BTRL ribbon

Remote Diagnosis of AC Equipment

Principal Investigators:
Dr. Afshin Afshari, Professor of Practice, Engineering Systems and Management, Masdar Institute
Dr. Peter Armstrong, Associate Professor in Mechanical Engineering, Masdar Institute

Brief:
Predictive maintenance programs incorporate some type of measurement and/or monitoring of equipment in order to observe or predict equipment degradation or failure. These measurements are able to detect the onset of problems or degradation of the equipment or a particular mechanism within the equipment before partial or total failure occurs.

Predictive maintenance bases maintenance requirements on the actual state of the equipment, rather than a preset schedule. Predictive maintenance can increase the life of the equipment and decrease downtime, parts and labor costs, while providing energy savings. Automated fault detection and diagnostics (FDD) provides a cornerstone for predictive maintenance of engineered systems.

Although FDD is an active area of research in many engineering fields, applications for heating, ventilating, air conditioning (HVAC) and other building systems have lagged those in other industries. We will start by assessing generic FDD and prognostics, providing a framework for categorizing methods, describing them, and identifying their primary strengths and weaknesses. We then move on to address research and applications specific to the fields of HVAC. We will evaluate the effectiveness of FDD methods for the considered applications and produce specifications for algorithms that could be incorporated within commercial products. Emulation and field tests follow.

Objectives:
To improve the operating efficiency of commercial HVAC systems by 15-30% through development and demonstration of the enabling technologies for detecting faults and control errors; specifically:

  • Identify abnormal conditions accurately.
  • Do not give false alarms of abnormal conditions (i.e., be robust).
  • Report the level of confidence associated with each diagnosis.
  • Rank the conclusions.
  • Be able to detect the necessary faults at the stipulated sensitivity level (i.e., at the pre-specified fault severity level).
  • As far as possible use sensors already installed by the equipment manufacturer.
  • Select sensors that provide the needed level of accuracy while minimizing cost.
  • Be able to handle insufficient data and uncertain situations.

Relevance:
Poorly maintained, degraded, and improperly controlled equipment wastes an estimated 15-30% of energy used in commercial buildings. The proposal aims at leveraging the untapped capabilities of modern building automation and control systems by developing embedded FDD tools that monitor the performance of subsystems and automatically detect faults. Advances in building automation technologies provide the prospect of very data–rich performance surveillance in buildings that can be applied on system-wide scales that are necessary to optimize overall system performance.