摘要 目前螺旋断层放疗设备先兆故障诊断模型,未曾分析螺旋断层放疗设备先兆故障信号,导致故障诊断模型诊断螺旋断层放疗设备效率低,误差率高,为此提出基于神经网络的螺旋断层放疗设备先兆故障诊断模型。根据螺旋断层放疗设备,时域和频域之间变换的桥梁,分析螺旋断层放疗设备先兆故障信号,利用神经网络,对故障信息的学习、记忆、识别和推理能力,建立设备先兆故障诊断模型,诊断螺旋断层放疗设备先兆故障。实验结果表明:确定螺旋断层放疗设备结构、模型运行环境、训练样本、故障诊断指标计算公式和设备故障位置,改变迭代次数和测试样本数据数目,对比三组模型,对螺旋断层放疗设备先兆故障诊断效果、诊断效率和诊断指标,此次研究的螺旋断层放疗设备先兆故障诊断模型,诊断螺旋断层放疗设备,具有较快的故障诊断效率,较低的故障诊断误差率,以及较高的召回率。 At present,the precursory fault diagnosis model of spiral tomotherapy equipment has not analyzed the precursory fault signal of spiral tomotherapy equipment,which leads to the fault diagnosis model.The diagnosis efficiency of spiral tomotherapy equipment is low and the error rate is high.Therefore,a neural network based precursory fault diagnosis model for spiral tomotherapy equipment is proposed.According to the bridge between time domain and frequency domain of spiral tomotherapy equipment,the precursor fault signal of spiral tomotherapy equipment is analyzed.Neural network is used to learn,memorize,identify and infer the fault information,and a diagnosis model is established to diagnose the precursory fault of spiral tomotherapy equipment.The experimental results show that:determine the structure of spiral tomotherapy equipment,model operating environment,training samples,fault diagnosis index calculation formula and equipment fault location,change the iteration times and the number of test sample data,and compare the three groups of models to diagnose the precursory failure of spiral tomotherapy equipment,diagnostic efficiency and diagnostic indicators,and the precursor of this study Fault diagnosis model,diagnosis of spiral tomotherapy equipment,has fast fault diagnosis efficiency,low fault diagnosis error rate,and high recall rate.
机构地区 唐山市人民医院
出处 《自动化与仪器仪表》 2021年第4期56-59,63,共5页 Automation & Instrumentation
基金 河北省卫生和计划生育委员会科研基金项目+AKR1C3、CTNNB1及LEF1与肿瘤放射抵抗的关系(No.20181219)。
关键词 神经网络 螺旋断层 放疗设备 先兆故障 诊断模型 neural network spiral tomography radiotherapy equipment precursory failure diagnostic model
