基于SVM的空调系统智能故障诊断研究

Intelligent Fault Diagnosis for Air-condition System Based on Support Vector Machine

空调系统作为武器装备管理系统的一部分,在国防工程中的作用至关重要,致使其结构和功能日益复杂,因此对自动检测提出了更高的要求。为了有效地提高空调系统故障诊断的效率和精度,本文提出了一种基于主成分分析(PCA)特征提取的支持向量机(SVM)混合诊断模型。该模型首先进行归一化预处理,消除奇异数据;然后利用主成分分析对数据属性进行简约,消除冗余信息并进行特征提取,支持向量机进行故障诊断;最后采用网格搜索法和交叉验证法对SVM的惩罚函数和核函数参数进行寻优。通过实例验证表明,整个处理过程将诊断识别率从58.018 1%提高到了99.953 6%,能有效地进行空调系统的故障诊断和定位,并可实现故障的准确区分。

As a part of weapon material management system,air-condition system is becoming more and more important to national defense engineering, resulting in the increasingly complex structure and function and putting forward higher requirements for automatic detection. For the purpose of effectively improving the efficiency and accuracy of fault diagnosis for air-condition system, this paper proposes a fault diagnosis model based on the support vector machine （SVM） of principal component analysis （PCA）. Firstly,the model normalizes data to eliminate abnormal data and applies principal component analysis to simple data attributes; then it eliminates redundant information for feature extraction and makes fault diagnosis by support vector machine; and at the end, it uses gridsearch method and cross validation method to optimize the parameters of the penalty function and kernel function of SVM. The results show that the diagnosis accuracy is greatly improved from 58. 0181% to 99. 9536%, and the faults can be effectively located and distinguished.