摘要
针对TRT系统中透平机结构复杂,故障特征表现及故障产生原因普遍具有模糊性、复杂性的特点,将现场采集的数据利用小波变换的软硬阈值折中算法处理小波系数,滤除噪声。通过建立量子神经网络(QNN)预测模型中网络结构的调整、网络的训练,得到预测结果。实验仿真结果表明:利用小波变换可以有效地滤除数据中的噪声,所建立的QNN预测模型可以有效地实现对TRT系统中透平机的运行状况预测。
In view of the characteristics of ambiguity and complexity of faults and causes for turbine in TRT system, the tur bine operation data collected from production scene are processed by the soft and hard threshold compromise algorithm in decompos ing the wavelet coefficient and filtering the noise to achieve the prediction of operation data of turbine. Then, the prediction of op eration condition model of Quantum Neural Network is established to adjust the structure of the neural network and train for the prediction results. The simulation results show that wavelet transform can filter the noise effectively, and model of Quantum Neural Network can realize prediction of operation condition of turbine in TRT.
出处
《电子技术应用》
北大核心
2014年第5期72-74,78,共4页
Application of Electronic Technique
基金
河北省自然科学基金钢铁联合基金资助项目(F2012209015)
关键词
TRT
透平机
小波阈值滤波
量子神经网络
预测
TRT
turbine
wavelet threshold filtering
quantum neural network
prediction