To solve current problems of difficulty,high cost,and low efficiency in data collection for power grid dispatch,a re-al-time data collection method for power grid dispatch automation based on similar days and BA-WNN is proposed to promote intelligent management of power grid dispatch data and ensure safe and stable operation of the power grid.This paper calculate the similarity index by using cosine similarity,select similar days,and establish a sample set of similar days for power grid dis-patch data.The sample set of similar days of power grid dispatch data is used as the input of wavelet neural network(WNN),the bat algorithm(BA)is used to replace the fastest gradient descent method in WNN to continuously optimize the weights and wavelet basis coefficients of WNN.It improves WNN's convergence effect,and outputs the best real-time data collection re-sults for power grid dispatch automation.The experimental results show that this method can accurately select similar days and effectively collect power grid dispatch data.
关键词
相似日/蝙蝠算法/小波神经网络/电网调度/自动采集
Key words
similar day/bat algorithm/wavelet neural network/power grid dispatch/automatic collection
维普
万方数据