Research On Construction Safety Warning System Based On T-S Network

The new type urbanization development incorporated into the 12th five-year plan provides new opportunities for the further development of the construction industry. Construction Industry makes a significant contribution to our country's economy, but at the same time is also faced with severe challenges that the construction accidents occurred frequently. In order to prevent the occurrence of construction safety accidents, the states of the construction safety must be predicted in advance. Compared with many developed countries, the level of our country's construction safety warning is still very backward. Therefore, it is very important to adopt scientific and effective methods of construction safety warning for warning the state of construction safety.From the perspective of systems engineering, man, material, machine, environment and management is regarded as a system. By using the accident-causing theory, this paper analyzes the influence factors of the construction safety. Select main influence factors in accordance with the principle of scientific, practical, etc. Then build the early warning index system of construction safety. In the process of selecting the indexes, take a quantitative evaluation of the indicators as far as possible in order to reduce the influence of subjective factors on the accuracy of the warning results.Through the in-depth study of fuzzy neural network, compared with the advantages and disadvantages of other early warning method, found that the fuzzy neural network inherits the advantages of fuzzy system and neural network, with adaptability and autonomous learning. Therefore, the innovative application of fuzzy neural network to the construction security early warning, for reference in other subjects in the field of application, based onT-S fuzzy neural network is constructed the construction of safety early warning model. After training and testing a large number of samples, and the results show that the early warning model constructed with better accuracy and generalization.Through warning lights expressing the warning results, not only directly reflect the security status of a particular period of construction, but also help macro-control construction safety overall situation within a certain period. According to construction safety warning results, proposed some from pre-control countermeasures from technical, training, management and other aspects.Take "Zheng Gang New Strict" project for example, using the safety warning model constructed, found construction safety status is "generally safe." Against the project results, combined with the actual construction of the security features of the project, proposed specific pre-control measures and suggestions from staff, machine materials, environment, management. Proved that safety warning model based on T-S fuzzy neural network has good adaptability.