Sensors and Fault Detection

The potential for connected lighting systems to easily incorporate sensors that monitor and report a wide range of environmental conditions has been widely touted, along with myriad data-driven use cases. However, in practice, the value of these use cases has been hard to quantify, and as a result, the cost of incorporating sensors has been hard to justify for most end-users. Similar to the case with energy data, the accuracy of the data reported by other types of sensors is an important consideration for use cases that rely on the data, and data reporting accuracy often is not specified or not well-specified. PNNL research on indoor occupancy sensors, environmental sensors, and automated fault detection and diagnostics is highlighted below.

Indoor Occupancy Sensor Performance. PNNL conducted a literature review exploring test methods for characterizing the performance of new and novel indoor occupancy sensors that might facilitate significant energy savings in buildings while minimizing occupant dissatisfaction. This review informed development of a new test method that leverages the best ideas discovered via the literature review. 

• A Review of Existing Test Methods for Occupancy Sensors (Report, August 2020)
• A Review of Existing Test Methods for Occupancy Sensors (Video with transcript, February 2021)
• Research Moves Towards Better Occupancy Sensors (Article, October 2020)

Environmental Sensors. Another PNNL study describes best practices for evaluating and calibrating environmental sensors and provides a related specification template that can be used when requesting calibration services.

• Specifying Calibration of Environmental Sensors (Report and Specification Template, September 2020)

Automated Fault Detection and Diagnostics. PNNL research in this area has evolved to focus on lighting issues associated with undervoltage conditions and associated faults, as well as background research on the use of data to support automated fault detection, diagnostics (AFDD) and prediction for lighting and other electrical systems. Researchers are also conducting a literature review on the use of machine learning techniques to perform AFDD for lighting and other electrical systems.

• Exploring the susceptibility of LED streetlights to power quality issues (Video with transcript, February 2021)
• Streetlight Faults Associated with Undervoltage Conditions and Methods for Automated Fault Detection (Report, in progress)
• On the Way: Automated Fault Detection and Diagnostics for LED Streetlighting Systems (Article, June 2023)
• A Review of Fault Detection, Diagnostics, and Prediction for Outdoor Lighting Systems and Associated Electrical Infrastructure (Literature review, in progress)