The Design About Alarm In Advance,Alarm And Fault Detection System Of Boiler Water Level In Power Plant

The development of computer and network quicken the step of digital and intelligent control system. Conventional control instrument's dominion is changed drastically. Intelligent computer control system and DCS replace it. So there is more high appeal for fault detection and alarm.In this thesis, fault detection, alarm and alarm in advance in water level system of water circulation system in bolter are studied.Combining the characteristic of water level system of in bolter, single signals were filtered. Put up a new arithmetic that has mutative coefficient. It can overcome the disadvantage of conventional arithmetic that has the settled coefficient.Based on strong association of plant's signals, find out association rules in the signals. Using the rules, we can detect and diagnose the fault of sensor, water pump.The association of the water level signal and other signals is very feebleness, so we use neural networks to detect the fault of water level sensor. Founding forecast model of control system of water level.Fuzzy math is applied in alarm in advance, put up the fuzzy arithmetic based on fuzzy reasoning. Making certain of the bound of difference of water flux and steam flux based on the water level signal. Judging if there have faults or not.Doing many experiments by Matlab, outcome is good. Realize the arithmetic by software. Get the real signal from DCS databank by net. Send out alarm and alarm in advance in computer, and put out the result of fault detection.Based on real engineering, the thesis made extensive research work on the alarm in advance and fault detection of bolter water level system in power plant. The software will be applied in power plant soon and improve operation and maintenance of power plant.