Study On Atrium Fire Detection And Smoke Exhaust Intelligent Control System

The atrium building have been developed greatly in the past years. This brings forwards new issues for the fire safety protection. Because of the special constructional features of atrium building, the fire in it is largely different from the common building. Moreover, nowdays criterion is not perfect in fire preventing, so it is an important aspect to study the exactness of early fire detction in atrium building and to study the smoke exhaust validily. The smoke filling process in large spaces is analyzed based on zone model concept. A simple model for predicting the height of smoke layer interface is deduced. On the base of analyses of the experiment data of fire in atrium building, the rule of fire smoke flowing is deduced as follows: firstly, in atrium building the temperature rise is not higher just more than 20 centigrade because of the smoke enroll great of cold air when it rising. Secondly, it just spend a few minutes for the smoke in descending from ceiling to floor because the smoke develop very quickly and in high speed. The performace, applicability and the cause of wrong alarm of fire detector were compared comprehensively base on the fire detector classification. The feasibility about how to detect complexly and how to deal with the detection result intelligently is discussed. And the fire detector──smoke detctor, gas detctor, infrared beam detector and image detctor──which are fit for atrium building is selected according to the constructure of atrium building and the feature of fire and etc.Intelligent control technology has been widely used in many fields. Neural networks technology and fuzzy control technology are avilable in processing information but each has defect. Fuzzy neural network has been comprised connected the two technology`s advantage. Input signals are first fuzzied and then processed on a neural network, which is a forword BP network with three layers. BP algorithm is used to train network parameters. Thereafter, fire is judged by means of fuzzy logic, The simulink result shows that this network can detect different standard test fire accurately and have strong anti-interference capability .The most valid method to control fire smoke in atrium is merchanical smokeexhaust. The smoke layer can been in definite height and not descend when the smoke exhaust rate is larger than that of generating smoke. The change of run smoke fan numbers and the adjust velocity by fuzzy neural control variable voltage variable frequency(VVVF) can be made. The simulink result shows that the smoke fan VVVF can validily keep the smoke layer in expected height.