International Conference on Electrical and control Engineering, p.p. 864-867, Sept. 2011. Study on online judgement model of boiler combustion stability.

International Conference on Electrical and control Engineering, p.p. 864-867, Sept. 2011.

Study on online judgement model of boiler combustion stability

OutlineAbstractIntroductionImproving fuzzy neural networkFuzzy neural network evaluation modelFNN method that determine the weights in fuzzy Comprehensive evaluationEpilogueReferences

AbstractIn this paper, seven parameters were selected as

judgement indexes. The model of improved fuzzy neural network and single factor veto has been established that was applied to online judgement of boiler combustion stability. The function of single factor veto can realize judgments when anyone of those factors can determine the stability of boiler combustion, so the judgement model is more ration. Through the neural network learning of the system, the learning samples of module were selected, and the network is trained and verified successfully

IntroductionThe pulverized fuel boilers are the main ones in power

boilers in China. The boilers’ operating stability determines the safety and the economy of the units operating in a large degree. And now, to judge and to adjust the boiler combustion state depending on the traditional operating rules and control instruments cannot satisfy the managing requirements of the modern power plants in a large degree

Improving fuzzy neural network

Improving fuzzy neural network

Fuzzy neural network evaluation modelThe input layer nodes are joined with the input

directly. The membership function generating layer deals with the fuzzy sets classification of every input vector.

The membership grade generating function is gauss

function. Every node of the consequence layer denotes a consequence rule. The clustering algorithms can be used to cluster the input samples to get the rules that can be adjusted in practice.

Fuzzy neural network evaluation model

FNN method that determine the weights in fuzzy Due to the position of primary air ducts is different in

furnace, the influence to the stability of combustion also be different. What we call the influence means that fire detection signal, pulverized coal concentration, primary air speed, primary air temperature effect on the stability by overall the measured values . In this paper we take method of input the sub- index weighted average, removing the position error , reflecting combustion status more objectively. The network can get Objective weight distribution by repeated learning

FNN method that determine the weights in fuzzy

FNN method that determine the weights in fuzzy

Epiloguethe system reflect the practical changing

tread of he combusting steady;make on-line judgment system of the combusting stability comes true

z the judgment can be made decidedly judgement by single factor veto module when some index clearly goes bad enough to cause the combustion stability clearly decline.

References