考虑配置的多源注水系统泵站运行方案优化

为了实现对多源注水系统中各站内运行注水泵的型号、数量及运行参数同时进行优化,以系统输入功率最小为优化目标,考虑水量平衡、泵流量、泵正常运行压力、泵型号等约束,建立了考虑泵配置的泵站运行方案优化数学模型.针对优化问题的特点,设计了双重编码改进遗传算法对模型进行求解.在双重编码中,第1行采用实数编码表示泵的流量,第2行采用整数编码表示泵所属注水站编号,实现了对优化变量的准确描述,并指出了解的编码为变长度编码.设计了根据流量确定泵型号的方法,泵的其他运行参数根据选择的泵型号确定.在改进遗传算法操作过程中,设计了初始解的产生方法及与问题相适应的多种交叉和变异方法,使部分约束条件得到了满足,减少了不可行解的数量.利用本算法对某深度注水系统进行了优化设计,优化后系统输入功率比现状降低了2.42%,表明在多源注水系统中对泵的配置及运行参数进行同时优化节能效果明显.

Optimizating pump station operation scheme and pump selection in multi-source water injection systems

In order to optimize the type and number of pumps, operation parameters of pumping sta- tions in multi-source water injection systems simultaneously, a mathematical model was established, in which the input power of pumps is objective function, the water balance in system, flow rate and ope- rating pressure as well as type of pumps are served as constraint conditions. Taking the features of this optimization problem in account, a dual coding was proposed to improve the existing genetic algorithm, and then was appied to solve the mathematical model. In that dual coding, a real number was adopted in the row in the coding to indicate the flow rate of a pump, an integer was put into the second row to present the number of a water injection station where the pump was installed; thus, an exact discrib- tion of the optimization variables was realized ; obviously, the coding of solution was in variable length. An approach was porposed, where the type of pump was determined initially according to a flow rate, then the rest operating parameters of the pump were decided based on the known type. In applications of the improved genetic algorithm, the method for generating of initial solutions was given ; meanwhile, a variety of crossover and mutation methods, which were adaptable to the optimization problem, were cast. As a result of this, a part of the constraint conditions were satisfied, and the number of infeasible solutions was reduced. The water injection system in an oil field with deep oil wells was optimized by using the algorithm proposed, the total input power was reduced by 2.42% , showing simultaneously optimizating pump seelection and pump operating parameters has a remarkable energy saving effect in the multi-source water injection system.