Safety Evaluation Of In-service High-piled Wharf Based On Fuzzy Neural Network

The safety evaluation model has been established by fully considering the integrity of structure,the randomness of the main influencing factors and the objectivity of evaluation method.Firstly,analyzing the characteristics and reasons of diseases of high-piled wharf,especially of inland river frame-pier wharf,damage model of the overall inland river frame-pier wharf is established by ANSYS,and the samples of the overall horizontal ultimate bearing capacity under each damage conditions can be obtained.Secondly,the safety of the inland river frame-pier wharf under each damage conditions is evaluated based on the reliability index.Finally,a fuzzy neural network is established,and the damage cases and the corresponding evaluation results are used as learning samples to train the network so as to obtain the safety evaluation model.The main conclusions are as follows:?1?According to the analysis of present safety evaluation of in-service high-piled wharf,it is considered that more work should be done towards integrity,reliability,objectivity and artificial intelligence.Analyzing the damage characteristics of inland river frame-pier wharf,reinforcement corrosion,concrete crack,concrete spalling and concrete carbonation are the mainly diseases to be considered,and the upper members and support components of the end bent are the main considerations.?2?Under load conditions of self-weight,heaped load and mechanical load,the greater damage degree of the high-piled wharf,the mean value of the overall horizontal ultimate bearing capacity is smaller while the coefficient of variation is greater.But the probability distribution of the overall horizontal ultimate bearing capacity remains the same,which is extreme value?distribution.?3?The safety evaluation based on reliability index and fully considered the randomness of the main influencing factors shows that the change range of grading standard ?_B and ?_C are less than 1% with different statistical parameters,under load conditions of self-weight,heaped load and mechanical load,so the grading standard can be ?_B=3.25 and ?_C=3.0 directly.?4?The damage cases and the corresponding evaluation results based on the reliability index are used as learning and test samples to train the network,then the safety evaluation model with error less than 10^-4 is established.Combining the engineering example,the safety evaluation model is used to evaluate the safety of the inland river frame-pier wharf with damages of the upper members and support components of the end bent.Compared with the assessment results using fuzzy comprehensive evaluation method,the result indicates that the safety evaluation model can reduce the influence of human factors,evaluate the safety quickly and accurately,and obtain a more realistic result.