Intelligentized Analysis And Research Of Air Leakage Of Rotary Air Preheater Based On The Artificial Neural Network

Air leakage rate is the most important factor of the efficiency in the air preheater. The air leakage rate is not only affected by the manufacturing process and sealed, but also affected by the speed and temperature of flue gas and air, the static pressure of flue gas etc, which are complicated to concerned. In short term and local view, operational parameters it's showed remarkable stochastic character, in long term and global view, it's showed distinct regularity character. As a computation intelligent method, the artificial neural network (ANN) is able to identify and simulate the nonlinear relationship of the data, approximate any nonlinear function, and find out the regularity among them. These play an important role in realizing and mastering the operating state of air preheater. This dissertation simply introduces the development and application of artificial neural network, then discusses the system of test and control of rotary air preheater, puts forward the research technique and research content. Base on analyzing the basic principle and structure of biologic neural network, puts forward the concept trait and basic principle of ANN, and introduces BP network and BP algorithm in detail.Expatiates the conception, quality and characteristic of rotary air preheater of station boiler and analyzes the factors of air leakage of air preheater. Determining the model of ANN through discussing the choice of network parameters, the choice and disposal of examples, and the choice and design of network structure, and puts forward the strategies of improving BP algorithm aiming at the limitation of traditional BP algorithm. Using neural network toolbox of MATLAB6.5 models and simulates, and determines the normalized transfer function , the training function, the number of hidden layer and the active function for ANN. According to the established structure of ANN, predictes the air leakage rate . Basing on the hot test of air preheater of 50MW coal-fired boiler in Shashi Thermo-power Plant, this dissertation has simulated and modeled the air leakage rate of air preheater, using neural network toolbox of MATLAB 6.5,and achieved a preferable prediction result.