Research On The Optimization Of The Recipe For Product Oil Blending In Petrochemical Industry

In china, petrochemical industry is the pillar industry withproduction line length and covers a wide range of features. As animportant part of the petrochemical industry, oil refining industryshoulders the responsibility to provide the petroleum products for variousprofessional domains. As the main products of oil refining industry,product oil occupies a considerable proportion in the wholepetrochemical industry. Product oil blending is the last step before theyleave the refinery, and it is also the key link of affecting refineryeconomic benefit, so it is very necessary to make a study aboutoptimization. In view of the significance of product oil blending for therefinery and based on the background of petrochemical industry, thispaper makes a research on the optimization problem about multistageproduct oil blending recipe. It includes the following contents:Firstly, the article introduces the background of the petrochemicalindustry and the significance of the subject, then makes a comprehensiveanalysis about the development and present research conditions ofproduct oil blending problems at home and abroad. And production planand inventory problem is elaborated by systematic expression.Secondly, the paper introduces three methods to solve theoptimization problem. Also it introduces the development, basic principleand broad application about genetic algorithm.Thirdly, through analyzing the market consumption conditions aboutthree kinds of product oil, the reasonable prediction about their quantitydemanded is made respectively by GM(1,1) grey forecasting method andindex smooth forecast method.Fourthly, based on the demand forecast about product oil, this paperconstructs the mixed-integer optimization model of the product oilblending recipe which targets minimized enterprise total cost ofproduction, under the product quality index constraints satisfied.In order to show the actual significance of the model, the paperdesigns the parameters based on the actual production data from an oilrefinery, and choose seven groups of experimental data to solve the parameters by genetic algorithm. When the unit inventory cost and unitfixed cost of alternately changes, through the comparison of results, itshows that the total inventory cost of enterprise has relationship with thenumber of starts, and it has internal relations with total cost of production.