Safe operation method of photovoltaic power generation system, training method for artificial neural network and real-time detection method in safe operation method, and real-time detection device

Abstract

The invention relates to a safe operation method of a photovoltaic power generation system. The safe operation method comprises a training method for an artificial neural network for detecting a direct current failure arc of the photovoltaic power generation system, and a real-time detection method of the direct current failure arc of the photovoltaic power generation system, wherein the steps in the real-time detection method can be completed by establishing a functional module frame and using a computer program instruction to control a computer system. In order to accurately, simply and conveniently test the direct current failure arc in the photovoltaic power generation system, direct current obtained by sampling is converted into a frequency domain, the harmonic energy of the frequency domain within a set frequency band is computed, and then the artificial neural network judges whether the direct current failure arc happens according to the harmonic energy.

Description

technical field [0001] The invention relates to a method for safe operation of a photovoltaic power generation system, which includes a training method for artificial neural networks for detecting a DC fault arc in a photovoltaic power generation system, and a real-time detection method for a DC fault arc in a photovoltaic power generation system, wherein each step in the real-time detection method It can be accomplished by establishing a functional module framework and controlling the computer system by computer program instructions. Background technique [0002] In recent years, photovoltaic power generation systems have been widely used. Most installations of photovoltaic arrays utilize long strings of high-voltage DC power sources, which increases the safety issues related to arcing. Arc is a kind of gas discharge phenomenon, and arc discharge phenomenon often contains huge energy, which poses a threat to the safety of surrounding equipment and staff. Arcs are divided i...

AI Technical Summary

Problems solved by technology

Arcs are divided into DC arcs and AC arcs according to current properties. Since the application of AC started earlier, the current detection methods for AC fault arcs are quite mature, and related AC arc fault protection devices have also entered the market. However, in photovoltaic power generation systems The properties of DC arc and AC arc are very different. DC arc is a random and unstable signal without the periodic "flat shoulder" characteristic of AC current. The traditional detection method based on waveform or time domain characteristics is no longer applicable, so The detection of DC arc in photovoltaic power generation system is relatively difficult

[0003] The early DC fault arc detection methods were identified based on the characteristics of arc luminescence, heat generation, and electromagnetic radiation. These methods generally use multiple sensors to collect arc occurrence information. The cost is high, the detection rate is low, and its application is limited.

At present, electrical characteristics such as voltage and current are used to detect DC fault arcs, and the occurrence of DC fault arcs is judged by threshold comparison. The flexibility is poor, and as the complexity of the system changes, the threshold determined after considering various factors may not be considered. Other factors lead to inaccurate detection results, it is difficult to determine an appropriate threshold

Images