核电设备无线振动测量冗余滤波算法研究

Research on wireless vibration measurement redundancy filter arithmetic for equipment of NPPs

为解决在核电站环境下,设备运行状况的振动测量存在的传感器故障率增加,噪声更复杂,测量参数和结构更容易变化,安全性要求更高等问题;选择了冗余测量结构进行算法设计。分析不同冗余测量的结构特点,证明了基于多模块处理器的分布式冗余测量结构比一般的集中式冗余测量结构有更高的可靠性;在该结构下结合振动信号变化特征,设计一种基于FFT与互相关相融合的冗余滤波算法;为使该算法具有更好的适应性,利用聚类分析中的距离计算方法和基于特征量的多参数ARMA,增加算法的故障判断能力。算法在MATLAB下进行仿真,并在MC13224V系统处理器,ADXL345传感器,ADUC7060的模块处理器的实际平台下进行了验证;实验与仿真结果表示在有故障情况,系统测量不受影响;而且其测试的相对误差从无该算法的10%以上减低到3%左右。

In order to solve problems of vibration monitoring system that under NPPs environment sensors having more fault rate, test having more complicate noise, measurement parameter and construction being more uncertain, and system need more critical for safety, redundancy measurement mode is selected to design suitable arithmetic. Based on the analysis of characteristic of different redundancy construction, it is proved that the reliability of dis- tributed redundancy test configuration （having many module processors） is higher than that of concentrate redun- dancy test configuration. Under this test structure, and considering the variable characteristic of vibration signal, redundancy filter arithmetic based on fusions of FFI＂ and correlation is designed. In order to extern the application field, the distance calculation of clustering analysis and multi parameter ARMA are adopted to increase the ability of fault diagnosis of this arithmetic. All the arithmetic is simulated by MATLAB. Meanwhile, it is test under the re- al hardware platform made by MC13224V system processor, ADXL345 MEMS acceleration sensor, ADUC7060 module processor. The simulation and test results prove that the faults of system have no influence with the whole system test, and the relative error of test reduces from 10% to lower than 3%