Perpustakaan Digital ITB

The oil and gas reserves keep decreasing during the production time. The company is required to operate the field in an effective and efficient manner, i.e., each well needs to be produced at an optimum production rate and minimum production cost. Using the current technology and software functionality, the production engineer works separately between well modeling and well optimization. Therefore, an engineer requires longer time and more steps to obtain an optimum operating condition for asset optimization. Careful and thorough steps need to be followed during well optimization. If the well has never been produced before, then a matching process between production data and model results does no need to be done. Unfortunately, due to the reservoir condition and tubing condition will change during the production time, then the matching process between the model and the actual condition cannot be avoided. Several parameters can be used to match the inlet condition of the liquid source, such as productivity index (PI), BHP value, etc. At the same time, parameters that can be used for matching on tubing conditions are tubing roughness, flow correlation, etc. Based on the thorough analysis, this paper proposes three parameters being used for the matching process: productivity index, head factor, and holdup factor. The optimization process aims to get the optimum condition based on several constraints and boundaries by using algorithms. It will maximize or minimize an objective function based on the problem relative to some constraints that represent the acceptable condition in the problem. In this study, the optimization process is used to find the maximum liquid rate using constraints of tubing head pressure (THP), pump frequency, pump number of stages, and maximum available power of the pump. The process will use particle swarm optimization (PSO) to get these variables that give maximum liquid rate. The workflow in this paper has several advantages from the traditional workflow. The workflow allows the modeling and optimization process to be automated from data input to optimization results. The optimization process also takes less time rather than traditional sensitivity analysis. Using the workflow, the engineer can minimize non-productive time and save time to further product design and analysis.