| The application of electric power is more and more extensive,but also brings greater safety risks,Among them,the greatest threat is fire.Relevant research data show that building electrical fire has become the fire accident with the greatest social harm,the highest casualties and the most serious property loss.And arc fault is the first major cause of electrical fire.So it can be seen that the research on arc fault is imperative.From the technical level,there are three common AC arc faults in building power distribution system: grounding,parallel and series.For the first two types of arc faults,the existing electrical protection means can effectively block them.However,there is a lack of more accurate device to detect the series arc faults,which is a great hidden danger to the safety of building power distribution.This paper takes the series AC fault arc in the building distribution system as the research object,and proposes a design scheme of an independent single-phase fault arc detector based on long and short-term memory network.In this scheme,the independent experimental platform is built to simulate the series AC fault arc under different conditions in the building distribution line firstly.Then the current harmonic content of2～31 times under the two working conditions of normal operation and the line with series fault arc are collected respectively by platform,and the data set is established to provide basic data for LSTM model construction of series fault arc.Secondly,the LSTM recognition model is constructed,compared and optimized to obtain the recognition model with the best recognition performance.The fault arc detector is designed from software and hardware,which mainly includes data acquisition module,core control module with built-in LSTM identification model,watchdog,display,alarm and communication response modules.Finally,it is verified that the design scheme can effectively identify the series AC fault arc through the experiments.This design not only provides a new idea for series fault arc detection in building power distribution system,but also provides some experimental experience for the construction and implementation of intelligent fire protection in the era of big data in China. |
