Abstract: Considered that the mixed oil would be produced in the batch transportation of the product oil pipeline, the highest revenue in the life cycle as the objective function, the diameter and the wall thickness as the decision variables, a product pipeline network optimization model was set up. The calculation formula based on diffusion theory and empirical formula were used to calculate mixed oil quantity. And the genetic algorithm was used to solve the model. The results show that the design of mixed oil costs account for 3% to 10%, and it’s related to the factors as the transportation batches per year, the number of species transported in each batch and the cost on management of oil mixture per unit volume. Generally mixed oil loss can’t be ignored, and the weight coefficient method can be used for a variety of design options. Using different formulation often got different result, so in practical engineering projects the closest formulation can be chosen to calculate.v

Key words: Mixed oil, Product oil, Pipeline network, Optimal design,  MATLAB , Genetic algorithm

摘要： 考虑成品油顺序输送中产生混油的特点，以管道生命周期内收益最高为目标，管径、壁厚为决策变量，建立了成品油管网优化设计的数学模型。分别利用扩散理论推导的变径管混油量计算公式和经验公式计算混油量，并采用遗传算法求解该模型。计算表明，混油损失费用占整个设计的3%～10%，并且与年输送批次数、每一批次输送的油品种类数、单位体积的混油处理费用等因素相关，一般情况下不能忽略不计，可以采用权重系数法生成多种设计方案以供选择。而不同的混油计算公式结果往往不同，在实际工程中应选择与本工程最接近的公式进行计算。

关键词: 混油 , 成品油 , 管网 , 优化设计 , MATLAB , 遗传算法