基于二维查找模型和ZigBee网络的CT扫描设备故障远程检测方法研究

Design of remote fault detection method for CT scanning equipment based on two-dimensional search model and ZigBee network

目的:设计基于二维查找模型和ZigBee网络的CT扫描设备故障远程检测方法,提高CT扫描设备故障检测准确率。方法:采用ZigBee网络技术从环境因素、电气因素和气路因素采集CT扫描设备运行状态数据,根据监测数据提取其中的时频域信号,将提取的故障信号特征参数进行融合,采用二进制反向变换将故障数据转化为平稳的时间序列,抑制数据的波动性,并通过建立相应的二维查找模型与匹配函数计算故障数据的依赖度,以此为依据,引入贝叶斯评价函数实现对设备故障的远程检测。选取医院临床在用的大型CT扫描仪为实验样机,同时随机选取使用与实验样机型号与配置相同的80台次CT扫描设备的550条实际故障数据为训练样本,对比CT扫描设备故障远程检测方法与长短时记忆网络(方法1)和深度信念网络(方法2)检测方法的准确率。结果:设计的CT扫描设备故障远程检测方法仅在状态标签5出现检测错误,但判断误差在可控范围。5次故障检测实验中,CT扫描设备故障远程检测方法得到的检测准确率均>80%,远高于方法1与方法2。结论:CT扫描设备故障检测方法具有较高的检测准确率,能够实现CT扫描设备故障远程检测的实际需求。

Objective:To design and construct a remote fault detection method for CT scanning equipment based on two-dimensional search model and ZigBee network to improve the accuracy of fault detection.Methods:ZigBee network technology was used to collect the operating status data of CT scanning equipment from three aspects:environmental factors,electrical factors and gas circuit factors.The time and frequency domain signals were extracted from the monitoring data,and the extracted fault signal feature parameters were fused.The fault data were transformed into stationary time series by inverse binary transformation to restrain the fluctuation of the data.The dependence degree of fault data was calculated by establishing corresponding two-dimensional search model and matching function.Based on this,Bayesian evaluation function was introduced to realize the remote detection of equipment fault.A large-scale clinical CT scanner in clinical use in the hospital was selected as the experimental prototype,and at the same time,550 real fault data generated from the 80 CT scanning scanners with the same model and configuration as the experimental prototype were randomly selected as training samples.The accuracy of CT scanning equipment faults remote detection method with that of the long short-term memory network(referred to as method 1)and deep belief network(referred to as method 2)were compared.Results:The designed remote fault detection method of CT scanning equipment only had detection error on status label 5,but the judgment error was in the controllable range.In the five fault detection experiments,the detection accuracy of the remote fault detection method of CT scanning equipment was more than 80%,which was much higher than that of method 1 and method 2.Conclusion:The fault detection method of CT scanning equipment has a high detection accuracy and can realize the practical requirement of remote detection of CT scanning equipment.