Parameter Optimization Of PC Continuous Rigid Frame Bridge

The prestressed reinforced concrete continuous rigid frame bridge is a kind of bridge which is nice wholeness, bearing property and has bright and pithiness bridge style. This kind of bridge is also fit for the comfortable steer with no other expansion joints except for those at the two ends, and also has the character of non bearer and construction convenience. Mainwhile, under the vertical load, the negative moment at the cross of girder and pier may make the midspan moment shorter. In a long run, the design of prestressed reinforced concrete continuous rigid frame bridge is mainly depends on experience of more work and full experience, so it is necessary to optimise the parameter of the rigid bridges. Based on the Longtanhe Bridge of Hurong national highway in the west of Hubei province, the parameter optimization of long span pretressed reinforced concrete continuous rigid frame bridge is proceet. The main work is summarized as follow:(1) The existing work of the development of the rigid bridge is introduced, and has discussed the amount of practical parameter. The design parameters of bridge which is finish already are also studed,(2) The cross test design is used in parameter optimization of rigid bridge. The result tells us that the cross test design is good in parameter optimization and can save us much time. The parameters are the Ratio of side-span and mid-span is 0.6, the index of web-line is 1.8, the ratio of mid-span and depth of beam at the button is 19 and the space of two thin pier columns is 11.4 meters.(3) The genetic algorithm, the combination of genetic algorithm and artificial neural network and the combination of genetic algorithm, artificial neural network and cross test design are all discussed. As a result when use the genetic algorithm lonely, we will cost long time, and the genetic algorithm may be premature. When use full specimen to educate artificial neural network, we can receive a better result. The parameters are the Ratio of side-span and mid-span is 0.55, the index of web-line is 1.77, the ratio of mid-span and depth of beam at the button is 18.96 and the space of two thin pier columns is 11.37 meters. But the time we spend is long too. Using full cross test design table to educate the artificial neural network can also get the good result and shorter time. It is an efficient optimum method.