Photovoltaic system fault detection and diagnostics using Laterally Primed Adaptive Resonance Theory neural network

Cost effective integration of solar photovoltaic (PV) systems requires increased reliability. This can be achieved with a robust fault detection and diagnostic (FDD) tool that automatically discovers faults. This paper introduces the Laterally Primed Adaptive Resonance Theory (LAPART) artificial neural network to perform this task. The present work tested the algorithm on actual and synthetic data to assess its potential for wide spread implementation. The tests were conducted on a PV system located in Albuquerque, New Mexico. The system was composed of 14 modules arranged in a configuration that produced a maximum power of 3.7kW. The LAPART algorithm learned system behavior quickly, and detected module level faults with minimal error.

[1]  S. Sathiya Keerthi,et al.  Evaluation of simple performance measures for tuning SVM hyperparameters , 2003, Neurocomputing.

[2]  Thomas P. Caudell,et al.  Acquiring rule sets as a product of learning in a logical neural architecture , 1997, IEEE Trans. Neural Networks.

[3]  J. Johnson,et al.  Photovoltaic prognostics and heath management using learning algorithms , 2012, 2012 38th IEEE Photovoltaic Specialists Conference.

[4]  Jay Johnson,et al.  PV faults: Overview, modeling, prevention and detection techniques , 2013, 2013 IEEE 14th Workshop on Control and Modeling for Power Electronics (COMPEL).

[5]  Stephen Grossberg,et al.  Fuzzy ART: Fast stable learning and categorization of analog patterns by an adaptive resonance system , 1991, Neural Networks.

[6]  Ganesh K. Venayagamoorthy,et al.  Comparison of a recurrent neural network PV system model with a traditional component-based PV system model , 2011, 2011 37th IEEE Photovoltaic Specialists Conference.

[7]  E. E. van Dyk,et al.  Assessing the reliability and degradation of photovoltaic module performance parameters , 2004, IEEE Transactions on Reliability.

[8]  Detlev Heinemann,et al.  Monitoring and remote failure detection of grid-connected PV systems based on satellite observations , 2007 .

[9]  Kevin P. Murphy,et al.  Machine learning - a probabilistic perspective , 2012, Adaptive computation and machine learning series.

[10]  Abraham Ellis,et al.  Low-cost solar micro-forecasts for PV smoothing , 2013, 2013 1st IEEE Conference on Technologies for Sustainability (SusTech).

[11]  Aissa Chouder,et al.  Automatic supervision and fault detection of PV systems based on power losses analysis , 2010 .

[12]  Simon Haykin,et al.  Neural Networks: A Comprehensive Foundation , 1998 .

[13]  Thomas P. Caudell,et al.  A neural architecture for pattern sequence verification through inferencing , 1993, IEEE Trans. Neural Networks.

[14]  B. Lehman,et al.  Decision tree-based fault detection and classification in solar photovoltaic arrays , 2012, 2012 Twenty-Seventh Annual IEEE Applied Power Electronics Conference and Exposition (APEC).