【摘要】 针对火炮供输弹系统故障信号所提取出的特征参量冗余复杂而导致故障类型不易识别的问题，设计一种基于蚁群优化支持向量机(ACO-SVM)的机械故障诊断方法。对信号进行特征提取得到信息熵，再采用ACO-SVM算法对其进行优化约简和故障诊断，并与支持向量机算法进行对比。结果表明：ACO-SVM算法具有更高的准确率，且准确率高达95%。更多还原

【Abstract】 A mechanical fault diagnosis method based on ACO-SVM is proposed to solve the difficulty in fault type identification caused by the complex redundancy of the characteristic parameters extracted from the fault signal of the ammunition ramming system. Signal feature extraction was conducted to obtain information entropy. ACO-SVM algorithm was used for optimization reduction and fault diagnosis, and compared with support vector machine algorithm. The results show that ACO-SVM algorithm has better accuracy with accuracy rate up to 95%.更多还原