TY - JOUR
TI - Control method of mechanical smoke emission in high-rise building corridor
AU - Chen Can
JN - Thermal Science
PY - 2021
VL - 25
IS - 6
SP - 4099
EP - 4106
PT - Article
AB - The traditional method has a large control error in the corridor mechanical  smoke control method. Therefore, a multi-task convolutional neural  network-based high-rise building corridor mechanical smoke control method is  proposed. Through the mechanical smoke exhaust principle of high-rise  building corridors, the threshold of mechanical smoke exhaust is set to  predict the mechanical smoke exhaust volume of high-rise building corridors.  The movement of mechanical smoke in high-rise building corridors is  simulated according to fire dynamics simulator to determine the turbulence  state of mechanical smoke in high-rise building corridors. Input the  mechanical smoke exhaust data of high-rise building corridors into the  multi-task convolutional neural network to complete the mechanical smoke  exhaust control of high-rise building corridors. Experimental results show  that the maximum accuracy of this method is about 98%, and the control time  is always less than 1 second.