Particle Swarm Algorithm To Optimize The Design Of Applied Research, In The Hot Oil Pipeline

Long distance oil & gas pipeline is a system engineering, which needs large investment. It is worth exploring and improving to decrease pipeline's bulild investment and operating fee by optimal design.This article introduced the actuality and development of worldwide oil&gas pipeline firstly, analyses and studies the former long distance pipeline optimal design.Taking the long distance heated oilpipeline as study object, considered the uncertainty of the throughout in the real operation , according to the long distance transportation theory and optimal design principle, a new oil pipeline optimal design model which has been founded. In the model, the variables include D(outside diameter of pipeline) δ_g (wall thickness of pipeline) δ_b (thickness of insult layer), n_R (amount of the heating stations) n_c (amount of the pump stations) t_R (the temperature at the outlet of every heating station) and Hc(the lift of every pump station).The total outlay expending on the pipeline system can be used as the norm appraising the pipeline's economic value.Considering the effect of investment yield,we choose the converted cost as the objective function.We roundly consider the thermal constraints, hydraulic constraints, strength constraints,and pump's performance constraints. In order to improve the calculate definition, the affections of fictional heat and thickness of wax deposition are considered, and the oil physical characteristics, hydraulic gradient, heat transfer coefficient and thickness of wax deposition are looked as variables.The different equations are used in different segments and different patterns.Mixed Discrete Complex Programming (MDCP) is a mature optimization recently . It is suitable for the questions containing discrete variable .But, its' astringency is slow and easy to be trapped in the partial solution . Mixed Discrete Random Programming (MDRP) is one kind of random arithmetic,this arithmetic's superiority is easy and quick, but it is easy to constringe partially.Particle Swarm Optimization (PSO) is a genic arithemetic,its'superiority is easy and quick and easy to be programmed. PSO is suitable for the successional optimal questions.According to the characteristics that in the model there are successional variables and discrete variables,we propose a synthetical arithemetic which combines the superiorities of PSO and MDRP. We use PSO optimal design pipeline first, when the searching is near the best successional solution, then use MDRP to search for the best discrete solution directly. This method effectively settles the limitation of the PSO arthemetic .At last, in a practical long distance heated pipeline design, the module is programmed by using the computer language Visual Basic 6.0 .We optimal design it by using the arthemeticusing the computer language Visual Basic 6.0 .We optimal design it by using the arthemetic PSO, MDCP, MDRP and PSO-MDRP respectively. The reliability and practicability of PSO-MDRP are approved by comparing the optimal results. The final design more accord with the reality that annual throughouts are various.This article is useful for the oil pipeline optimal operation and long distance gas pipeline design to some extent.