Lead Performer: Lawrence Berkeley National Laboratory – Berkeley, CA
May 19, 2020
minute read time
Lead Performer: Lawrence Berkeley National Laboratory – Berkeley, CA
Partners:
-- Pacific Northwest National Laboratory – Richland, WA
-- National Renewable Energy Laboratory – Golden, CO
-- Oak Ridge National Laboratory – Oak Ridge, TN
-- Drexel University – Philadelphia, Pennsylvania
DOE Total Funding: $1,506,000
Project Term: October 1, 2019 – September 30, 2022
Funding Type: Lab Award
Project Objective
Algorithms developed to perform automated fault detection and diagnostics (FDD) use building operational data to identify the presence of faults and (in some cases) isolate their root causes. As buildings become more data rich, and as data science comes to buildings, FDD is of increasing relevance to the building community. A persistent challenge has been the lack of common datasets and test methods to benchmark the performance accuracy of FDD methods, and gauge improvement over time. On the other hand, users of FDD technology commonly report a gap in robust techniques to manage and prioritize the large number of faults that are flagged for resolution in their portfolios. When confronted with dozens to hundreds of faults and constrained operation and maintenance (O&M) resources, balancing considerations of energy, cost, occupant comfort and equipment health is needed in deciding when and where to act first.
The project team is working to create the world’s largest publicly available dataset with verified ground truth information on the presence and severity of faults. The data will comprise time series of HVAC operational data (e.g. temperatures, pressures, control signals, component status, etc.) under a diversity of operating conditions, combined with information on the presence and absence of faults and their associated intensity. The dataset will span a wide range of commercial building HVAC systems and configurations. The team at National Renewable Energy Laboratory (NREL) will leverage Modelica to develop simulation-based RTU fault data.
The secondary outcome of this work comprises best practice heuristic guidance for FDD users to prioritize faults when confronted with hundreds of faults and constrained resources. This guidance will be paired with a detailed set of recommendations for out-year R&D to develop analytical prioritization solutions.
Project Impact
In commercial buildings, it is estimated that annual building energy use can be cut by an average of 29%, or approximately 4-5% of overall national energy consumption, through corrections to existing buildings controls infrastructure and resulting improvements to operating efficiency. For heating, ventilation, and air conditioning (HVAC) systems, for example, BTO aims to achieve 20% end-use energy savings through research and development of improved automated FDD solutions. Outside of the research community, building owners and operators at the leading edge of technology adoption are using automated FDD to enable median whole-building portfolio savings of 7%. A diversity of techniques is used for FDD spanning physical models, black box, and rule-based approaches, and researchers continuously strive to develop new and better algorithms.
The test dataset can be used by FDD developers, FDD users, and research funders to compare and contrast performance accuracy across FDD algorithms; identify performance gaps to focus future development efforts and resource investment; and develop an understanding of how FDD technology overall is improving over time.
Contacts
DOE Technology Manager: Erika Gupta
Lead Performer: Jessica Granderson, Lawrence Berkeley National Laboratory