摘要
抽油井故障诊断系统是油井系统产量的关键,为了更好更快地对当前油井系统进行诊断以保证石油的产量,人们利用各种各样的技术来完成这一目标。BP神经网络技术就是新兴的油井故障诊断手段之一,但是由于其容易陷于局部最优解、收敛速度慢以及泛化能力差等缺点,运用融入免疫算法浓度概念的遗传算法对BP神经网络进行优化,并将优化后的BP神经网络算法应用于抽油机现场的故障诊断过程中,结果显示优化后的BP算法有更高的诊断率,更快的运行速度,因而优化后的BP算法的寻找全局最优点的能力更强,收敛速度更快。
Pumping well system fault diagnosis system is the key to production,for better and faster diagnosis of the current well system to ensure the production of oil,people uses a variety of technologies to achieve this goal.BP neural network technology is one of the wells fault diagnostic tool,but because of its easy trapped in local optimal solution,slow convergence and poor generalization ability,the paper uses concept of concentration in immune algorithm into genetic algorithm and uses this improved genetic algorithm to optimize BP and which used in the process of pumping site fault diagnosis,the results show that the optimized BP algorithm have a higher detection rate and faster speed,so the ability of optimized BP algorithm is more better than BP algorithm in finding global optimal point and convergence.
出处
《计算机与现代化》
2010年第12期182-185,共4页
Computer and Modernization
关键词
示功图特征
BP神经网络
浓度
遗传算法
模式识别
character of indicator diagram
BP neural network
diversity
genetic algorithm
pattern recognize
