BLENDING AND OPTIMIZATION OF CRUDE OIL WITH MULTI-QUALITY AND MULTI-PROPERTIES BASED ON INTELLIGENT ALGORITHM
Aiming at the problems of refineries relying on traditional experience to blend crude oil to meet the quality requirements of processed crude oil and the low profit of enterprises,a multi-objective optimization mathematical model based on the maximization of crude oil similarity and the maximization of production profit was established,and the Gray Wolf Algorithm,which combined the Levy Flight and Randomized Swimming Strategies(LRGWO),was chosen to solve the model after comparing the performance of five robust algorithms.The results showed that in LRGWO,Gray Wolf Algorithm(GWO),second-generation Non-dominated Sorting Genetic Algorithm(NSGA-Ⅱ),third-generation Non-dominated Sorting Genetic Algorithm(NSGA-Ⅲ)and Particle Swarm Algorithm(PSO),the coverage of the optimal solution set obtained from the LRGWO algorithm and the overall performance of algorithm were optimal.After optimization,the average similarity between the blended crude oils T1-T4 and the corresponding target crude oils W1-W4 reached 95.82％,which indicated that the blended crude oils with similar physical properties to the target crude oils could be obtained by using the constructed crude oil selection and blending model.According to the prices of the 10 kinds of crude,under the premise of the maximum purchasing quantity of 175 kt,the production profit maximization model optimizes the processing quantity of blended crude T1-T4 to be 50,55,40 and 30 kt,respectively.The maximum production profit of the enterprise is 33.5819 million Yuan.
crude oil blendingsimilarityproduction profitmulti-objective optimizationGray Wolf Optimization Algorithm
万方数据
维普
