Design And Research Of Automatic Fire Alarm System In Intelligent Building

The automatic fire alarm system (FAS) is vital to the modern intelligent buildings (IB) for fire safety. As technology of the sensor, the wireless communication, the integration circuit and the micro-electronics gradually progresses, the FAS meets an excellent opportunity. The purpose of this study was to reduce the distortion alarm rate and failure alarm rate, and improve the system intelligence level of FAS. The characteristics of the fire process, the fire modeling design approach and the principle of fire detection are presented. The fuzzy neural network (FNN) algorithm is introduced to complexity fire detection. Furthermore, the wireless communication node is designed, and a wireless communications network (WCN) is set up to fill up the deficiency of wire network.According to the requirement of FAS, a two-level architecture system composed of a FAS controller and detection node is designed. The FAS controller is designed based on embedded system for achieving the fire detection algorithm and managing the detection nodes designed based on MCU for collecting and transforming the smoke concentration signal, temperature signal and CO concentration signal.Analysis of the fire signal characteristic of IB is described. The FNN intelligent algorithm is applied to fire detection which is a complex non-linear structure system. The expertise, reasoning ability of fuzzy system and learning, self-adaptability ability of neural networks are utilized to deal with fire detection. The FNN training and testing are accomplished based on the sample set supplied by expert knowledge and fire tests. The result indicates the FNN model has a satisfactory generalization ability. Furthermore, the accessional momentum and the adaptive learning rate method which can accelerate the convergence rate of the FNN are presented.The universal communication node is designed based on nRF905, and the WCN is set up and applied to FAS of IB to realize the communication between the FAS controller and the detection nodes, system integration of the FAS, and fill up the defect of wire communication networks.The FAS design and some key technologies have been completed. Research proves that FNN can improve the accuracy of fire detection, reduce distortion alarm rate and failure alarm rate, improve the intelligent level of FAS, and fill up the deficiency of domestic fire detection. The flexibility of FAS design and integration is further enhanced by utilizing the WCN, which is of great value to further research and has enormous room for growth. In conclusion, the prospect of FAS in IB and the further study and issue are proposed, respectively.