In most cases, oil wells are completed and perforated to produce from multiple reservoirs. Thus, it is crucial to monitor production from each individual reservoir. Therefore, a new prediction method for multi-reservoir production by using temperature logging data is presented in this paper. Firstly, by coupling the wellbore and formation heat transfer with reservoirs heat transfer, a new comprehensive heat transfer model for the oil well with multiple reservoirs is established in this paper. Both the heat convection of two-phase flow in reservoir and vertical heat transfer between different reservoirs was considered in the model, and the relationship between flow rate distribution and temperature distribution could be specified by the model. Then based on the model, the stochastic approximation method was proposed to predict the multi-reservoir production through the temperature logging data. Lastly for a water injection well with five layered reservoirs and depth of 3500m, the production of each reservoir was predicted by the presented method. The results show that the temperature distribution simulated by the built model has a good agreement with the reference data. The estimation error of the total flow rate is less than 9% compared with the reference data, and the production of the five reservoirs accounted for 36%, 35%, 17%, 7% and 5% of the total production respectively.

在大多数情况下，油井已完工并打孔以从多个油藏中开采。因此，监控每个单独油藏的产量至关重要。因此，本文提出了一种利用温度测井资料预测多储层产量的新方法。首先，通过将井眼和地层的传热与储层的传热耦合，建立了多储层油井的新型综合传热模型。该模型既考虑了油藏两相流的热对流，又考虑了不同油藏之间的垂直传热，可以通过模型确定流量分布与温度分布之间的关系。然后根据模型，提出了一种随机逼近方法，通过温度测井资料预测多储层产量。最后，对于一个具有五个层状储层，深度为3500m的注水井，通过所提出的方法可以预测每个储层的产量。结果表明，所建模型模拟的温度分布与参考数据吻合良好。与参考数据相比，总流量的估计误差小于9％，五个储层的产量分别占总产量的36％，35％，17％，7％和5％。结果表明，所建模型模拟的温度分布与参考数据吻合良好。与参考数据相比，总流量的估计误差小于9％，五个储层的产量分别占总产量的36％，35％，17％，7％和5％。结果表明，所建模型模拟的温度分布与参考数据吻合良好。与参考数据相比，总流量的估计误差小于9％，五个储层的产量分别占总产量的36％，35％，17％，7％和5％。