Research on fast control of distributed photovoltaic countercurrent based on multidimensional data mining

Abstract

To speed up the control of distributed photovoltaic countercurrent prevention, a fast control method of distributed photovoltaic countercurrent prevention based on multidimensional data mining is investigated. This approach uses a multidimensional photovoltaic output power data mining method based on association rule mining. After defining the support and confidence of multi-dimensional photovoltaic output power data mining, an FP (Frequent Pattern) tree is constructed on the basis of Apriori algorithm to generate frequent itemsets. Partial redundant nodes of frequent itemset association rules are eliminated through pruning, and the final reserved node data serve as the result of multidimensional photovoltaic output power data mining. Combined with the multi-dimensional photovoltaic output power data mined, the anti-reverse current cont rol action is rapidly started by the anti reverse current control method based on the anti reverse current safety automatic control device, combined with the difference between the photovoltaic output power data and the fixed value status. Experimental testing demonstrates that this method requires no >1.5 s to control the counter current of the distributed photovoltaic system, exhibiting efficient countercurrent control capabilities for distributed photovoltaic systems.