柴油机润滑油在使用过程中, 会受到水份、 柴油或冷却液等的污染, 受高温或燃烧室气体的作用还会产生氧化、 硝化、 磺化等产物, 严重时, 这些污染和产物可导致装备的失效。 在用润滑油分析能可靠地反映其工作状态, 从而确定最佳换油时机和故障源, 以避免柴油机在使用过程中产生异常磨损及腐蚀等问题。 目前的传统方法, 是用卡尔菲休法测定石油产品中水份, 气相色谱法测定柴油机油中的柴油稀释剂, 因分析时间长和费用高等因素影响其普及应用。 傅里叶红外(FTIR)光谱法从分子水平分析润滑油, 从而能更有效地监测在用油, 由于光谱的复杂性, 目前尚不普及。 通过在实验室条件下配置多种污染程度的油液样, 分别为润滑油含水份百分比浓度为0.11%, 0.22%, 0.44%和0.88%的油样; 润滑油在230 ℃下的氧化时间为299, 323, 371和395 h的油样; 润滑油被柴油稀释的质量百分比浓度为1.5%, 3%, 6%和12%的油样; 润滑油含乙二醇质量百分比浓度为0.1%, 0.2%, 0.4%和0.8%的油样, 包括配置的润滑油油样共计20个。 实验采用安捷伦Agilent Cary 630红外光谱仪, 使用100 μm厚的液体池, 防潮ZeSe型配置的光谱范围为4 000～650 cm-1。 分析了所有油样, 获得了FTIR光谱图, 确定了水、 氧化产物、 柴油、 乙二醇污染所对应的特征波段范围分别为3 150~3 500, 1 670~1 800, 745~755和1 030~1 100 cm-1; 监测参数包括特征波段的实测中心点、 左边界、 右边界、 左基线、 右基线等。 建立了一种在用润滑油污染物定量分析模型, 通过水份、 柴油、 乙二醇的污染比例和峰面积, 在230 ℃下的氧化时间与峰面积分别建立拟合方程, 水、 柴油、 乙二醇含量百分比与对应峰面积的相关系数分别为0.977 9, 1.000 0和0.989 5, 氧化时间与对应峰面积的相关系数为0.999 6。 分析计算预测值与实际值的相对误差可知: 水份和乙二醇在含量大于0.2%时, 预测的最大相对误差为10%; 氧化时间和柴油预测值的最大相对误差为1%。 通过按比例稀释, 可对在用油进行监测, 对3个日常工作的实际油样进行FTIR光谱分析。 结果表明: 一个油样的水份含量为0.38%, 已经超过标准; 一个油样的燃油稀释为19%, 已经超标; 另外一个油样正常。 水份超过标准门阈值的油样经卡尔菲休法测量, 其相对误差为4.6%。 燃油稀释超过标准的油样经黏度测量, 其相对值变化同样超过了标准要求, 即在判断润滑油是否需要换油方面是一致的。 利用FTIR光谱法分析在用润滑油, 选定适当的吸收峰, 并计算吸收峰的面积, 借助于已经建立的拟合公式, 可快速可靠地监测在用油的污染种类及程度, 该方法可在一定程度上满足工程需要。

Marine diesel engine oil in service will be contaminated by water, fuels, coolants, and other degradation products such as oxidation, nitrification and sulfonation, which can also be produced under the influence of high temperature or combustion chamber gas. In serious cases, these contaminants and products can lead to the failure of equipment. Analyzing the used oils can reveal its condition, therefore determine the optimum oil changed period and fault source, and avoid the abnormal wear and corrosion of diesel engine during operation. At present, the traditional methods are as follows that the coulometric Karl Fischer titration be used to measure the water content in oils and gas chromatography to determine the fuel dilution. Due to long time and high cost for oil analysis, these two methods mentioned above are not widely used in monitoring oils. For analyzing the lubricating oil at the molecular level, Fourier infrared (FT-IR) spectroscopy can be used to monitor oil more effectively. Unfortunately, it is not yet widely used at present because of the complexity of the spectrum. The oil samples sum up to 20, which are contaminated by water, fuel, ethylene glycol coolant, high temperature oxidation as well as new oil. Water concentrations in oil are 0.11%, 0.22%, 0.44% and 0.88%. The durations of high temperature for oil oxidation are 299, 323, 371 and 395 h. The fuel dilutions of oil are 1.5%, 3%, 6% and 12%. Ethylene glycol coolant concentrations in oil are 0.1%, 0.2%, 0.4% and 0.8%. The oil samples were analyzed by the Agilent Cary 630 FT-IR with a factory set pathlength of 100 microns and with the spectral range of 4 000～650 cm-1. FT-IR spectra of all oil samples were obtained with FT-IR spectrometer. It is determined that the corresponding characteristic band ranges of water, oxidizing products, fuel dilutions and ethylene glycol coolant respectively are 3 150~3 500, 1 670~1 800, 745~755, 1 030~1 100 cm-1. Monitoring parameters of FT-IR spectra include center point, left boundary, right boundary, left baseline and right baseline. A quantitative analysis model for contaminants in used oil was established. The fitting equations are about the concentrations of contaminants and the peak area of FT-IR spectra. Correlation coefficients between the contaminants of water, fuel dilutions and ethylene glycol coolant and the corresponding peak area of FT-IR spectra are 0.977 9, 1.000 0 and 0.989 5, respectively. The correlation coefficient between oxidation time and the corresponding peak area of FT-IR spectra is 0.999 6. The maximum relative errors between the predicted from the fitting equations and the actual value are 10% for the content of water and glycol more than 0.2%, and are 1% for oxidation time and fuel dilution. By new oil proportional dilution, 3 routine used oil samples were monitored with FT-IR spectrum analysis. The results showed that the water content of one oil sample was 0.38%, which exceeded the standard; the another oil sample was diluted to 19%, which exceeded the standard, and the other oil sample was normal. In the case of oil samples, which exceeded the threshold of the water, the relative error of the FT-IR measurement is 4.6% when compared with the Karl Fischer method. The fuel dilution of the oil sample, which exceeded the requirement of the standard, was verified by the change of viscosity. The FT-IR spectrum analysis and the change of viscosity are consistent in judging whether the used oil is or not to be changed. FT-IR spectroscopy was used to analyze the lubricating oil in use, the peak absorbances was selected, and the area of peak absorbances was calculated. By means of the established fitting formula, the contaminants types and degrees of oil can be monitored quickly and reliably. FT-IR spectroscopy method can meet the engineering requirements of used oil monitoring to a certain extent.