The oil chromatographic data processing and analysis method based on the robustness model

In this paper, the analysis and test of dissolved gas in oil of oil-immersed type equipment, are proved to be difficult to avoid two kinds of errors, random error and non-random error. Aiming at the non-random error, a data processing and analysis method of oil chromatogram based on robust model is proposed in this paper. Firstly, the gross error or abnormal data that will distort the measurement results are discriminated and eliminated according to certain criteria. In order to avoid or suppress the influence of gross error, the improved least square method is proposed to deal with the gross error in the test data, which improves the fitting accuracy, and the balance criterion method is proposed in the data analysis process of the model,to conduct a comprehensive analysis of the data. Therefore, by using the robust estimation of the improved least square method, the singleness and subjectivity of the traditional method can be overcome, and the model is effective.