一种高效的光伏监控方案及发电量预测

An Efficient Photovoltaic Monitoring Solution and Generated Energy Prediction

针对市场上光伏监控系统运行效率不高、发电量预测忽略太阳辐照度以及缺少嵌入预测功能等问题,提出一种以串口服务器为数据采集网关,以面向对象的方法描述设备、以多线程的方式处理数据、以模块化的思想布局,采用自定义数据队列、以嵌入的RBF神经网络模型预测发电量的高效光伏监控方案;以在广西大学内搭建的一个小型的光伏发电站为实验对象,在该方案的基础上实现多总线的光伏远程监控系统,并在该系统中嵌入发电量预测功能;经过长时间地运行系统并预测发电量,结果表明该系统具有运行效率高、可扩展性强、发电量预测精度高等优点。

An efficient photovoltaic monitoring solution is proposed by taking the serial port server as data collection gateway, using the object--oriented approach to describe devices, taking the multithreading technology to process data and using modularized layout, user--de- fined data queue and the embedded RBF neural network generated energy prediction model for the photovoltaic monitoring system to solve the problems such as low efficiency, ignoring the solar irradiance in output prediction and lacking of prediction function. A small PV station used as the experiment object is set up in Guangxi University. A photovoltaic monitoring system with more bus network is implemented on the ba- sis of the solution which is embedded the generated energy prediction function. The situation of system running and generated energy predic- tion turn out that the system has the advantage of high running efficiency, powerful expansibility and precision generated energy prediction.