摘要：【中文】本发明公开了一种基于云边融合架构的小水电发电量预测系统及其实现方法，从外部系统采集所测区域的实时气象数据和小水电发电数据，通过VPN通道直接上送云计算服务器，通过互联网实现系统的人机界面；云计算服务器将气象数据和小水电发电数据传给边缘服务器储存，构建出小水电发电量预测模型，并利用该模型进行小水电发电量预测，然后将预测结果回传云计算服务器，完成了数据分析、挖掘、预测等系统核心分析功能，最终将预测、分析结果通过互联网传输到终端设备上展示。本发明是云计算及边缘计算技术的有机结合应用的典型案例，在提高系统可靠性的同时，有效的解决了电网敏感数据储存、传输、展示的问题。 【EN】The invention discloses a system for predicting the power generation capacity of a small hydropower station based on a cloud-edge fusion framework and an implementation method thereof.A real-time meteorological data and small hydropower station power generation data of a measured area are acquired from an external system, and are directly uploaded to a cloud computing server through a VPN (virtual private network) channel, and a human-computer interface of the system is implemented through the Internet; the cloud computing server transmits the meteorological data and the small hydropower generation data to the edge server for storage, a small hydropower generation prediction model is constructed, the model is used for small hydropower generation prediction, then the prediction result is transmitted back to the cloud computing server, the core analysis functions of the system such as data analysis, mining and prediction are completed, and the prediction and analysis results are transmitted to the terminal equipment through the internet for display. The invention is a typical case of organically combining cloud computing and edge computing technologies, and effectively solves the problems of storage, transmission and display of sensitive data of a power grid while improving the reliability of a system.

摘要：【中文】本发明公开了一种基于大数据驱动的小水电群发电量预测方法，针对小水电群点多面广、无序管理的现状，将其按照区域划分；获取各区域的历史气象数据和发电量数据，采用数据挖掘技术对坏数据进行预处理，然后再做归一化处理；利用两层长短期记忆神经网络和一层全连接神经网络，对历史气象数据和发电量数据进行训练，并以测试数据的发电量真实值与训练模型发电量预测值的均方根误差最小为优化目标进行优化计算，最终得到小水电群的发电量预测模型。本发明的方法可对小水电群发电量进行精准预测，实现了小水电站的综合有序管理，保证了电网的安全稳定运行。 【EN】The invention discloses a method for predicting the power generation quantity of a small hydropower station group based on big data driving, which is divided into areas according to the current situation of multi-surface, wide and unordered management of the small hydropower station group; acquiring historical meteorological data and generating capacity data of each area, preprocessing bad data by adopting a data mining technology, and then performing normalization processing; and training historical meteorological data and generated energy data by using two layers of long-short term memory neural networks and one layer of fully-connected neural network, and performing optimization calculation by taking the minimum root mean square error between the true generated energy value of the test data and the predicted generated energy value of the training model as an optimization target to finally obtain a generated energy prediction model of the small hydropower station group. The method can accurately predict the power generation quantity of the small hydropower station, realizes the comprehensive and ordered management of the small hydropower stations, and ensures the safe and stable operation of the power grid.