A novel real-time monitoring system for photovoltaic (PV) generation is presented in this paper. Internet of Things (IoT) integrated with cloud servers and terminal applications allow the remote monitoring of centralized or distributed photovoltaic systems. The proposed system could realize the networking communication of multiple nodes based on ZigBee

and upload the operational data to the cloud server via a 4G communications network. Then

the server selects three-phase current as the sample sets from the uploaded data to set up a composite current characteristic combining wavelet packet energy with waveform parameter

and establishes a fault diagnosis model based on the probabilistic neural network to analyze the health status of the PV inverter online. This method requires little sampling frequency without an extra signal and saves the cost of local devices by placing online diagnostics on the server-side. In addition

the user could query and store the running information of the PV systems through the personal computer-side web or mobile phone with remote control. This article presents the hardware design of ZigBee and the 4G module

and composition of the diagnosis model for open circuit failure of the PV inverter through the cloud server

building the user software at the application layer. The final experimental results show that the proposed system has better communication performance and higher diagnostic accuracy in real-time monitoring.