Coastal Environment Multivariate Coupling Conditions Remaining Life Prediction

In coastal areas, due to the deterioration of many factors, the durability of concrete structures sharp decline, deterioration factors include chloride ion erosion, carbon dioxide in the atmosphere,chemical corrosion, freeze-thaw cycles and acid rain. In actual projects, due to many factors, so the data is difficult to obtain, therefore estimation of structure damage that cannot be accurate, and that could provide useful information to decision makers, missed the best maintenance of coastal structures and consolidation time., in this paper he extent of damage to the coastal structures can be by using four kinds of models and detailed experimental data, while taking advantage of the actual damage coefficients estimated life of the project, and according to this paper, the structure of the projected maintenance of coastal structures to extend the durable life of the structure..(1) In order to study the durability of concrete, we must first understand the various causes of the deterioration of silver and mechanism of erosion, and then find the appropriate countermeasures, and ultimately achieve durability of concrete structure design requirements to ensure the life of the structure. Therefore, this paper describes the impact of the coastal environment of the durability of concrete structures deterioration of various factors, and various damage mechanisms and detection methods deterioration factors in a brief description, at the same time establishing coastal environmental factors index system of concrete structure, with a species intuitive way to understand the various deterioration factors. Finally solved the problem of quantifying the extent of damage by concrete damage coefficient, provide the basis for the later resolve the remaining life of the concrete structure prediction problem.(2) After a brief introduction to the basic ideas and philosophy of gray system theory, there is a detailed discussion about basic gray prediction model GM(1,1) model. While the application GM(1,1) model is made to extend from single factor can predict the deduced multivariate prediction of GM(1,N) model. In the face of the most basic of GM(1,1) model, we were explored both cases such as periodic and non-periodic forecasts, etc, the rich variety of projections GM(1,1) mode.(3) Through the history of the neural network of the network, and gradually introduces the structure of neurons, the basic principles of neural networks and network structure. After analyzing the advantages and disadvantages of BP neural network, through and GM(1,1) model combined with each other to form a single factor gray neural network, which network is derived GM(1,N) gray neural network model, at last parameters and reverse the spread of error and the model was set to complete the gray neural network model.(4) Detailed description of the basic situation of Dalian big Northbridge, combined with the local hydrometeorological history and collecting over the years from a local bridge management parameters at the bridge on the overall damage done overall rating. Invite experts to rate the bridge deterioration factors, the main factors determining the use of principal component analysis, the formation of gray neural network. In order to solve the initial weights of the neural network initialization problem, we introduce genetic algorithm to optimize the initial weights and thresholds through the principle of genetic algorithm and calculate damage coefficient, the final calculation of bridge damage coefficient by real life.(5) Finally summarizes the advantages and disadvantages of the model to provide direction for the future of the coastal environment and damage quantification of residual life.