基于改进的粒子群算法的钻进参数多目标优化研究

发布时间：2018-01-11 23:17

本文关键词：基于改进的粒子群算法的钻进参数多目标优化研究　出处：《西安石油大学》2015年硕士论文　论文类型：学位论文

【摘要】：钻进参数是影响和制约钻进速度、成本和质量等的重要因素。将钻进参数优化组合可以提高机械钻速、减少钻头磨损、降低钻井成本。原有的钻进参数优化理论和方法主要以钻速为核心,把降低成本作为单一目的,忽略了对其它因素的考虑,而目前的钻井工程不仅要提高钻速,更要综合考虑钻头寿命、比能等因素,需要多目标优化。另外,近年来应用智能优化算法来有效解决实时参数优化已成为钻井优化的关键技术之一,引入优化算法新理论对钻进模型进行求解是不可避免的趋势。本文针对单目标钻进参数优化的局限性,建立了一定约束条件下,综合考虑机械钻速、钻头寿命和钻头比能的多目标优化模型。针对现有优化方法求解效率低、易陷入局部最优等问题,提出一种改进的、基于混沌的粒子群优化算法。基准测试函数检验结果表明:改进算法具有良好的分布性和收敛性,而且跳出局部最优的能力较强。最后,将改进算法应用于某油田井段钻进参数优化模型的求解。利用回归分析法确定钻速模型的待定系数,讨论并选择了算法相关参数。仿真试验得到了较多且分布均匀的Pareto前端,从而验证了模型和算法的有效性,为工程实际优化方案的选择提供了有效依据。
[Abstract]:Drilling parameters are important factors that influence and restrict drilling speed, cost and quality. Optimizing drilling parameters can improve drilling speed and reduce bit wear. The original drilling parameter optimization theory and method mainly take drilling speed as the core, regard reducing the cost as the single purpose, ignore the consideration of other factors, but the drilling engineering at present should not only increase the drilling speed. In addition, the intelligent optimization algorithm has become one of the key techniques in drilling optimization in recent years. It is inevitable to introduce the new theory of optimization algorithm to solve the drilling model. Aiming at the limitation of single objective drilling parameter optimization, this paper establishes a certain constraint condition and synthetically considers the mechanical drilling speed. Multi-objective optimization model of bit life and bit specific energy. An improved method is proposed to solve the problems of low efficiency and easy to fall into local optimization. Particle swarm optimization algorithm based on chaos. Benchmark function test results show that the improved algorithm has good distribution and convergence, and the ability to jump out of the local optimization is strong. Finally. The improved algorithm is applied to solve the optimization model of drilling parameters in a certain oil field, and the undetermined coefficient of drilling velocity model is determined by regression analysis. The relevant parameters of the algorithm are discussed and selected. The simulation results show that the Pareto front end is more and more evenly distributed, which verifies the validity of the model and the algorithm. It provides an effective basis for the selection of practical optimization schemes.
【学位授予单位】：西安石油大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TE24;TP18

【参考文献】