Carbon Steel Erosion Corrosion Laws In Sour Water System And Prediction Method Research

Sour water system erosion corrosion in oil refining unit is a common failure style with thedeveloping tendency of more inferior raw oil, large-scale equipment and rigorous operationcondition. Especially with the increasing proportion of high sulfur, high acid, chlorine-containinginferior raw oil in processing, corrosion failure accidents caused by sour water erosion are moresevere, which is a serious threat to the device running for a long period. Contrapose the frequentfailure phenomenon of sour water system, how to identify its inner failure mechanism and form amethod for quantitative prediction is always one focus of academia and engineering.An erosion failure case of air cooler outlet piping at sour water stripping overhead wasselected as the object in this paper, to study the erosion failure mechanism with morphology offailure, the multiphase flow physical parameters, corrosion medium equilibrium distribution andthe multiphase flow state by dissection analysis, process simulation and flow state calculations.Typical pipe material carbon steel erosion law in sour water system was affected by NH4HSmass percent concentration, temperature, H2S partial pressure, chlorine ion content and shearstress, which were explored by using rotary erosion experiment device and simulating corrosivemedium based on sour water composition. Carbon steel erosion law was studied combiningerosion electrochemical characteristics and corrosion product film properties by tafelpolarization curve and scanning electron microscopy analysis. Finally, the erosion corrosiondatabase of carbon steel corrosion rate change with multiple parameters was formed.Establishing BP neural network prediction model based on the database, to achieve the accurateprediction of carbon steel corrosion rate in sour water system. The research results were showedas follows:1) The source of tube wall thinning quickly was that the NH4HS concentration washigh and the mean flow velocity was quick;2) Severe corrosion locating at the upstream insideof elbow and extrados of elbow was due to bias flow caused by unbalanced outlet piping andcorrosion pits on the elbow inside wall was formed because of liquid membrane discontinuity.3)Erosion rate was controlled by the interaction of corrosion reaction rate and the diffusion rate ofcorrosion medium;4) The film was divorced from the matrix and became non-protective whenthe shear stress was over some value because the corrosion product film was loose, porous at thesurface of film and poor adhesive;5) The mean error was within8.63%while the prediction model was used for predicting the existing experimental data, thus, the model had a goodprecision and can be applied in the prediction of carbon steel corrosion rate under otherconditions.Innovation points of the article were summarized as follows:1) The failure mechanism oferosion corrosion was analyzed synthetically combining the dissection analysis, processsimulation and calculation of flow state;2) Combined the study of corrosion product filmproperty and erosion property, inner erosion corrosion mechanism under different influencefactors were analyzed;3) Based on BP neural network, corrosion prediction model in sour watersystem was established.