一种面向深度信念网络的故障诊断加速方法

A Fault Diagnosis Acceleration Method Oriented Deep Belief Network

针对深度信念网络(Deep Belief Network,DBN)故障诊断模型训练时间过长、效率低下等问题,提出一种面向深度信念网络的故障诊断加速方法,该方法通过构造一种指数损失函数来放大损失代价,通过梯度放大达到分类模型训练加速目的,同时提出基于指数损失函数的一种面向深度信念网络的故障诊断模型构建加速算法.故障诊断模型对比实验表明所提出的方法具有较好的加速效果,并具有更好的准确率,验证了方法的有效性.

Aiming at the problems of long training time and low efficiency of deep belief network(DBN)fault diagnosis model,a fault diagnosis acceleration method oriented deep belief network is proposed.In this method,the exponential loss function is constructed to amplify the loss cost,and the automatic gradient amplification method is used to accelerate the training of the classification model.Then,a fault diagnosis model construction acceleration algorithm oriented deep belief network is provided based on the exponential loss function.Through fault diagnosis models comparison experiment,it shows that the proposed method has better acceleration effect and better accuracy,which verifies the effectiveness of the method.