Computer Room Fire Alarm Research Based On Multi-Sensor Information Fusion

With the rapid development of Internet and Internet of Things technology,the world has entered the time of DT(Data Technology).At present,the information center of units at all levels is equipped with a computer room for data storage and transmission.Once a fire broke out in the computer room,the result would be disastrous.At present,the effective and accurate detection of fire in the computer room and the timely adoption of corresponding fire protection measures have become a hot research topic at home and abroad.However,the computer room fire is different from the traditional fire.It is mainly caused by various kinds of electrical faults.It has the characteristics of high concealment,rapid development and great destructiveness in the early stage of development.The traditional fire alarm system has a low recognition degree of early fire and disturbance in the computer room,and the alarm decision is single,and it is difficult to combine the fire state and site characteristics to put forward appropriate fire alarm decision.In order to solve these problems,this paper proposes a fire alarm model based on multi-sensor information fusion using Simple Recurrent Unit(SRU)network and fuzzy logic reasoning.It is indicated in the simulation data based on Pyro Sim that the proposed scheme can accurately identify the fire in the computer room and the hierarchical output of alarm decision,and verified in the simulation experiment of the fire in the computer room.Accordingly,the main contents of this paper are as follows:(1)According to the time sequence features contained in each fire feature quantity and in order to shorten the training time,the feature layer based on the improved SRU network is established to identify the current fire situation as smoldering,open fire or interference,and obtain the fire state of the detection area.Compared with the RNN(Recurrent Neural Network)and LSTM(Long Short Term Memory)algorithm,it proves that the proposed algorithm has a good recognition effect and short recognition time.Therefore,it is selected as the fire identification algorithm of the feature layer.(2)Due to the single alarm decision factor in the current fire alarm,the decision result cannot represent the severity of the fire.In this paper,the detected fire signal duration and the protection level of the computer room are introduced into the decision factor,combined with the probability of the identified open fire and smoldering by using fire fuzzy logic fusion to output alarm decision.It makes the fire alarm decision in the computer room more reasonable and is also conducive to the effective use of fire protection resources.(3)In order to verify that the model can be applied to the fire in the computer room,the simulation of the fire in the computer room is designed by taking the computer room of somewhere.The simulation data is brought into the model,and the simulation results show that the model can accurately identify the current fire state and output reasonable alarm decisions.(4)A fire test-bed is designed based on the fire simulation of the computer room,and a combustion experiment is carried out.The validity of the model for fire identification and alarm in the computer room is verified by the combustion experiment data,and it provides a certain reference for the fire alarm in the real computer room.