Double-layered big data analytics architecture for solar cells series welding machine

The rapid and extensive pervasion of big data has enhanced the revolution of the society. A great interest has arisen in the past five years for mining and sharing the massive data. However, implementation of the big data analysis is facing many challenges, such as the storage, transmission, and computing. How to make the decision more intelligent in latency time becomes a crucial requirement for many researches. In this paper, a double-layered architecture ATWDP of online and offline analytics for solar cells series welding machine industry is proposed and the distributed and parallel computing system can handle the above challenges. The ATWDP offers an approach to analyze the data gathered from various sensing devices stably and efficiently. Some key implementation technologies in the system are discussed, especially the Hadoop Apache framework. The evaluation of service quality based on Support Vector Machine-Dempster Shafer Theory (SVMs-DS) and Spark is an application scenario to illustrate the mechanism of the ATWDP. And a data set is used to verify the rationality of the ATWDP in the storage and processing. Test results show that the ATWDP platform has a good performance and is a suitable solution for dealing with the big data of solar cells series welding machine.