摘要
针对功角测量方法在极对数多、转速慢的水电机组中误差较大,且在暂态过程时需要进行复杂的离线计算,难以满足功角高精度的在线监测要求,提出了一种改进的RBF神经网络功角软测量方法。将电压、电流及功率作为软测量模型训练集的输入,离线计算所得的稳态及暂态过程的功角值作为训练集的输出,并利用PSO算法对RBF神经网络参数进行优化。在单机无穷大系统中进行验证,验证结果表明,功角软测量的均方误差在0.004 9以内,且测量值与功率的映射曲线和理论功角特性曲线基本保持一致,所提出的软测量模型能够很好地拟合机组响应参数与功角的关系,实现了水电机组功角精准的在线监测。
In view of the large error of the load angle measurement methods in hydroelectric generating unit with many pole pairs and slow speed, and the complex off-line calculation is required in the transient process, which is difficult to meet the requirements of high-precision on-line monitoring of load angle, an improved RBF neural network load angle soft sensing method was proposed.The voltage, current and power were used as the input data of the training set of the soft sensing model, and the steady-state and transient load angle values calculated off-line were used as the output data of the training set.The parameters of RBF neural network were optimized by PSO algorithm.It is verified in a single machine infinite system and the verification results show that the mean square error of load angle soft sensing is less than 0.004 9,and the mapping curve between the measured value and power and the theoretical load angle characteristic curve are basically consistent.The proposed soft sensing model can well fit the relationship between unit response parameters and load angle, and the accurate on-line monitoring of load angle of hydroelectric generating unit is realized.
作者
蒋小平
雷佳玉
肖业祥
JIANG Xiao-ping;LEI Jia-yu;XIAO Ye-xiang(School of Mechanical Electronic and Information Engineering,China University of Mining and Technology-Beijing,Beijing 100083,China;Department of Energy and Power Engineering,Tsinghua University,Beijing 100084,China)
出处
《仪表技术与传感器》
CSCD
北大核心
2022年第8期108-114,共7页
Instrument Technique and Sensor
基金
国家工业和信息化部高原水轮机建设项目(W03010)
中央高校基本科研业务费专项(2020YJSJD15)。
关键词
水电机组
功角
在线测量
软测量
暂态功角
径向基神经网络
hydroelectric generating unit
load angle
online measurement
soft measurement
transient load angle
radial basis neural network