Orthonormal Basis Filters for gas turbine fault diagnostics system design: A review

Gas turbines have become the dominant technology for power generation. They can be quickly assembled and put to service. They are convenient for engine exchange during system overhaul. The emission of NOx, SOx, CO, and particulates are also significantly law as compared to coal fired power plants. However, their maintenance cost is relatively high. The perceived best approach to reduce the cost is by using a proactive maintenance strategy in which a real-time diagnostics system plays a key role. The purpose of this paper is to review application of Orthonormal Basis Filters (OBFs) to fault detection and diagnostic systems design. The types of OBFs studied include Laguerre filters, Meixner filters, Kaurtz filters, Generalized OBF, and Markov-OBF. The combination of OBFs and computational intelligence methods (artificial neural network, fuzzy systems, and evolutionary optimization) are also highlighted. The review shows that, even though OBFs have been around for more than a decade, their application is limited to model identification only. As such, the only diagnostic problem revealed so far is that concentrating on stirred tank reactor. Therefore, to extend the use of OBFs to power plants, there needs to be further study in the context of power plants or specifically gas turbines. © 2006-2016 Asian Research Publishing Network (ARPN).