Implementation of the IDAIC framework on an air handling unit to transition to proactive maintenance

Abstract

Industry is stuck in reactive mode regarding the maintenance of their heating, ventilation and air conditioning (HVAC) systems, which leads to inefficient operation. To avoid wasteful use of natural resources a proactive approach is needed. To enable this transition combined knowledge-based and data-driven approaches such as the Industrial Data Analysis Improvement Cycle (IDAIC) framework have been proposed, but there is a lack of practical implementation in the literature. In this paper, we implement the IDAIC framework on an Air Handling Unit (AHU) in a real industrial facility to address the lack of practical and reproducible real world analysis. The lack of domain knowledge-driven data mining solutions is addressed primarily in two parts. Firstly, data and control issues are addressed to enable the application of knowledge-based techniques to detect faults that are easily identifiable. Secondly, data-mining based techniques are applied which focus on the detection of more subtle faults which are indicative of component degradation. This approach enabled the development of data-driven analysis that visualises the effect of the different modes of operation on the energy consumption of the AHU, which allow the onsite facilities team to make more informed decisions regarding the maintenance of this AHU.