Research On Fault Detection For Industrial Big Data

Process safety,product quality and environmental protection are the core objectives of the modern industrial processes.As the important part and one of the key technologies in process automation systems,fault detection plays an important and realistic role in meeting these objectives.With the extensive application of information technology in industrial system and the rapid development of computer storage technology,massive data can be collected and stored,which makes fault detection based on industrial big data possible.Consequently,the fault detection based on industrial big data has received more and more attention.In the existing detection algorithms,Q statistics and T~2 statistics are usually used to identify faults.However,these algorithms generally have high leakage rate of fault data,which cannot meet the precision requirements of the industrial system.To overcome this default,in this thesis,we design a novel detection algorithm by using the compound statistics of residual and score(CRS).Based on the proposed fault detection algorithm,a fault detection system is designed.Experiments show that the CRS algorithm has better detection accuracy than Q statistics and T~2 statistics detection algorithms.The main contents are as follows:(1)The algorithms based on the Q statistics and T~2 statistics exist fault data omission problem in fault identification.In order to solve this problem,a fault detection algorithm called CRS is designed.First,principal components analysis(PCA)is used to reduce the dimension of high-dimensional data and extract the main elements.Then the CRS is built.Based on the process,the fault detection algorithm is constructed.Finally,according to the contribution value of CRS,the contribution rate of fault attribute is analyzed.(2)Based on the proposed fault detection algorithm,a fault detection system is designed.The system uses model-view-controller(MVC)architecture,where model module is primarily used to realize data storage,view module is used for implementation of the UI of the system,and controller module is mainly used to achieve our proposed fault detection algorithm.