Under Data Deletion Based On Spline Transform Ancient Timber-framed Building Health Assessment

Ancient wooden buildings contain China's own characteristics in the architectural design, construction technology and other aspects. In the process of long-term preservation and maintenance, because of all kinds of damage, such as long-term load, earthquake, fire, man-made, these ancient buildings suffered damage more or less, or even destroyed. Life Monitor for ancient wooden buildings has great significance for ancient timber-framed building maintenance and repair.According to the existing detection methods and expertise, combined with the actual situation, in the process of data collection will be produce data leakage phenomenon. Considering the economic and practical application of different sensors have different sample rate. In this paper, we applied fault diagnosis method to study these questions, in order to get better maintenance for the ancient wooden buildings, and then prolong the life of wooden ancient buildings. The main factors affecting the service life of wooden buildings is load factors, environmental factors, structure, material, etc. Through the use of sensor networks to get the object monitoring information, we use NDT periodic testing to obtain the target body condition,performance status and other information, on the basis of these data, we introduced into fault diagnosis method based on the data feature extraction to the wooden ancient life monitoring study. The main work has the following two points :(1) Considering different factors of economic, environmental and the sensors set the different sample rate, while traditional principal component analysis model in view of the data set is the sampling rate is consistent, therefore, the traditional principal component analysis model can't accurately to sampling rate inconsistent data set for anomaly detection. Introduced variety rate PCA fault diagnosis model, in order to improve the sensitivity of the model, adding spline transform, Multiple sampling rate is put forward based on spline transform principal component analysis model of fault diagnosis.(2) Because of the complex nonlinear relationship between the actual data, and linear relationship between for collecting data of partial least squares algorithm cannot predict very well. Therefore, the introduction of the partial least squares algorithm spline transformation, the complex nonlinear relationshipinto quasi linear relationship; additional leak due to equipment failure or human data collection, resulting in the lack of data collection, therefore on the basis of the two kinds of problems, we establish the missing data based on nonlinear partial least squares estimation algorithm model, and estimate the missing data,then to life prediction.In this paper, we provide a new monitoring methodology for timber-framed ancient architecture, and to predict the life of the ancient timber-framed buildings in the wood components. There has a research significance to effectively prolong the life of wooden buildings in the view of sustainable development.