The Research Of The Optimized Running Of Oil Field Injection System

Water injection was one of the most widely used oil production modes in onshore oil fields in china. This mode could effectively supply stratum energy, which played a key role in improving the recovery ratio and guaranteeing high and stable yields. Energy consumption by injection system in large-scale oilfields was enormous, and increased as the water content in crude oil rose. Accordingly, it was significant to optimize the operation of injection system to reduce the production cost in oilfields.This thesis summarized and analyzed the network of injection system in oilfields, and educed the computational equation set of injection network based on the mass conservation equation and energy conservation equation. The solution based on simple iteration process was also demonstrated. By use of this technique, combined with the known node parameters, the pressure and flowrate in any node of the water injection network could be calculated, which laid foundations for the water injection system optimization.On the basis of establishment and calculation of the network model, this thesis elicited the optimized mathematical model of the operating parameters of injection system and advanced the processing method under constraint conditions. With the application of dross-cascade binary encoding method this paper encoded the output of the water injection pump and executed crossover and mutation operation utilizing the corresponding pointcrossover operator and simple mutation operator. In the choosing of crossover and mutation probability, the adaptive thinking was brought in, and the crossover and mutation probability was chosen according to the contemporary adaptive level of population, thus AGA was established. In the choosing of selection operator, the thinking of hybrid genetic algorithm was brought in and Boltzmann selection operator was used because this operator consulted the main idea of simulated annealing. In the choosing process individuals with less adaptivity were selected and imported into the next generation group with a specified probability to sustain the diversity of group, so premature convergence was avoided. Because of the inherent truncation error of binary encoding, some of the constraint conditions in the optimized model were hardly satisfied. Therefore, the operations of adjustment and fine adjustment of the output of the water injection pump were set. By adjusting the pump output the constraint conditions were satisfied.On the basis of optimization of operating parameters of water injection system, this thesis completed the self-adaptive hybrid genetic algorithm simulated annealing and educed a genetic algorithm adequate for optimization of water injection system. The 0-1 coding was used to depict the pump programme of the water injection pump, and then input this to output coding, thus this paper displayed a uniform multiparpmeter cross-cascade binary encoding programme. In the optimization of the pump programme, some obviously unfeasible pump programme emerged. So it was necessery to assess the new individuals generated in each genetic algorithm as to feasibility. If the programme was unfeasible, the implement of optimization was unnecessary, in this way time saved.On the basis of work above, utilizing software ergonomics, the establishment ,calculation and output process of the model got execution of computer operation in Visual Basic programming environment. The software wrote for optimization of the water injection system had a friendly interface and was convenient to operate.