Research On Corrosion Prediction Technology Of CO₂ Wet Gas Gathering And Transportation Pipeline Based On Random Uncertainty Method

The accurate prediction of internal corrosion is a long-term research topic to be solved in CO2-containing wet natural gas gathering pipelines.The on-line detection of internal corrosion is limited by the physical and geometric factors of the pipeline and the environment,and is difficult to implement in the gathering pipeline.However,the corrosion process in the corrosion prediction model of the classical CO2-containing wet natural gas gathering pipeline is complicated.Due to its idealized conditional assumptions and limited experimental conditions,the fitted model itself has strong aleatory uncertainty.The multiphase flow calculation of the parameters along the pipeline also has similar problems due to its calculation method.So the input parameters of the internal corrosion prediction model also have aleatory uncertainty.Based on the internal corrosion direct assessment and the Kriging aleatory uncertainty analysis method,this thesis has solved the aleatory uncertainty of the model and the parameters,that improves the accuracy of the corrosion prediction of CO2-containing wet gas gathering pipelines.This thesis has studied the internal corrosion rate prediction model and analyzed the model's aleatory uncertainty.The study found that the internal corrosion process is a complex microscopic process.Due to the complexity of the internal corrosion mechanism,the classical prediction models are mostly based on certain experimental data and mathematically fitted.When the distribution law of the test data is different from the data distribution law under the actual conditions,aleatory uncertainty is generated.Then this thesis studies the principle and the parameter uncertainty of multiphase flow calculation of CO2 wet natural gas gathering pipeline which is inseparable from internal corrosion prediction.It mainly studies the key parameters of the internal liquid holdup rate and the multi-phase flow pattern,which have the greatest influence on internal corrosion,and uses the PIPESIM software to simulate the multi-phase flow process.And this part analyzes the calculation theory of the liquid holding rate and multi-phase flow pattern in pipeline.It is found that there are some aleatory uncertainties in the calculation of the basic flow process and the calculation of the flow pattern.Next part of this thesis studies the aleatory uncertainty analysis algorithm and the Python implementation of this algorithm.The Kriging method,an aleatory uncertainty analysis algorithm,is applied to the internal corrosion prediction of CO2-containing wet natural gas gathering pipeline,and the whole algorithm is optimized by self-sampling method and combined with genetic algorithm and particle swarm optimization.This algorithm is programmed by Python.Based on the algorithm research,the De Waard-Milliams internal corrosion rate prediction model is improved and the hyperparameters are optimized.Through a calculation example,the calculation process and result of the whole algorithm are shown in detail.Finally,based on the internal corrosion direct assessment,this thesis applies the algorithm to the real internal corrosion prediction process,and predicts the internal corrosion of the No.1 gathering pipeline of Su 120-3 station at the third gas production plant of CQ oilfield.The results consists of the inclinations of the pipeline,the multiphase flow simulation results and the corrosion rate prediction results based on Kriging method.We can get the mean value of the internal corrosion rate of the target pipeline,the variance of the predicted internal corrosion rate,and the probability of corrosion in the pipeline to various degrees.In the example of this part,the error rate of the classical algorithm is 43%.The algorithm based on the aleatory uncertainty analysis proposed in this paper can control the error rate to about 10%.It has been tested and verified that the internal corrosion prediction method based on aleatory uncertainty can improve the accuracy of the internal corrosion prediction of CO2-containing wet natural gas gathering pipeline.