基于非线性滤波的智能故障诊断方法

Intelligent Fault Diagnosis Method based on Nonlinear Filtering

针对大型机械设备运行中存在的强扰动、多干扰噪声存在的问题,建立合理的故障混合智能诊断模型,确定不同模型的分界阈值,深入分析多噪声相关情况下噪声联合概率密度的数学分解表达式及噪声统计特性的分布函数,以高斯噪声为背景,推导最优建议分布函数的具体解析表达式,实现强扰动、多噪声干扰情况下故障诊断的精确、鲁棒滤波。

In view of the problems of strong disturbance and multi disturbance noise existing in the operation of large-scale mechanical equipment,a reasonable fault hybrid intelligent diagnosis model is established,the boundary thresholds of different models are determined,the mathematical decomposition expression of joint probability density of noise and the distribution function of noise statistical characteristics under the condition of multi noise correlation are analyzed in depth,and the optimal suggestions are derived based on Gaussian noise.The specific analytical expression of the distribution function can realize the accurate and robust filtering of fault diagnosis in the case of strong disturbance and multi noise disturbance.