| The thesis discusses the hardware composition and software programming of the system in detail. This system, which is composed of three industrial computers and one common PC, is connected as a peer-to-peer network through a hub. We can monitor any of the others, using any of the computers. The software was developed in object-oriented programming thinking and method, and it visits database by MFC ODBC classes. And it has many functions, such as data acquisition, data processing, graph protracting, report forms deposing etc. Besides, both system monitoring and fault diagnosing are the important functions too. According to the experiences of the experts, this thesis summarizes a series of deducing rules, and then sets up a fault forecasting and diagnosing fussy neural network on line based on them. The initial weighs are not arbitrary but given, and theyl be adjusted in a small range. So the speed of the network convergence is extremely quick, which has overcome the low convergence speed of the BP arithmetic. A fault diagnosis expert system is also discussed in the thesis, which is helpful for the engineers to check out the causes of the fault, and to make further decisions. This system has applied in the Xia Hua Yuan power plant, Hebei province. It is very valuable to apply this system to practical applications.
|
PDF Full Text Request