基于WELCH算法集成学习模型的滚动轴承故障诊断#br#

针对滚动轴承振动信号具有非平稳性、非线性且易受背景噪声干扰，故障特征难以提取等问题，提出一种基于WELCH功率谱算法的集成学习模型的故障诊断方法。首先使用WELCH算法对轴承的原始振动信号进行预处理，从中提取峭度、偏度、波形因子、峰值因子、脉冲因子和裕度因子6 个参数，作为支持向量机的特征向量；然后结合集成学习算法构造Bagging-SVM集成学习模型。实验结果表明，与单一的SVM分类器相比较，Bagging-SVM 集成模型对于轴承的故障诊断性能更优；在不同电机转速下的轴承故障诊断中，诊断率分别为97 %，98 %，98 %和99.5 %；说明了该集成模型在不同工况下的适用性强，诊断性能优秀。

Traffic Noise Pollution and Its Prevention in Residential Area of Qingdao City

Traffic noise was measured in Chongqing South Road and East-West Motorway of Qingd ao. The results indicate that the noise seriously exceeds the critical value in the regions of 15 meters on both sides of Chongqing South Road and 25 meters on both sides of East-West Motorway respectively at night. To protect inhabitation environment, it is suggested that new technique should be utilized when the pavement were re-constructed, trees or shrubs should be planted to form green noise isolation belts, and sound barriers along the motorway should be constructed. In addition, soundproofing windows can be installed for inhabitants in the adjacent area of the motorway.