Investigation On Key Technologies Of Oil Analysis Spectrometer And Analysis Of Spectral Data

Using the type and content of wear particles in lubricating oil as an important index to evaluate the safety and health status of aviation equipment and to judge the wear parts,the reproducibility and repeatability of the detection are very important for the safe operation and maintenance of mechanical equipment.Therefore,the accurate detection of the content and type of wear elements in lubricating oil is of great significance for improving the operation safety of mechanical equipment and prolonging the service life.At the same time,in certain machinery operating places and occasions that are prone to cause major accidents,accurate detection of lubricating oil spectrum can not only be based on the type and content of detection elements of mechanical equipment health status of routine evaluation and detection,but also can predict the time period and fault occurrence in advance The parts are repaired and maintained immediately to avoid accidents..Most of the existing equipment for measuring abrasion particles in aviation oil is an oil analysis spectrometer,which mainly depends on imports,and the test results are not stable enough.In this paper,a self-developed spectrometer can be used to detect various metal or non-metallic wear elements in lubricating oil,and the results are accurate and reliable.The main contents of this paper are as follows:1.In view of the many problems that often occur in the excitation light source of traditional oil analysis spectrometers and improve the excitation intensity and stability of the light source,improvements to the light source are proposed.The IGBT device is used to replace the traditional auxiliary electrode.The improved light source excitation voltage is stable and reliable.The pulse peak voltage is adjustable at 11-15 k V,and the peak current is adjustable at 10-50 A.The DC spark light source excited by the power supply has good excitation conditions,full spectrum excitation and low overall calorific value.2.Combining with a large amount of disordered spectral data obtained by excitation,this paper proposes a method for normalizing the spectral intensity.The detection results are stable and reliable,and achieve the expected goal;the algorithm is tested with Cu and Mg as examples,and the maximum value of the relative standard deviation(RSD)of the test results is 6.21%.Test data shows that the use of intensity normalization and summation is beneficial to improve the detection accuracy.3.In view of the uncertainty of the spark excitation method,the fluctuation of the spectrum cannot be avoided during the actual detection process,so the detection results at different concentrations are also very different.In this paper,carbon is used as the internal standard element.Without the need to add additional internal standard elements,combined with related algorithms to process and analyze the detection data of the standard oil at different concentrations,the measurement errors caused by the fluctuation of the light source are reduced,and the reproducibility and repeatability of the test results are improved.4.Based on the BP neural network and MATLAB,the historical data of the two groups of engines were modeled and analyzed,and the concentration of iron was predicted as an example.First,establish a time series from the historical data of oil samples,and then use this time series as the training object to train the BP prediction model and predict the iron content in a period of time in the future.The prediction results show that the average error of both the predicted and actual values is less than15%,which can meet the need of forecasting the wear trend of aircraft engines.