The Research On Intelligence Modeling And Control Method Of Fouling In Condenser

The condenser is a kind of large-scale heat exchanger, widely used in the power,chemical, metallurgy, machinery and other industries, it makes the turbine orcompressor work after the exclusion of the steam condenses into water, reducing theexhaust steam pressure and exhaust temperature. Because cooling water is always notclean, heat exchange will occur a chemical reaction and other reasons, then thefouling will deposit on the pipe of the heat exchanger, with the condenser operating,the fouling will accumulate. Our country is a big country of energy consumption,saving energy and reducing consumption have been the theme of the development inour country. The fouling is a poor conductor of heat exchanger, generally, the heatconductivity is only ten percent of carbon steel. As soon as the fouling deposited onthe surface of heat exchanger, the heat resistance will increase, the performance ofheat exchanger will deteriorate, resulting in the unit economic benefits reducing, somake study on the condenser is necessary.First of all, the subject analyzing the predicting and controlling of fouling oncondenser at home and abroad. Elaborating the condenser fouling formation processand formation mechanism, and the influence on condenser of fouling factors，the maineffects of fouling formation, and the periodic fouling characteristics are summarized.Secondly, in order to overcome the slow convergence in the neural network ofthe generally algorithm, the subject combining the particle swarm optimizationalgorithm and genetic algorithm, the improved algorithm has quickly convergence.Due to the fouling on condenser is periodic and the process of fouling is dynamics, sothe subject using Elman neural network make model, but the standard Elman neuralnetworks for complex system identification degree is reduced. Thus put forward theupswing Elman neural network concepts, and using particle swarm optimizationgenetic algorithm optimizing its weights，the result shows that the improvingalgorithm has quickly convergence and the training speed, and compared to the othersalgorithms, but also highlight its advantages.In the end, about fouling prediction modeling on condenser, in order to controlthe fouling much better, we must optimize and control the fouling chemical cleaningparameter. To put forward several kinds of condenser cleaning technology. On thefouling cleaning, using model predictive control, intelligent algorithm to optimize improved Elman neural network method, in determining its cleaning cycle control ofchemical cleaning of the main parameters to achieve the cleaning purpose.In order to verify the validity of the proposed algorithm, by using MATLABsimulation，the results show that the algorithm is effective and feasible.