论文部分内容阅读
针对基本果蝇优化算法(FOA)寻优精度不高和易陷入局部最优的缺点,融合混沌算法对果蝇优化算法的进化机制进行改进,提出混沌果蝇优化算法(CFOA)。将CFOA算法应用于最小二乘支持向量机(LSSVM)惩罚因子和函数参数的选择中,可以改善参数选择的随机性和盲目性,从而建立基于CFOA-LSSVM的故障模式预测模型。应用该模型对变压器油中溶解气体故障模式进行预测,结果表明,CFOA方法在在收敛速度、收敛可靠性及收敛精度上均比基本FOA有较大的提高,依此而建立的CFOA-LSSVM故障模式预测模型具有较高的准确率。
In order to improve the evolutionary mechanism of Drosophila optimization algorithm, the Chaos Drosophila Optimization Algorithm (CFOA) is proposed to solve the shortcomings of the low accuracy of the basic fruit fly optimization algorithm (FOA) and its vulnerability to local optimization. The CFOA algorithm is applied to the selection of penalty factors and function parameters of least square support vector machine (LSSVM), which can improve the randomness and blindness of parameter selection, and establish the failure mode prediction model based on CFOA-LSSVM. The model was used to predict the failure modes of dissolved gas in transformer oil. The results show that the CFOA method has a larger improvement in convergence rate, convergence reliability and convergence accuracy than the basic FOA. The CFOA-LSSVM failure The model prediction model has higher accuracy.