基于GSA-BP神经网络的电厂风机运行状态评估

State Evaluation of Fan in Power Plant Based on GSA-BP Neural Network

风机的运行状态在很大程度上影响了火电厂生产的安全性和经济性。针对其健康状态的监测问题,提出了一种基于改进BP神经网络的风机运行状态评估方法。首先,通过BP神经网络建立正常状态下风机振动的预测模型;然后,根据风机的实时运行参数得到健康状态下风机振动的预测值;最后将振动的预测结果与实际值的相对误差来反映机组的退化程度和健康状态。为提高BP神经网络的收敛速度和泛化性能,将引力搜索算法(Gravitational Searching Algorithm,GSA)应用于神经网络权值的优化。现场实测数据表明,所提方法有较高的预测精度,能有效应用于风机运行状态的智能化评估。

The operating state of the fan in a power plant greatly affects the safety and economy of power generation.Aiming at the online monitoring problem of its health status,this paper proposes an intelligent evaluation method based on the improved BP neural network for the operating state of the induced draft fan.Firstly,the BP neural network is used to establish the vibration prediction model based on the operating parameters under normal conditions.Then,according to the real-time operating parameters of the fan,the predicted value of the induced draft fan is obtained.Finally,the predicted value of the vibration is compared with the actual value to evaluate the health state of the fan.In order to improve the convergence speed and generalization performance of BP neural network,Gravitational Searching Algorithm(GSA)is applied to the optimization of neural network weights.The measured data show that the vibration prediction model has higher prediction accuracy and can be effectively applied to the intelligent evaluation of the operating state of the fan.