基于QPSO-BP 神经网络的火灾预警算法

为了进一步提高油库消防系统的安全性，针对其火灾报警信息系统进行了改进，构建基于量子粒子群算法优化BP神经网络的火灾智能预警算法，以温度、烟雾浓度以及CO 浓度数据作为神经网络的输入，以无火、明火以及阴燃火的概率作为神经网络的输出。使用量子粒子群算法优化BP 神经网络运行中随机产生的权值和阈值，加快神经网络收敛到期望误差的速度，增强全局搜索能力。通过MATLAB 软件对智能火灾预警算法的模型进行仿真，模型输出的火情概率与实际值基本吻合。设计了多传感器数据采集设备，获取火灾现场数据，输入网络模型，能够有效识别明火、阴燃火和无火情况，验证了该算法可提高消防预警系统的准确性。

Fire early warning algorithm based on QPSO-BP neural network

In order to further improve the safety of the fire protection system of the oil depot, the fire information system has been improved, and a fire intelligent early warning algorithm based on the quantum particle swarm optimization optimized BP neural network is developed, with temperature, smoke concentration and CO concentration data as the input of the neural network, the probability of no fire, open fire and smoldering fire is used as the output of the neural network. The quantum particle swarm optimization algorithm is used to optimize the weights and thresholds randomly generated during the operationof the BP neural network, accelerate the convergence rate of the neural network to the expected error, and enhance the global search capability. The model of the intelligent fire early warning algorithm was simulated by MATLAB software. In the result, the fire probability output by the model was basically consistent with the actual value, and a multi-sensor data acquisition device was designed to obtain fire scene data to further verify the effectiveness of the algorithm. Entering the experimental data into the network model can effectively identify open flames, smoldering fires, and no- fire conditions, proving that the algorithm has achieved the purpose of improving the accuracy of the fire warning system.