Research On Intelligence Analysis Of Running State And Operation Optimization In Water Flooding System

Oil field flooding system is a large scale, closed and complex hydrodynamic system, which covers large area and delivers water deep into ground continuously everyday, the power consumption of the system is also enormous. The system need to be simulated and analyzed to give scientific instruction or make reasonable policy no matter for design of the system or the running of the system, else, the dynamic devices running situation may not match better with pipe network, which will limit or disturb fluid delivering in the network, the system efficiency will be effected and the power consumption of dynamic devices may be high. Since the water flooding system is increasing, the pipes is updating and the application of new control technology of pump, the study on water flooding system is enhancing. The intelligentized methods used on the analysis and optimization of water flooding system are key to realize advanced management and scientific decision of water flooding system. The following research is carried out in the dissertation.Water flooding network modeling and solution. According the principle of finite-element method the water flooding network is discretized, the hydraulic unit models for normal pipeline, hot water pipeline and valve are set up, a thermodynamic model for hot water pipeline is also set up. The unit models are conversed and linearized to obtain the same expression, and then the unit models can be assembled to general equation sets for water flooding networks, which can be solved through iterating algorithm, and thus to attain better simulation for complex networks.Study on water flooding network partition according to pressure. The high pressure water flooding wells is now mixed with low pressure ones in the water flooding system, and the network pressure should be rise to meet few high pressure wells, which will power consumption rise. A network dividing method is presented based on fuzzy logic theory, the fuzzy field and inferring rules are set up, water flooding network is divided into high and low pressure area, and different pressure is applied to different area. The fuzzy dividing model is simulated in Matlab and applied on an actual network, the result is reasonable.Water flooding system operation optimization study. The operation features of water flooding system and its constrain conditions are analyzed, the output water optimization model for stations of multi-source water flooding system is set up, the operation parameters optimization model of large scale complex water flooding system is set up, and the optimization models for frequency control of motor speed and pump starting schema of large scale complex water flooding system is also set up. The operation optimization model for water flooding system concerning many variables, many equality constraints and inequality constraints, what's more, the objective function is nonlinear function with multi-peak, which often cause solution fall into local minimum. Among these intelligent optimization algorithms Genetic Algorithm and Ant Colony Algorithm have characteristic of better global searching ability, suitable for distributed computing and better robustness. But these algorithms have their shortcoming. A new self adapting algorithm mixed Ant algorithm and Genetic algorithm (MAGA) is presented, which can effectively improves rate of convergence. MAGA is used in the three operation optimization models, the results show the effectiveness and its feasibilityThe research achievements have been applied on oil field, and better effectiveness is achieved.