Study On Diagnosis Of The Reasons For Vacuum Down Of Thermal Power Plant Air-cooling Condenser

In recent years,with air-cooling units increasing in China, units in operating state often appear in vacuum down, which results in units can not operate at full load and power plant economy cause losses. In order to improve the economy of air-cooling units and energy utilization ratio, the reason and location which results in direct air-cooling units vacuum reduce need finding out.This paper first analysis traditional back propagation (BP) neural network according to the questions of BP neural network, an improved BP neural network algorithm with self adaptive learning rate is proposed. Then the improved algorithm is applied to air-cooling condenser faults diagnosis. Secondly, norm temperature rising and norm terminal temperature of cooling air in condenser is obtained by formula derivation; further norm vacuum is obtained too. Take 600MW air-cooling units for example to establish variable condition mathematical model, then characteristic curve of air-cooling condenser is obtained by using computer. Compare the value of actual operation with the norm vacuum,if difference between the value of actual operation and the norm vacuum is large, which means vacuum of air-cooling condenser is abnormal. Then, through comparing the value of actual operation of temperature rising and terminal temperature with norm temperature rising and norm terminal temperature,thus analyses the reasons causing vacuum reduce. Finally, further analyses the reasons causing difference between the temperature rising and terminal temperature and diagnose the reasons and location which results in direct air-cooling units vacuum reduce, which provides basis for improving turbine output and mining energy saving potential.