Experimental Investigation On Monitoring Tangential Fluid Field In A Furnace With Acoustic Method Using BP Neural Network

The combustion of pulverized-coal in large-scale boilers of power station is a complex physical and chemical reaction, which is occurred under the condition of large space, frequent fluctuation and having conspicuous three-dimension character. The monitoring of combustion status is very important to safety, economy and cleaning of boilers in operation. The measurement of combustion gas velocity is a complex and important issue in combustion area. When the combustion gas status in furnace is known, the optimized control of combustion and well-organized production can be achieved. A measurement style of not touching is used during monitoring air dynamical field in a furnace with acoustic method, which is modern high-tech control means. The investigation on BP neural networks and the simulation on monitoring dynamical field in a furnace with acoustic method make acoustic monitoring technology more practical . In the article, firstly , we introduce the basic method of reconstructing furnace dynamical field with acoustic method, principle of artificial neural networks and far-ranging applications ,and discuss the study of Artificial Neural Networks and BP model, and bring forward that BP neural networks ,as a nonlinear mapping means, has a lot of advantage that traditional optimizing methods don't have .And then, we build BP network model by making a model boiler furnace as simulation object. During choosing data , network layer ,study step and momentum coefficient , we all us the method of as training network as modifying it, BP neural network that adapts to monitoring tangential fluid field with acoustic method comes into being. Finally, we do a series of simulation and analysis with trained network , and validate the network by the experiment of monitoring furnace dynamical field with acoustic method ,the result is satisfied. BP network is the most comprehensive being applied one of all the artificial neural networks, whose framework is simpler, the speed of mapping data , the stability and accuracy is also better than the other. It is feasible that we us BP neural network in monitoring tangential fluid field in a furnace online with acoustic method.