The Luliang Production System Corrosion Experimental Study Of Corrosion Rate Prediction Techniques

Accompaning with LuLiang oil field is developed continuously and it's comprehensive water content ratio rises year by year, there are comparatively more serious corrosion appearance in tubes, casings and surface pipelines of LuLiang oil field, corrosion becomes the significant hidden danger that effects safe production and economy of LuLiang oil field, so we should analyze carefully the corrosion mechanisms, corrosion actions and effect elements of tubes, casings and pipelines in LuLiang oil field product system, in the meantime we should take some anticorrosion measures and some corrosion prediction scheme, these works not only have the profound theory sense but also have wide-ranging pragmatic value.First of all, in the paper the author analyzed carefully the corrosions and corrosive environment in the oil field product system, mastered roundly it's ion composition of output water, the content of corrosive gas in output gas , the temperature and pressure of oil well in the bottom , corrosive types and corrosive mechanisms, these supply the powerful base support for the follow-up work.Secondly, by using weight-loss method of steel slice, the author study some factors that affect corrosion rate of N80 and 20# steel ,and then put forward some laws of the corrosion rate that is relate to temperature, pH, Cl^-, SO₄^2-, Ca²⁺, Mg²⁺ and HCO₃^-, subsequently by using dynamic corrosive experimental data and grey system and mathematical statistics theory, some corrosive factors are sorted and the main factors are found.In the paper, by the HITACHIS-450 scaning electronic microscope, the structure of the N80 oil tube steel and 20# low carbon steel is analyzed. The releration between the structure of corrosion layer and the corrosion of N80 steel and 20# low carbon steel is analyzed too.The prediction of corrosion rate in oil field product system is always a hot problem which oil scientific and technical personnel comparatively concerned, by synthetically comparing some kinds of prediction method ,finally genetic algorithm and neural network is adopted to predict the corrosion rate of N80 and 20# steel ,in this process genetic algorithm optimized initial stage weight and valve value and corrosion rate is approached by using BP net, in the end the corrosion prediction software is programed, examined with some examples, the software has high prediction precision, so it can be widely used widely in LuLiang oilfield product system.