Rolling Parameters Analysis And System Implementation Based On DBN Model

In the process of dam construction in China,the quality is the most critical part.For the evaluation of rolling quality,the existing method of sampling tests cannot obtain the situation of the overall storehouse surfaces.It will affect the construction schedule and cannot carry out the quality feedback effectively.Based on the application of the real-time construction quality monitoring system,there is a new way to evaluate quality from the entire construction process of storehouse surfaces on site.We can analysis the relationship between the rolling parameters and the compaction quality when obtaining running speed,rolling times,rolling thickness and other data from the system.In this paper,we propose a method based on the deep belief network,extract feature of rolling parameters and classify the values of compaction quality.Thus,it can provide reference value for quality evaluation on the rolling compacted site.This paper mainly completes the following aspects:1.According to the present situation of quality evaluation and real-time construction quality monitoring system,we analysis the influence of rolling parameters on compaction quality and determine the factors including running speed,rolling times and rolling thickness.Through the understanding of the development process of machine learning,we choose the DBN method in deep learning,and introduce the structure and training process of DBN model in detail.2.We design a feature extraction and classification method based on DBN.First,we preprocess the rolling parameters obtained from the monitoring system.Then,we use the DBN model to extract the parameters of rolling data.Through training the model about the relationship between the rolling parameters and compaction quality,we can get the compaction quality classification model under the DBN model.At last,we use two methods,PCA and manual feature extraction method,to compare with DBN model,and find DBN model has superiority in feature extraction.3.We design and implement a rolling parameters analysis system.On the basis of DBN compaction quality classification model,we achieve the system requirements analysis,system design and system implementation.In this system,we use Spring MVC framework to design the architecture,the structure of database and detailed function models.Also,we make the necessary explanation with functional structure,network topology and timing diagrams.After the implementation of the system,user can analysis the rolling parameters and compaction quality easily and quickly.It is very easy to operate and provide the help to assess the quality of compaction on the construction site.The analysis of the rolling parameters and compaction quality based on the DBN model can provide theoretical and practical significance for the compaction quality classification on site.It can help people control the quality during the rolling process combining this rolling parameter analysis system.To a certain extent,this paper can play a good guiding role on the quality evaluation of the whole surfaces of the construction site.