Time Series Analysis Based On Complex Network And Application Research In Wind Records

The research on the source location and tracking of toxic gas leakage is an important issue which is related to the social security,the national economy and the people's livelihood.Although the sustained development of science and technology,it is still difficult for flow field model to match the actual environment.And the efficiency of path tracking algorithm is also low.Therefore,we should not only improve the mobile robot perception accuracy and sniffer tracking algorithm,but also strengthen the analysis of dynamic characteristics of natural wind field.Based on the classic complex work analysis results of time series,this thesis described new method of mapping time series into directed weighted complex network.The related experimental analysis is carried out.Arnd the main contents are as follows:Firstly,in order to adapt to the actual system analysis in time series with large amount of data and meet high efficiency requirements of network analysis and network visualization,this thesis puts forward a new method of mapping time series into weighted directed complex network.The new method is then applied to Logistics mapping and a set of wind speed records.Compared to the classical visibility graph method,this new method can distinguish different time series state more efficiently.And the corresponding network statistics can characterize different types of time series signals.Secondly,in order to realize multivariate time series complex network analysis in the actual system,this thesis described a method of bridging multivariate time series into weighted directed complex network.The established method is more efficient and robust than the classical undirected unweighted complex network.By using this method for indoor and outdoor 3D wind speed time series analysis,we can effectively distinguish the indoor and outdoor 3D wind speed time series,and provide much information for the study of the intrinsic dynamic characteristics of the three-dimensional wind field.Thirdly,in order to explore the deeper information of 3D wind speed time series,this thesis uses complex network community structure mining technology to analyze indoor 3D wind speed time series.By application community mining technology to indoor three-dimensional wind speed time series in different experiment time under the same environment,the experimental results show that the communities of multi-group indoor wind field overlapped with each other.That is to say that the three-dimensional wind speed time series has similar evolution and fluctuation law.And this will be a certain guiding significance for the evolution of the indoor wind field.