A fault data analysis method for the landing gear based on association rule mining
J. Lai, Xiaobo Zhou, Shengjun Li · 2023
The landing gear system is a typical fault-prone system with a complex structure. By mining the Association Rules among fault diagnosis results and fault experimental parameters from the historical data of landing gears, it is possible for the rules to support fault location and provide the existing diagnosis database with knowledge supplements. Thus, an efficient method to mine association rules hidden in the landing gear fault data was proposed. Firstly, this paper constructed the organization mechanism of historical fault data and formally defined the target data set for data analysis. Secondly, to enhance the efficiency of data analysis, an improved Apriori algorithm was proposed by reducing the frequent item sets’ elements and the times of calculations to mine rules. Then, this paper used the evaluation index— fitness, which integrates confidence, completeness, and simplicity, to select association rules with higher prediction accuracy. The effectiveness of the proposed method was verified by experiments based on the landing gears’ historical fault data.