Skip to main content
European Commission logo print header

GEneric diagNosis InstrUment for SOFC Systems

Article Category

Article available in the following languages:

Fuel cell defect diagnosis potential

Solid oxide fuel cells (SOFCs) are among the most promising fuel cell technologies with great promise for combined heat and power (CHP) systems. Scientists investigated approaches to online and offline defect detection to support commercialisation.

Energy icon Energy

Fuel cells are electrochemical conversion devices, converting the energy stored in chemical bonds into electrical energy, but without combustion. SOFCs can operate with a variety of fuels, and their high operating temperature produces hot exhaust that can be used in combined cycle operations such as CHP systems. One of the key challenges to commercial implementation in micro-CHP systems is the lack of SOFC diagnostics methodologies to reduce failures, prolong service lifetimes and decrease operating costs. A large European consortium launched the GENIUS (Generic diagnosis instrument for SOFC systems) project to investigate the feasibility of a generalised approach to fill the gap. With EU support, scientists sought to create a generic tool to diagnose the state of health of a SOFC using simply obtained measurements and their deviation from control parameters. Currently, the complexity of the SOFC system makes it difficult to test parameters associated with the stack's internal components. Further, the resulting models are too complicated for real-time use. Research partners compared two different diagnostic approaches. Both were designed to be compatible with both online and offline modes. In the first case, the main purpose is plant safety and compliance with regulations. Offline diagnostics are important in scheduling maintenance and repairs. Electrochemical impedance spectroscopy data from a real commercial SOFC system provided important input for development of the two modelling approaches. The first approach was based on analysis of residuals between experimentally measured values and those predicted. If the residual value exceeded a previously determined threshold, a fault was detected. The second approach relied on pattern recognition. Health indicators were extracted from measured signals and classified according to their mathematical behaviour or pattern of changes. Research highlighted the fact that each industrial partner had its own set of demands for such a tool, making it very difficult to create a generic application. In addition, its targeted usage for safety critical diagnoses requires significant additional research to determine thresholds, one reason that current diagnostic systems are quite conservative. Partners concluded that such a tool is quite valuable, but should be developed on an individual rather than generic basis.

Keywords

Fuel cell defect, SOFCs, defect detection, CHP systems, diagnostics

Discover other articles in the same domain of application