随机共振模型结构参数自寻优方法与应用  

Parameter Self-Optimizing Method and Its Application of Stochastic Resonance Model

作　　者：明廷锋[1] 张永祥[1] 李婧[1] 

机构地区：[1]海军工程大学动力工程学院,武汉430033

出　　处：《天津大学学报（自然科学与工程技术版）》2014年第10期886-891,共6页Journal of Tianjin University：Science and Technology

基　　金：国防预研基金重点资助项目(9140A27020413JB11001);海军工程大学自然科学基金资助项目(HGDQNJJ12009)

摘　　要：在随机共振微弱周期信号检测过程中,如何确定结构参数值非常关键.已有的结构参数选择方法在应用上存在着局限性,如由于模型输入信号的干扰噪声未知,通常定义的信噪比无法获得.针对该问题,首先改写了随机共振模型的数学表达式;然后定义了一种随机共振模型输出信噪比,并给出了计算方法;最后在此基础上,以输出信噪比为评价指标,提出了一种结构参数自寻优方法,用于构建性能优良的随机共振模型.仿真信号分析表明该方法能较好地解决输入信号背景噪声未知的问题,所构建的随机共振模型可有效检测出低频率和高频率的微弱周期信号.通过对转子试验装置上的转子系统早期不平衡故障分析的应用,验证了所提出方法的有效性和实用性.Determining the value of the structure parameters plays a key role in the weak periodical signal detection by the stochastic resonance technology. The existing parameter-choosing methods show some limitations in applica-tion. For example,because the interference noise of the model input signal is unknown,the signal-to-noise ratio cannot be obtained. To solve this problem mentioned above,firstly,a new mathematical expression of the stochastic resonance model is proposed. Secondly,the signal-to-noise ratio of the model is redefined and its calculation process is provided. Finally,a structure parameter self-optimizing method using the output signal-to-noise ratio as the evalua-tion index is presented for designing the stochastic resonance model with excellent performance. The simulation data processing results indicate that the difficulty in obtaining the signal-to-noise ratio can be resolved by the method pro-posed. And the constructed model can be used to successfully extract the weak periodical signals with both the low and high frequency. Also the early imbalanced fault diagnosis in the rotor system demonstrates the effectiveness and prac-ticality of the presented self-optimizing method.

关 键 词：随机共振 结构参数寻优 微弱周期信号检测 故障诊断 

分 类 号：TP206[自动化与计算机技术—检测技术与自动化装置]