Abstract: In this paper,an oilfield is taken as a dynamic system from the view of a grey box system control.A self-containing model for the prediction of the production decline of a water flooded candstone oil reservoir is established by the application of a method of multilayer staged prediction (1) combined with the method of muleiple variables,multiple steps self-checking recurrence prediction (2) based on the analysis of the statistics of changes in the withdrawal rate in a seguence of time.The future output of Daqing oil field without any adjustment programs has been predicted by means of such a model and a dynamic programming model is established and solred by the simplified one-step method of optimization to give an optimized control strategy based on these informations.An economically feasible plan with stable production for the oil field is thus worked out.

摘要： 本文从灰色系统控制角度出发,将整个油田看作是一个动态系统。在分析研究系统的产量变化时间序列统计规律基础上,采用多层递阶预报和多变量多步自校正递推预报器,分别建立了水驱砂岩油田产量递减自适应预报模型。并应用该模型预报大庆油田无措施条件下的产量变化。在此基础上建立了油田动态规划模型,采用简捷的一步优化方法,求解最优控制试编了满足油田稳产要求和经济上合理的五年规划。