基于Bitmap的油水井采注优化实时推理引擎

A Real-time Reasoning Engine for Injection-production Optimization of Water and Oil Wells on Account of Bitmap

针对油田油水井采注优化业务中,油水井数据量大、地层结构复杂以及人类经验多的特点,分析了传统推理方法在油田采注实时优化处理过程中的不足,采用事件处理思想,提出了一种基于Bitmap事件编码与匹配机制的推理引擎,有效地实现了对无效事件的过滤并提升了事件与规则的匹配效率.在油田实际数据试验平台上对该方法进行了验证并与RETE算法、LFA(Linear forward-chaining)算法的性能对比,结果验证了本文方法在实时推理能力上的有效性.

In the process of optimizing the injection-production in oil and water wells, complex stratigraphic structure and large amount of business data and human experience will be involved. In this situation, traditional reasoning methods cannot be effective. By introducing the event processing theory, a reasoning engine using bitmap event encoding and matching is proposed, which can filter out invalid events efficiently and improve the matching performance between events and rules. The proposed reasoning engine is implemented in a real oilfield data experiment platform. Compared with RETE algorithm and LFA（Linear forward-chaining） algorithm, the proposed method shows a better reasoning capability.