The state of the art in PV system monitoring is relatively simplistic, relying either on comparisons of outputs between various parts of the system (e.g., inverters) or on an evaluation of a performance metric that normalizes output to available irradiance and other environmental conditions. However, neither of these methods is very effective in discovering the source of identified problems or in identifying component-level failures, especially if they occur at the module or string level and thus have only a small, proportional effect on system output at the inverter or plant energy meter: hence the need for new system-monitoring methods.
Maximizing a PV system’s power production is critical, because nearly all of the investment is made before system activation. In their article in PV Power Tech: PV Power Plant Technology and Business Vol. 2, Sandia scientist Joshua Stein (in Sandia’s Photovoltaic & Distributed Systems Integration Dept.) and collaborator Mike Green from M.G. Lighting Electrical Engineering explore why PV power-plant monitoring—to identify performance and safety problems early—commonly fails to pick up on the very problems it is supposed to detect.
PV solar energy technology is usually static in nature. Aside from fans used to cool inverters, and when tracking systems are used, there are no moving parts, and these systems run cool and quiet compared with conventional energy-producing systems. It is not surprising, then, that PV systems tend to suffer from a lack of monitoring, because, in principle, no danger is involved, no serious safety issues exist, and monitoring is easily overlooked. However, PV systems have high-energy parts counts and are characterized by numerous identical pieces (e.g., modules). Even low failure rates for individual components are more likely to occur when there are many components, and the failure of one component can put more stress on other components and thus lead to cascading failures.
The future is ripe for innovative PV monitoring. The authors believe that monitoring systems of the future will be able to collect data from inexpensive sensors and use it in conjunction with data collected from other systems in the region, including weather and irradiance stations and/or satellite data, to determine the health status of a PV system. In addition, if multiple data streams from inexpensive sensors can be automatically collected at a low cost without disrupting PV generation, such information would be invaluable for detecting module degradation, locating system faults, and providing diagnostic information for O&M activities, including commissioning. In other words, be on the lookout for new monitoring products and services in the near future.