红外光谱评价内燃机油抗氧化性能的研究

红外光谱快速检测石油产品性能是近年来发展的新技术,目前国内外在该领域的研究仅限于测试燃料油性能,由于润滑油组成、结构复杂,红外光谱技术测试润滑油性能的研究还未见报道.文章研究了润滑油组成、结构的红外光谱特征,提出了根据内燃机油组成、结构对抗氧化性能的贡献来提取其光谱信息的技术路线.结合BP神经网络和自组织神经网络的优点,发展了量化自组织神经网络数学模型,该数学模型具有自组织神经网络的定性聚类功能和BP神经网络的定量分析功能,与BP神经网络相比较,量化自组织神经网络具有更好的鲁棒性,测试结果优于BP神经网络,该论文的研究为润滑油性能的快速检测提供了一种新的技术手段.

Study on the Antioxidation Property of Engine Oil by Infrared Spectroscopy

The relationship between the engine oil's infrared spectroscopic characteristics and its antioxidation property was studied,and the technical method to obtain the spectral information in accordance with the contribution of lubricating-oil composition to antioxidation property is presented.Quantitative self-organization neural networks was developed by combining the advantages of BP neural networks and those of self-organization neural networks.Quantitative self-organization neural networks possesses both the qualitative clustering function of self-organization neural networks and the quantitative analysis function of BP neural networks,and features better robustness,so the quantitative analysis result of quantitative self-organization neural networks was better than that of BP neural networks.The result has provided a kind of new technology means for fast testing lubricating-oil performance.