Energy Data Book. With the improvement of new construction's efficiency, the share of equipment's energy use increases more and more compared to the overall building energy use. This article proposes to study a new approach in building's systems Fault Detection and Diagnosis (FDD), so as to provide an intuitive FDD tool for every operator of the maintenance staff regardless of their qualifications. This new approach uses a method of multivariate statistics, provides easily understandable outputs allowing a quick comprehension of the equipment fault by building maintenance staff. Thereby the number of unsolved problems can be minimized and the intervention time would be considerably reduced, avoiding unexpected energy use and equipment's premature obsolescence.

Experimental dataset for an AHU air-to-air heat exchanger with normal and simulated fault operations

“…Two recent studies may be cited since they are based on the same AHU equipment. In [46], a multivariate statistics method was employed to elaborate a map of the commissioning of the equipment. One year of experimental data were collected from the automation system of an office building.…”