Implement Of Distributed Monitoring System Combining Multivariate Statistical Process Control

Process monitoring and fault diagnosis is an important issue in process industries. It is also a key technique for safe operation and productivity improvement for process plant. Once the process monitoring using multivariate statistical process control technology is implemented with the distributed object software technique, it has many advantages over traditional centralized status monitoring technology. This thesis analyzed the multivariate statistical process and distributed object technique. About multivariate statistical process, three methods are introduced: Principal Component Analysis, Partial Least Squares, Kernel Density Estimation. About distributed object technique, the system info and development of CORBA are introduced particularly. Then, a multi-client distributed process monitoring system based on CORBA is developed w using multivariate statistical process control technology for process industries. In terms of functionality, it consists of three modular: data collection, computing server, and client graphical interface. Due to the opening characteristics of the system structure, it offers a good maintainability and extensibility.