A statistically based fault detection and diagnosis approach for non-residential building water distribution systems

•PCA and SVM were combined to build a fault detection and diagnosis method for non-residential building water networks.•Data from a non-residential building was used to validate the developed fault detection and diagnosis model.•Improved performance monitoring of non-residential water distribution systems can be achieved using limited process data.•A comparative study illustrated that the proposed approach detected and diagnosed more actual system alarms than a conventional approach, while also utilizing less computational resources.