基于灰色组合模型的校准间隔优化仿真

针对测量仪器校准间隔的优化问题,分析了历史校准数据的特征,描述了校准数据动态发展的数学模型,提出一种灰色组合模型进行校准间隔的预测仿真。灰色组合模型的思想是:采用灰色GM(1,1)模型预测历史校准数据序列的趋势性成分,同时引入AR模型、BP人工神经网络模型和马尔可夫模型三者的组合模型预测随机性成分。仿真实验表明:灰色组合模型适应了校准数据小样本、非线性的特点,适合用于校准间隔的预测仿真。

Optimum Simulation of Calibration Interval Based on Grey Combined Model

For optimization of a measuring instrument's calibration interval, firstly, characters of historical calibration data were analyzed and the mathematic model of calibration data was described. Then a grey combined model was proposed to optimize calibration interval. Trend components of calibration data are predicted by GM (1, 1) model. Furthermore, random components are predicted by the combined model which is comprised by AR model, BP neural network model and Markov model. Experiments show the grey combined model adapts to small sample and non-linear characters of calibration data and it is fit for the prediction of calibration interval.