| 摘要: |
| 提出一种基于近红外光谱技术的在用润滑油闪点快速检测方法。通过比较样品光谱和混入燃油光谱之间的光谱差异进行波段筛选,利用人工神经网络方法(ANN)和偏最小二乘方法(PLS)进行建模并比较,最终确定针对3个特征波段建立的ANN在用润滑油闪点的数学模型性能较优,模型的R2和SEP分别达到0.918 3、3.06 ℃。实验结果表明,ANN方法作为一种处理非线性问题的化学计量学方法,能较好地实现对在用润滑油的闪点测定。利用近红外光谱分析技术对快速判断润滑油是否混入轻质油料,为及时准确找到设备故障所在提供依据具有重要的指导意义。 |
| 关键词: 在用润滑油 闪点 近红外光谱 人工神经网络 |
| DOI：10.3969/j.issn.1007-3426.2013.05.018 |
| 分类号: |
| 基金项目: |
|
| Rapid quantification of flash point in using lubricant by near-infrared spectroscopy |
|
Xing Zhina, Wang Juxiang, Liu Jie, Shen Gang
|
|
(Department of Airborne Vehicle Engineering, Naval Aeronautical and Astronautical University, Yantai 264001, Shandong, China)
|
| Abstract: |
| A rapid NIR measurement was applied to flash point of using lubricant. The spectra band were chosen by comparing sample’s spectra and mixed fuel’s spectra. The different calibration models of flash point were built by the means of ANN and PLS respectively. Then these models were compared through model valid parameter. In result, the ANN model with three characteristic bands was optimal. Its R2 and SEP were 0.918 3,3.06 ℃ respectively. It was indicated that ANN method can determinate flash point of using lubricant as a non-linear chemometrics means, which could take place of traditional method for rapid measurement of diesel lubricant quality. That’s to say, NIR technology was an important technology to judge using lubricant whether mixed with light oil and provide instruction for confirming equipment failure in time. |
| Key words: using lubricant flash point near-infrared spectroscopy artificial neural network (ANN) |
