An algorithm for judging flame state by using a variety of sensors is proposed. According to carbon monoxide sensor

smoke sensor and temperature sensor data

flame state is classified under the neural network algorithm. And use the L-BFGS optimization algorithm to improve the traditional BP algorithm to speed up the convergence process of neural network

and effectively improve precision of system. The experiment proves that this algorithm can effectively reduce the false alarm rate of the fire alarm system

enhance the sensitivity and reliability of the system security alarm

and realtime and effective fire warning can be carried out.