Research Finding: Automated Fault Detection and Diagnostics in AHUs


The Pattern Matching Principal Component Analysis (PCA)-based fault detection method developed by CBEI consistently detected faults at a detection rate of 94% with no false alarms.

This automated fault detection and diagnostics (AFDD) methodology is a low cost minimal touch method of detecting and diagnosing AHU faults because:fault-free-diagnostics-fiugre

  • Passive detection strategy,
  • No requirement for fault data,
  • No modeling requirements,
  • Automated threshold generation, and
  • No requirement for additional sensors beyond what is installed.

Energy impacts of various faults can vary greatly depending on temperatures, loads, and operational conditions. Only via a pattern-matching framework is it possible to identify a suitable baseline for calculation of the energy impact.

This AFDD methodology offers service providers the ability to quantify customer value for FDD in existing buildings and therefore generate a value proposition that can pay for the small incremental software approach and generate customer savings to make the transaction work.

Successful uptake of this new AFDD tool has the potential to improve HVAC performance in commercial buildings.


Publication Title: Research Finding: Automated Fault Detection and Diagnostics in AHUs

Consortium Member(s): Drexel University

Project Contact: Jin Wen

Date: April 07, 2015