The Research Of Calibration Methods On Pipe Friction Factor In Oilfield Water Injection System

Oil field water-injection network plays a role of allocating water in water- injection stations to different oil wells in accordance with the requirements of production. If pipe friction factors are known in oil field water-injection network, the pressure values of the nodes can be obtained through solving a system of nonlinear equations. We can also achieve the purpose of energy saving through operation optimization studying in oilfield water-injection network. In simulation and operation optimization, raw pipe friction factors are applied in mathematical model now. Oil field water-injection network is the high-pressure system, and the pipelines'diameter are relatively small, and they have been corroded seriously,and they have been built for a long time,so the pipe friction factors has been changed. The practical pressure and calculated pressure would be inconsistent if adopting original friction factors. So we need to correct friction factors. The following aspects are researched in this paper:The constrained least squares model of correcting friction factor under single operating mode is established, and an effective numerical method is given. It is proved that the model has infinite solution for unconstrained least squares problems. Sensitivity analysis can also reduce number of calibration variables,improve accuracy of the correction,and obtained exact values of the actual friction under certain conditions.On the premise of knowing pressure values of all nodes, the optimization model of correcting friction factor with multi-mode data is established, in which the operating mode number is obtained by the number of pump stations, pressure values as basic data, the range of friction factor value as constraint condition. For the larger network, the grouping method reduces dimensions of optimization variables, and increases the computation speed and precision of correcting. Finally optimization model is solved by improved PSO.Under the condition of knowing a part of nodes pressure values, the unknown pressure values are as random parameters. The expected value model is established with analyzing random parameters. Through analyzing limitations and reliabilities of the expected value model, chance-constrained programming model is established. The two types of random models are solved by GA and stochastic simulation.Under the condition of knowing a part of nodes pressure values, the unknown pressure values are as fuzzy parameters. The fuzzy optimization model is established with analyzing membership function of the fuzzy parameters. The fuzzy optimization model is solved by GA and fuzzy simulation.Particle swarm algorithm is improved in the aspects of inertia weight, learning factor, speed variation, optimal-worst-particle and enhancing local searching capabilities according to its characteristics; Genetic algorithm is improved in the aspects of retaining history best individual, periodic replacement of the current best individual and periodic axial searching according to its characteristics. The two algorithms increase the ability of local searching to improve computational efficiency.Finally, on the basis of the theory studied above, a set of software for friction factor calibration for water injection system is made with VB builder language. The software can correct friction factors under single operating model, multi-model, random environment and fuzzy environment. The software can also be used to solve friction factors of water supply network.