Grey predication of corrosion on oil atmospheric distillation equipment

Shengli refinery plant process Middle-east crude oil, the equipment is exposed to wash erosion seriously. The elbow on the entrance of the atmospheric distillation tower top exchangers always leak for eroding. Measured the thickness of elbow by ultrasonic thickness meter every month, established GM(1,1) model with grey system theory, the average relative error is 1.8959%, model passed error check, is eligible. Adding a number 100 to the thickness of the elbow, then subtract it from the result, established improved GM (1, 1) model, and the average relative error is 1.3084%, increased prediction precision. Used full data to create improved GM(1, 1) model to predict elbow´s life. Predicted the thickness of elbow in the future, the elbow life can reach to 24 month, but can not reach to 36 month. It shows that the grey system model is feasible; the improved full data GM (1, 1) model can be used for long-term forecasting and is more accuracy.