摘要
本设计提出了同时采用烟雾传感器、温度传感器和湿度传感器来探测火灾现场,并使用神经网络对多个传感器探测到的数据进行处理和识别。由于火灾报警器与RTU传输距离较远,而且需要多台火灾报警器与RTU的通信,所以该文采用RS-485串行总线的通信方案。实现了智能漏电火灾报警器与RTU的双向通信功能,以此组成火灾监控网。实验结果表明,该多传感器的智能火灾报警器能够比较准确的识别火灾信号、准确性较高,误报率较低,单个传感器达到了93.8%。
This design uses smoke sensors,temperature sensors and humidity sensors simultaneously to detect the fire scene.The system also uses neural networks to process and recognize the data which have been detected by multiple sensors.In communications,RS-485 standard interface can be used to ensure the communication distance,accuracy and reliability of data transmission,which achieves a two-way communication between intelligent fire alarm leakage and RTU,as composition of fire monitoring network.The experiment result shows that,the multisensor＇s intelligent fire alarm equipment can identify the fire signal accurately,the accuracy of each single sensor reaches 93.8%,and the rate of false alarm is relatively low.
出处
《电子器件》
CAS
2011年第5期600-603,共4页
Chinese Journal of Electron Devices
基金
浙江省重大科技攻关项目(2009C11043)
关键词
多传感器
火灾探测
神经网络
串行通信
multi-sensor
fire detect
neural network
RS-485 serial communication
