文档介绍：分类号:
密级:
公开
学号:
6720100342
单位代码:
10407
硕士学位论文
论文题目:
基于ARM的机车火灾智能报警灭火系统
研究方向
嵌入式系统及应用研究
专业名称
控制工程
研究生姓名
赵志成
导师姓名、职称
王祖麟教授
2012年3月15日
江西·赣州
摘要
目前,机车内发生火灾情况仍然比较严重,尤其是在内燃机车内,行车时内部温度比较高,起火现象常有发生,现在即使在机车内配置了自动报警系统,但所用的探测器往往只是单一的烟雾探测器,误报、漏报现象严重。
为了及时发现机车内各种火灾隐患,避免火灾事故的发生,本文在讨论传统火灾报警系统的基础上,提出了基于多传感器信息融合的火灾自动报警灭火系统。本系统主要由复合智能探测器、主控制器、灭火装置等三部分组成。系统探测器与主控制器的核心处理器采用32位ARM Cortex-M3。在机车内由于空间小,为了简化系统结构、合理布线,通信部分采用CAN现场总线贯穿整个系统,使系统的可靠性大大提高。
智能探测器采用烟雾传感器、温度传感器和CO气体传感器三种传感器实时采集现场信息,把多传感器信息融合技术应用到火灾探测器中,使探测器更加智能化。由于在机车内空间狭小,灭火装置采用了悬挂式七氟丙烷灭火装置,用继电器来实现弱电控制强电,工作人员通过控制面板来确认是否启动灭火,这使得报警系统功能更加全面化和功能化。
该系统在算法控制上,利用人工智能中神经网络BP算法,通过自主学习能力将多传感器数据进行训练,综合处理分析得出判断结果。克服了传统探测器进行单一阈值的判断问题,更加提高了系统最终判断的准确性和可靠性。利用MATLAB仿真软件对我国标准明火SH4、标准阴燃SH1等两种火灾信号进行了仿真,仿真结果表明该系统具有较强的识别能力和抗干扰能力,因此本文基于多传感器融合的智能火灾报警灭火系统有很好的可行性。
关键词:ARM;多传感器信息融合;CAN总线;报警灭火系统;神经网络
Abstract
At present,a fire occurs in a otive is still more serious, especially in bustion engine otive. A fire happens quite often because internal temperature is higher when it is driving even though it have automatic fire alarm system, the phenomenon of false alarms and missed alarms is serious for single smoke detector.
For in time and exact fire detection and avoid fire accident, this paper analyze traditional fire alarm system and present automatic fire extinguishing system, based on multi-sensor data fusion. The system posed of three parts: pound Detector, main controller. The core processor of system detectors and main controller system is made of 32-bit ARM Cortex- order to simplified system structure and reasonable wiring in the small size, ponents is based on CAN the field bus throughout the entire system, and it significantly improve reliability of system.
Intelligent Detector is made of the smog sensor , temperature sensor and CO gas sensor. It have ability of the real-time data acquisition field information and multi-sensor data fusion system applied to fire detectors. Fire extinguishing devices use the hanging FM-20