Data-driven Intelligent Evaluation of Injection-Production Well Pattern

Aiming at the time-varying complexity of heterogeneous reservoirs and the current problem of massive data and tiny information in oilfields, an intelligent evaluation system of injection-production well pattern is developed based on data-driven method and computing with words idea in the paper. At first, based on the digital filtering approach, the production data of oil field is processed and the estimation of injection-production relationship is calculated. Secondly, the enhanced interval approach is used to encode the descriptive words of geological information and injection-production relationship so as to obtain the code book. Thirdly, According to the useful rule bases and perceptual reasoning outputs of oil field production evaluation obtained in the preliminary work, the recommended output of decision is obtained by using the decoder according to Jaccard similarity measure based interval type-II fuzzy sets. Finally, the effectiveness and practicability of the proposed scheme is testified by production example of inverted nine-point well-pattern of heterogeneous reservoirs with channel sedimentation with the aid of expert verification.