产量预测是油田开采方案设计的一个基本前提。精确的产量预测能够为油田的开采方案调整提供指导方向,能够决定油田的开采规模。由于地质储层性质不一,使得油田产量预测误差很大。小波神经网络因其更好的收敛性和处理复杂地质状况的特性,能够对非均质油藏提供较常规预测更加精准的预测。经过实际参考数据进行模拟计算,计算结果误差更小,证明其产量预测可作为油藏实际开采的参考。 Production forecasting is a basic prerequisite for the design of oilfield recovery plans. Accurate production forecasting can provide direction for the adjustment of the production plan of the oilfield and determine the size of the oilfield. Because of the different nature of geological reservoirs, oilfield production prediction error is large. Because of its better convergence and the ability to deal with complex geological conditions, WNNs can provide more accurate predictions for heterogeneous reservoirs than conventional prediction. After the actual reference data is simulated, the calculation error is smaller, which proves that the prediction of production can be used as a reference for actual reservoir exploitation.