首页|基于ARIMA-GM(1,1)模型的温度二次仪表误差修正方法

基于ARIMA-GM(1,1)模型的温度二次仪表误差修正方法

扫码查看
针对化工设备的温度二次仪表在超高温阶段存在的误差无法检测的问题,文章提出了一种基于ARIMA-GM(1,1)模型的温度误差修正方法.首先,将温度二次仪表和标准器静置在稳定的环境中,通过标准器输出电信号来计算标准器与二次仪表之间的误差;其次,通过误差序列建立ARIMA模型,得到一组预测值;再次,计算实际误差与预测值之间的残差,为了提高ARIMA模型的预测精度,截取残差序列的尾部数据建立GM(1,1)修正模型;最后,将ARIMA模型与GM(1,1)模型相结合,建立了 ARIMA-GM(1,1)修正模型,预测下一个阶段仪表可能存在的误差,实现温度二次仪表的自修正功能.对比实验证明,本研究提出的ARIMA-GM(1,1)模型的相对平均误差最低,预测精度较高,能够较好地预测出下一个阶段仪表可能存在的误差,为后续仪表的自修正提供科学依据.
Error Morrection Method for Secondary Temperature Instruments Based on ARIMA-GM(1,1)Model
the problem of undetectable errors in secondary temperature instruments for chemical equip-ment during ultra-high temperature stages,this paper proposed a temperature error correction method based on ARIMA-GM(1,1)model.Firstly,put the secondary temperature instrument and standard in a stable environment,and calculate the error between the standard and secondary instrument by out-putting electrical signals from the standard instrument;Secondly,an ARIMA model was established through error sequences to obtain a set of predicted values;Then,the residual between the actual error and the predicted value is calculated,to improve the prediction accuracy of the ARIMA model,the tail data of the residual sequence is truncated to establish a GM(1,1)correction model;Finally,the ARI-MA model was combined with the GM(1,1)model to establish an ARIMA-GM(1,1)correction mod-el,which predicts the possible errors of the instrument in the next stage and realizes the self-correction function of the secondary temperature instrument.Comparative experiments have shown that the ARI-MA-GM(1,1)model proposed in this paper has the lowest relative average error,higher prediction ac-curacy,and can better predict the possible errors of the instrument in the next stage,and this method provided a scientific basis for the self-correction of subsequent instruments.

error correction modelARIMA modelGM(1,1)modelARIMA-GM(1,1)model

冯鑫、李舒、李磊、封海兵

展开 >

南京市计量监督检测院(南京210036)

误差修正模型 ARIMA模型 GM(1,1)模型 ARIMA-GM(1,1)模型

2024

滁州学院学报
滁州学院

滁州学院学报

影响因子:0.235
ISSN:1673-1794
年,卷(期):2024.26(5)