摘要
针对石油开采中“电动机长期处于相对轻载状态”的问题 ,提出了规则自调整的RL模糊神经网络模型 ,并将其应用于抽油机的间歇控制中。该模型具有良好的自适应性能 ,在不断向环境学习的过程中 ,自动调节自身参数及结构 ,实现了模糊推理和模糊规则的构建。基于该模型开发的抽油机节电控制器已在大庆、胜利等采油厂投入使用。实践表明 ,抽油机启停时间合理 ,在保证采油量的前提下 ,节电率达 3 0 %以上 ,有效地解决了抽油机“相对轻载”的问题。
Aiming at long-term relatively light condition of motor in petroleum exploitation, we proposed a rule self-regulating RL fuzzy neural network model, that was applied to the intermittent control of pumping units. The model possesses favorable self-adaptability. In the course of learning from environment, it can carry out fuzzy inference and support structure of fuzzy rules by regulating parameters and structure by itself. Pumping unit controllers based on the model have been used in Daqing, Shengli oil recovery plant, etc. Experiments showed the pumping unit start-stop time was reasonable and brownout rate ran up to 30% upwards with the precondition of oil output equality. The problem of relatively light load of pumping unit is effectively solved.
出处
《控制工程》
CSCD
2002年第6期57-59,共3页
Control Engineering of China
基金
北京威英智通有限公司横向联合项目
关键词
RL模糊神经网络
采油控制
应用
油田开采
neural network
fuzzy rules
fuzzy Takagi-Sugeno model
reinforcement learning
pumping unit
