Damage Identification Method For Beam Bridge Structure Based On Curvature Mode

This paper combined the curvature mode theory and neural network theory to expl ore the bridge damage identification characteristics,and focused on identifying the locat ion of simply supported beam damage,aimed at the problem that curvature mode differe nce for minor structural damage identification inaccurate,proposed the method to explo re the tiny structure damage location method curvature mode difference combining neur al network model,extended the curvature mode difference from the one-dimensional to two-dimensional,proposed difference Bridge Damage Identification dimensional positi on based on two-dimensional curvature mode difference,and expanded this method to t he crack shape recognition.Finally,proposed solutions to the two-dimensional random d amage identification in case of temperature loads,its main idea as follows:1.For one-dimensional simply supported beam structure,firstly using curvature mo de difference to its damage identification,including single place damage,multiple dama ge and damage to identify the location of different locations,and verified its effectivene ss;for currently no quantitative lesion size conditions,put forward damage quantization concept which blend the volume of the fusion rate and damage degree,provided a basis for the quantitative analysis of minor injuries;in the case of minor damage,curvature m ode difference is not sensitive to the damage location,for this feature,proposed the met hod blend neural network classification algorithm to identify simply supported beam da mage to the tiny local location,the results show that,when the curvature mode differenc e can’t recognize the small damage,this method can effectively identify the structural d amage location.2.For two-dimensional simply supported beam structure.Firstly,extended curvature mode theory to two-dimensional simply supported beam structure;secondly.Secondly,u sing Gaussian curvature mode difference to identify two-dimensional damage location f or simply supported plate beam,the results show that it is not obvious;proposed using t he method of dimension modal curvature difference to identify damage location,using n umerical simulation to study simply supported plate beam damage identification,the res ult significantly improved than Gaussian curvature mode difference,it can accurately id entify the damage location in the same condition;explored the relationship between the two-dimensional curvature mode difference between the damage location and identification effect,research shows that the local damage remains insensitive still exist;when the t wo-dimensional curvature mode difference can’t recognize the local damage,blend the neural network classification algorithm to analyze its damage identification,the results s how that this method can effectively solve the problem of insensitivity to local damage.Finally,based on two-dimensional curvature mode difference,proposed crack shape rec ognition technology,using numerical simulation and successfully identified shape crack s.3.Combined with the effect of temperature on the elastic modulus of concrete,usin g elastic modulus reduction to simulate the effect of temperature load on the concrete.F irstly,used two-dimensional curvature mode difference to identify damage location on u niform temperature load in simply supported beam,which identification effect is obviou s;proposed impact factor method to simulate the difference of temperature load on dam age position and undamaged position and explore the Impact factor influence on identifi cation effect,when the impact factor at a reasonable range,two-dimensional curvature m ode difference can accurately identify the damage position;due to the randomness of th e distribution of temperature,used normal distribution to simulate temperature random l oad effect on a simply supported beam structure,in this state using two-dimensional cur vature mode difference to identify its damage,analysis the impact of temperature rando m discrete degree on the damage identification,when the discrete level within a certain range,two-dimensional curvature mode difference still accurately identify the damage l ocation;combined gradient curves of temperature on the outer surface and the inner surf ace of the concrete,simulated sunlight conditions,explore its effect on two-dimensional curvature mode difference damage location location identification,and the results show in the sunshine conditions,the two-dimensional curvature mode difference can identify damage location.Finally,aimed at the condition that it can’t be identify damage locatio n at the temperature random load and impact factor,blended neural network classificatio n algorithm identification,studies have shown that this method can be more accurate da mage identification under the ambiguous position.