基于报警时序数据的水泥生产过程故障预警方法

Fault Warning Method for Cement Production Process Based on Alarm Time Series Data

实际的生产过程中报警系统往往仅用于给监控人员提供预警信息,缺少进一步的深入分析。在传统水泥生产过程中,会产生电流、温度、压力等报警时序数据,当某段工艺或设备发生异常报警时,这可能是由于设备故障、工艺问题、材料变化或其他不可预测的因素引起的。利用智能化数据检测手段,在生产数据发生跳变、斜率等报警时,将报警数据提取为报警时间序列,然后使用InceptionTime网络对报警时间序列进行水泥生产过程的故障预警,指引生产人员针对异常设备、工艺、材料进行及时分析和调整,有效防止水泥生产过程中的故障发生。

In actual production processes,the alarm system is often only used to provide warning information to monitoring personnel,lacking further in-depth analysis.In the traditional cement production process,alarm timing data such as current,temperature,and pressure will be generated.When an abnormal alarm occurs in a certain process or equipment,it may be caused by equipment failures,process issues,material changes,or other unpredictable factors.This paper uses intelligent data detection methods to extract alarm data into alarm time series when there are jumps,slopes,and other alarms in production data.Then,the InceptionTime network is used to provide fault warning for the cement production process using the alarm time series,guiding production personnel to analyze and adjust abnormal equipment,processes,and materials in a timely manner.