www.bewet.fora.pl

loczytzn

Field Calibration temperature after turbine technology

Method is reasonable. 2) Field calibration technique described in this article, you can make up for lack of laboratory conditions, the calibration device, and then solve the aero-engine turbine gas temperature sensor accuracy after calibration problem. References [1] Zhao Zhennan. Heat Transfer [M]. Beijing: Higher Education Press, 2002. [2] Pan Jinshan. The basis of gas dynamics [M]. Beijing: National Defence Industry Press, 1983. M to m Mimi Mimi Mimi Chai 米米米米米米米米米米 mining meters to adopt to 米米米米米米米米米米 to Mimi Mi Mi Mimi Mimi Zhang 米米米米米 ( on page 7) 2.3 Fault pattern recognition of network training is completed, to meet the requirements of fault diagnosis system to the network structure,[link widoczny dla zalogowanych], weights, threshold values exist in the network. When the input feature value of any fault identification, through the fault diagnosis system, the system can interpret the corresponding fault type. Table 2, Table 3 shows, the normalized value after the fault identification features the end of the input has been trained neural network, results show that fault-based group number 1, the generator rotor winding inter turn short circuit, group number 2 failure the main generator stator winding insulation damage. Fault output calculated with precision in the error and allow the actual fault line within the range, so the network training is successful, has a high diagnostic accuracy. Table 2 Group No. fault identification eigenvalue eigenvalue 116.66,30.21,0.75,36.16,6.7514.05 fault identification. A 1.73, a 3.62. 0.748,2.56115.29,27.65,1.37,36.28, a 2.292-6.61, a 1.85, a 3.74 , 0.784,0.16 Table 3, the output value of the group number of fault diagnosis fault diagnosis output 10.00050343, -0.034099,0.99159,0.037682-0.18132,[link widoczny dla zalogowanych],1.0152,0.16228, a 0.0852143 article describes the elastic gradient descent algorithm is applied to neural network training, and the gradient descent method based on neural network application to fault diagnosis of air power training samples,[link widoczny dla zalogowanych], the samples through the same training, diagnosis and analysis showed that under the same conditions, neural network pattern recognition in the fault has a faster training speed. And the trained neural network applied to power system fault diagnosis, fault output calculated with the error and accuracy within the allowed range consistent with the actual fault, has a high diagnostic accuracy. References [1] Liang Shan, Liu Nian, Yue Liangshun. Wavelet neural network based brushless synchronous generator rotating rectifier fault diagnosis [J]. Sichuan Electric Power Technology, 2008,31 (3) :51-53. [2] Chen Wei-kan,[link widoczny dla zalogowanych], Pan amidships, et al. Algorithm based on improved wavelet neural network for power transformer fault diagnosis method [J]. Scientific Instrument,[link widoczny dla zalogowanych], 2008,29 (7): 1489.1483. [3] NikunjChauhan, V. Ravi. Differentialevolutiontrainedwaveletneuralnetworks [J]. ExpertSystemswithApplications, 2008,09 (19) :1-7. [4] BaskshiBR, StephanopoulosG. Wavenet: amuhiresolutionhi-erarchiealneuralnetworkwithlocalizedlearning [J]. AICHEJ, 1993,39 (1) :57-81. [5] Daubechies1. Orthonormalbaseofcompactlysupportedwavelets [J]. CommunicationsonPureandAppliedMathematics, 1988,41 (2) :909-996. [6] MastorocostasP pill Resilientbackpropagafionlearningalgorithmforrecurrentfuzzyneuralnetworks [J]. ElectronicsLettens, 8thJanuary, 2004,4 o (1) :27-31. [7] yellow root of the whole. Air power failure feature extraction and fault diagnosis [D]. Northwestern Polytechnical University, 2005.

相关的主题文章：

ugg stivali 正多面体角误差的全组合互测比较测量_417

[link widoczny dla zalogowanych]

[link widoczny dla zalogowanych]