多分支水平井参数智能优化技术

多分支水平井参数优化设计是一个多目标最优化问题,采用传统的优化方法求解复杂程度依赖于优化对象数目,且容易产生局部收敛。基于多分支水平井的油藏数值模拟技术,以生产净现值为目标函数,应用遗传算法建立了多分支水平井参数智能优化设计方法,并编程实现了优化设计的全程自动化。优化过程中,利用正交设计原理生成种群初值,避免了初始种群的随机盲目性;根据个体适应值大小选择交叉和变异概率,保证了种群的多样性和算法的全局收敛能力。以珠江口盆地某海上低渗透油藏为例进行了多分支水平井参数优化设计,结果表明:遗传算法优化具有全局智能搜索寻优的特点,优化结果比传统优化算法有较大提高,具有较强的优越性和实用性。

Intelligent optimization technology for parameters of multi-branch horizontal well

Parameter optimization design for multi-branch horizontal well is an issue of multi-objective optimization. The complexity of optimization method solution using traditional method depends on the number of optimized objects, and local convergence may occur easily. Based on reservoir numerical simulation technique for multi-branch horizontal well and taking the net present value of production as the objective function, the intelligent optimization design method for parameters of multi-branch horizontal well was established using genetic algorithm and the whole-course automaton of optimization design was achieved by programming. During optimizing, the orthogonal design principle was used to generate the initial value of the population, which avoided the random blindness of initial population. The probability of crossover and mutation was selected according to individual fitness value, which ensured the diversity of the population and global convergence ability of the algorithm. An offshore low-permeability oil reservoir at Zhujiang River Mouth Basin was used as an example to carry out optimization design for parameters of multi-branch horizontal well, and the result showed that the genetic algorithm optimization had a feature of global intelligent search optimization, and the optimization results were improved greatly compared with traditional optimization algorithm. So this method has significant superiority and practicability.