基于EMD和分形盒维数的旋转机械耦合故障诊断方法研究

针对旋转机械耦合故障的诊断问题，提出一种基于EMD（Empirical Mode Decomposition）和分形盒维数的诊断方法。该方法结合EMD对非线性信号处理的自适应性和分形盒维数能对非线性行为定量描述的特点，先对故障信号进行EMD处理，得到含有故障特征的本征模式函数(Intrinsic Mode Function,简称IMF)，然后求出各IMF的盒维数，通过盒维数的比较分析进行故障诊断。构造了含有裂纹-碰摩-松动耦合故障的转子-轴承系统动力学模型，用龙格库塔法求出故障模型振动信号。通过对耦合故障信号进行分析，得到耦合故障特征向量，并与传统的边界谱诊断方法比较，证明该方法对旋转机械耦合故障诊断的有效性和优越性。

Study on coupling faults of rotary machinery diagnosis method based EMD and fractal box dimension

Aiming at coupling faults diagnosis of rotary machinery, based on EMD (Empirical Mode Decomposition) and fractal box dimension method of faults diagnosis is proposed. With the method combining adaptability of EMD processing nonlinear signal and characteristic that fractal box dimension can quantificationally describe nonlinear behaviors, the method through the first to process fault signal by EMD and obtain Intrinsic Mode Function (IMF) containing fault characteristic, than to solve box dimension of IMF and to diagnose fault through comparing and analyzing the box dimension. A dynamic model of the rotor -bearing system is established consider in coupling fault of crack rub-impact and pedestal looseness, vibration signal of fault model is obtained with Runge-Kutta method. Through to analyze coupling fault signal, getting coupling fault feature vector, meanwhile, comparing with the traditional boundary spectrum diagnosis method, proving the method for coupling fault diagnosis of rotating machinery validity and advantage.