METHOD OF DIAGNOSING FAULTS IN PHOTOVOLTAIC ARRAYS, BASED ON AN IMPROVED RANDOM FOREST ALGORITHM

摘要

FIELD: measurement technology.;SUBSTANCE: invention relates to a method for diagnosing failures of a photovoltaic array based on an improved random forest method. Parameters of common main lines and each branch of photovoltaic array are used to reflect the operating state of each branch in photovoltaic array, and the voltage difference between the arrays between the different taps reflects the operating state of each photovoltaic battery unit in the branch, i.e. determining location of failure of photovoltaic array. At the same time, using unselected samples, the method is optimized and improved in three aspects: transfer of authority for voting of decision trees, processing of even count during voting and degree.;EFFECT: technical result is higher accuracy of fault diagnosis.;4 cl, 13 dwg, 8 tbl