Maintenance of heating, ventilation and air conditioning (HVAC) systems in building portfolios becomes increasingly
challenging as systems become more complex, and as the number of systems increases across a managed portfolio. Data-driven maintenance approaches employ multiple data sources to analyze the system’s operation and maintenance
(O&M) status, and hence can effectively support decision making for complex systems’ maintenance. Automated fault
detection and diagnostics (FDD) tools are used to identify abnormal operations and resolve the types and locations of
problems in HVAC systems. Data generated by FDD tools contain essential information in terms of the system's
abnormal operation such as fault causes, fault location, fault occurrence, and duration. Therefore, the integration of FDD
tools’ output data into data-driven maintenance tools can significantly support the maintenance decision-making
procedure, and streamline HVAC system’s O&M processes. However, the semantic heterogeneity and the structural
heterogeneity in FDD data lower data interpretability and interoperability, and hence hinder the integration of the data
by other maintenance tools. In this paper, we propose a framework to organize and integrate FDD data, so that the
data can be efficiently queried by or integrated into other maintenance tools. The framework includes the FDD data
model, the fault taxonomy library, and organized FDD data structure. The case study demonstrates that the FDD data
reorganized under the framework can be efficiently analyzed to assist HVAC system maintenance.