状态监测数据流时间乱序问题建模与研究

Modeling on time out-of-order problem of condition monitoring data stream

复杂装备的状态监测数据在传输和处理过程中可能出现时序混乱,其结果会影响设备状态的监控和分析,针对这一问题进行了建模与研究。提取问题域中的主要元素进行分析建模。针对不同分析监控任务的要求和数据发送频率,结合对数据传输解析过程延时的概率密度拟合,确定对乱序数据重新排序所需的缓冲窗口的大小和数据处理策略,以满足不同分析监测任务对时序容错率和数据实时性的要求。对模型中的参数进行了计算和分析,为设计满足指定时序容错率和实时性要求的监测分析系统提供了理论依据。通过仿真实验,验证了模型的有效性。

The time out-of-order problem may appeared m the process ofcondition monitoring data＇s transmission andmanipulation, which could affect the equipment condition＇s monitoring and analysis. To solve this problem, a series of key elements were extracted for modeling. On this basis, according to different requirements of analysis tasks and different data transmission frequencies, the data processing strategies and the size of sliding window for sorting were determined to satisfy different demands of analysis tasks on time-out-of order fault tolerance and data real-time per formance by combining with probability density fitting of data transmission delay. The parameters in the model were calculated and analyzed, which provided a theoretical basis for designing monitoring and analysis system to meet specified requirements of fault tolerance and real-time performance. Simulation experiments were conducted to vali date the effectiveness of the proposed model.