Fault Detection & Diagnostics
Performance over time graph showing the impact of good maintenance.
Fault detection and diagnostics (FDD) is an area of investigation concerned with automating the processes of detecting faults with physical systems and diagnosing their causes. FDD tools distinguish themselves from analytic software and troubleshooting tools by automating the actual process of reaching conclusions from empirical data. They produce easily-used information from raw data, in contrast to traditional engineering tools, which usually convert data to more data. These tools can also be contrasted with tools that provide assistance in diagnosis, for example, by plotting data in various ways so performance problems can be detected and diagnosed by a knowledgeable expert. The availability of such expertise is limited and applying it manually takes considerable time. Automation reduces the time required for performance-problem detection and the associated costs. Coupled with methods for automated correction of problems (such as sensor drift, out-of-tune controllers, incorrect control algorithms and improper set points), automated fault detection and diagnosis will provide the basis for automated continuous monitoring and commissioning in the future, increasing the operating efficiency and performance levels of building systems and equipment.
PNNL has been involved for over 15 years in developing automated fault detection and diagnostics for building systems. The accomplishments of that work and descriptions of on-going projects are provided on this web page. Use the links that follow to connect to detailed information.