Damage Identification Under Temperature Effect Of Simply Supported Bridge Based On Neural Network

With the development of China’s reform and opening up and the technique ofbridge construction, bridge construction in our country has become powerfully. Inthe pursuit of a bridge across the ability to continuously improve, the security andstability in the process of operation have been paid more and more attention. Due tomaterial aging, increased traffic, the influence of environmental factors, the bridgestructure has varying degrees of damage, which is easy to cause catastrophicaccidents. If the damage is not identified timely, it will pose a great threat to people’slife and property safety.The research results showed that the temperature is the main influence factor ofbridge modal frequency, temperature changes can cover damage in practice, thedamage identification of the structure will be obtained by the error and wrong results.Therefore, the damage identification of bridge structure under the action oftemperature effect was carried out, and following works were conducted:(1) Based on the theory of modal vibration, the relationship between the modalfrequency and structure damage is described in detail; From several neural networkmodels, the neural network learning algorithm, neural network activation functionand the improved algorithm, a detailed introduction to the neural network theory andmethods are presentied. They can provide a theoretical reference for the subsequentnumerical simulation and indoor test.(2) Taking a simply supported beam bridge as a numerical example, the damageidentification method using modal frequency and BP neural network model underthe influence of the temperature of the bridge are proposed. In the neural networkmodel of damage location, the frequency change ratio and temperature are used asinput of neural network, the damage location can effectively be recognized; In theneural network model for damage degree, the change of modal frequency and temperature are adopted as input parameters, the damage extent identification valueis identified. Compared with the traditional identification methods, this methodconsiders the influence of temperature on the damage, which can eliminate the effectof temperature on the submerged bridge structural damage.(3) With C30concrete mixture ratio, three concrete slabs with the sizes150mm*600mm*50mm are constructed. Application of dynamic load test is used todetermine the natural frequencies of slabs at the temperature of-20℃,0℃,20℃,40℃,60℃under5different temperature conditions, so as to analyze thetemperature change of the elastic modulus of concrete and the influence on modalfrequency. The change trend of different damage conditions to explore the frequencyin the cross beam, known as the frequency of the damage location and extent ofdamage to lay the foundation. Finally, based on the neural network theory, thedamage identification method is suitable for the concrete slab and eliminating theinfluence of temperature effect.