A STOCHASTIC PROCESS APPLICATION FOR PREDICTING PITTING CORROSION AND SCALE GROWTH IN OIL/GAS PIPELINE

• Main Author: Kurniawan Santoso, Ryan
• Format: Final Project
• Language: Indonesia
• Subjects: Pertambangan dan operasi berkaitan
• Online Access: https://digilib.itb.ac.id/gdl/view/54705
• Institution: Institut Teknologi Bandung

This paper presents a method to predict pitting corrosion and scale growth in oil/gas pipeline using a stochastic model. Pitting corrosion and scale growth are chosen as the points of interest here because they are complicated to be deterministically estimated. A literature study was conducted to choose and select several probabilistic models to describe the pitting and scale growth process. Markov chain and Poisson model were chosen as the best probabilistic models to describe pitting corrosion and scale growth. Then, those two models were combined together to create a simple equation so that the calculation will be easier. Experimental data using iron plate in white water system were used to confirm and simulate the combined model to predict pitting corrosion. A field data of oil pipe in unburied and bare system was used to verify the suitability of combined model to be used to evaluate the scale growth. The scale used in the field data was paraffin wax scale. The results of the model for pitting corrosion and scale growth are then compared with JSCE/Gumbel Distribution Model results. By combining the two probabilistic models before, a combined function is derived to determine pitting and scale thickness. A gamma function plays an important role in determining the calculation result. For predicting pitting corrosion, the combined model was simulated using experimental data with 2, 3, 4, 5, 6, 7, 8, 9, 10, 20, 50, 80, 100, 150 and 170 pipe-thickness partitions scenario. Minimizing error method was applied to choose the best scheme. Thus, for pitting corrosion, 20 partitions scenario is the most suitable one. Using the same stochastic model for scale thickness prediction, 50 pipe-inner-radius partitions scenario gives the minimum error. From the predicted results, pitting corrosion and scale growth are following logarithmic function. Thus, logarithmic trendline was applied to the stochastic result to help to roughly predict the failure time for pitting corrosion and scale growth. The failure time is then plugged in to the stochastic model to confirm the probability result. Failure time prediction is very useful for planning preventive actions. Many measurements are conducted in the field to measure the pipe thickness or effective diameter during the production period. The measurement is so expensive. Laboratory experiments can be the alternatives. However, laboratory data are always obtained with many assumptions and cannot represent the real situation. Thus, this model was developed to give an accurate prediction.