Research On Wireless Sensing And RBF Recognition Of Mountain Fire

The electric power industry of our country has developed rapidly; the grid density and scale has expanded continually. More and more power grid has across the mountains and forests that the frequent fire caused transmission line tripping accident. The accidents seriously threatened the normal operation of the transmission line, to guarantee the normal work of power transmission line becomes the key to the safe operation of power grid. In recent years, many transmission line faults occur in Yunnan power grid owing to the forest fire, and they bring great pressure to the safe operation of the transmission line.Tripping in fire conditions is the result of flame temperature, the flame electrical conductivity, the ashes and smoke caused gap insulation level dropped sharply. In order to effectively forecast and monitor fires, to reduce the impact of fires on the power transmission lines normal operation, this paper does the following researches:1. By studying the fires effects on the normal operation of the transmission line, we select several important environmental parameters (temperature, humidity, carbon monoxide, hydrogen sulfide, ion smoke and rainfall) that cause and reflect the fires as monitoring objects. Based on the working principle of several kinds of sensor and wireless sensor network, and according to the characteristics of transmission line wide distribution, monitoring regional dispersion, high cost of cable transmission, wiring and maintenance difficult, we monitor environmental parameters around the substation transmission lines adopting smart wireless sensor networks.2. According to the monitoring system requirements design and develop an online monitoring system for the power fire. Using multi-threading technology writes the communication protocol analysis procedures, and analyzes the data frame of wireless transmission. The monitoring data will be stored in the database and displayed on a real-time monitoring interface. On the basis of the actual situation and equipment installation requirements of Chuxiong 220kV waist station, install sensor network monitoring and communications equipment, build Chuxiong 220kV waist station online monitoring system of power fire, and access to electricity private network.3. According to "The people’s Republic of China forest fire meteorological grade standard (QX/T 77-2007)", select the network training samples, and substitute into supervised learning RBF neural network for training. Choosing monitoring data that Chuxiong 220kV waist station power fire online monitoring system collected into the RBF neural network, and simulate dangerous grade of fire happening.