基于改进布谷鸟搜索优化RBF神经网络的抽油机故障诊断

Pumping machine fault diagnosis based on improved cuckoo search optimized RBF neural network

针对目前油田抽油机故障诊断存在耗时低效、无普适性等问题,提出通过动态自适应布谷鸟搜索(PSCS),优化径向基函数(RBF)神经网络的诊断方法。首先对示功图进行特征提取,作为RBF神经网络的输入层信息;接着引入动态发现概率和自适应步长,令布谷鸟搜索根据目标函数的收敛速度自动调整步长,确保不同搜索阶段的效率和精度保持平衡;最后改进的布谷鸟搜索优化RBF神经网络,获取其宽度、权值等最优相关参数,建立PSCS-RBF故障诊断模型。将模型应用于抽油机不同故障类型的诊断,并与当前主流的5种方法比较,所提出的PSCS-RBF故障诊断方法的平均检测精度达到95.9%,精度最高且耗时最短,验证了其实用性和优越性。

To address the current problems of time-consuming and inefficient fault diagnosis of oilfield pumping machines,which are not universal and require high hardware resources,this study proposes an improved cuckoo search(PS cuckoo search,PSCS)by introducing dynamic discovery probability and adaptive step size,and then an optimized diagnosis method for radial basis function(RBF)neural network.Firstly,the max-min algorithm is used to normalize the schematic power map,and then the shape invariant moments combined with Fourier descriptors are used for feature extraction as the input layer information of the RBF neural network.Then,dynamic discovery probability and adaptive step size are introduced to make the cuckoo search easier to retain the better solution,and the step size can be automatically adjusted according to the convergence speed of the objective function to maintain a balanced state of efficiency and accuracy in different search stages.Finally,the RBF neural network is optimized by the improved cuckoo search to obtain the optimal relevant parameters such as width and weight of the RBF to establish the PSCS-RBF fault diagnosis model.The model is applied to the diagnosis of different fault types of pumping machines and compared with the current mainstream five methods,the average detection accuracy of the proposed PSCS-RBF fault diagnosis method reaches 95.9%with the highest accuracy and the shortest time,which verifies the practicality and superiority of the proposed method in this study.