潜油柱塞泵MMAGA-RBF故障诊断方法

发布时间：2018-07-07 10:20

本文选题：潜油柱塞泵 + 特征参数　；参考：《哈尔滨工业大学学报》2017年09期

【摘要】：针对潜油柱塞泵无法使用传统地面示功图方法进行故障诊断的问题,提出了一种适用于该抽油机的故障诊断方法.该方法首先利用多变异位自适应遗传算法(MMAGA)对RBF神经网络进行参数寻优,然后从潜油直线电机的运行参数和油井井口的仪表参数等综合数据中,提取出反映油泵运行状态的特征参数,并将其作为故障诊断模型的输入向量,从而实现潜油柱塞泵故障工况的诊断.结果表明:MMAGA-RBF故障诊断方法能够在较少的训练样本下达到较高的综合诊断准确率,在训练样本集达到1 000组以上时,其综合误判率可低于4%,相对于普通遗传算法优化的模型,其泛化能力更强.MMAGA-RBF故障诊断方法符合潜油柱塞泵的工况特点,能够达到其综合诊断准确率的要求.
[Abstract]:Aiming at the problem that the submersible piston pump can not use the traditional ground dynamometer to diagnose the fault, a fault diagnosis method suitable for the pumping unit is put forward in this paper. In this method, the parameters of RBF neural network are optimized by MMAGA, and then the parameters of the submersible linear motor and the instrument parameters of the well head are integrated into the data, such as the operating parameters of the submersible linear motor and the instrument parameters of the well head. The characteristic parameters reflecting the operating state of the oil pump are extracted and used as the input vector of the fault diagnosis model to realize the fault diagnosis of the submersible piston pump. The results show that the fault diagnosis method of: MMAGA-RBF can achieve higher diagnostic accuracy with fewer training samples. When the training sample set reaches more than 1 000 groups, the synthetic misjudgment rate can be lower than 4 parts, compared with the model optimized by common genetic algorithm. Its generalization ability is stronger. MMAGA-RBF fault diagnosis method accords with the working condition characteristics of submersible plunger pump and can meet the requirements of its comprehensive diagnostic accuracy.
【作者单位】：哈尔滨理工大学电气与电子工程学院;大庆油田采油工程研究院;哈尔滨工业大学电气工程及自动化学院;
【基金】：黑龙江省自然科学基金(E201305)
【分类号】：TE933.3

【相似文献】