Font Size: a A A

Fuzzy Inversion And Application For Three-dimensional Steady-state Heat Transfer

Posted on:2018-05-19Degree:MasterType:Thesis
Country:ChinaCandidate:Q HuaFull Text:PDF
GTID:2322330536969151Subject:Power Engineering and Engineering Thermophysics
Abstract/Summary:PDF Full Text Request
The inverse heat transfer problem refers to some unknown parameters,such as boundary conditions,thermal physical parameters and geometric shapes,etc.At present,the inverse problem of heat transfer is widely used in aerospace engineering,power engineering,mechanical engineering,construction engineering,biomedical engineering and engineering thermophysics.The study of inverse heat transfer problem of three-dimensional model is more close to the practical application of engineering than the study of low dimension.It is of great scientific significance and engineering significance to study it deeply.Due to the inherent ill posedness of inverse heat conduction problem,the current commonly used heat transfer typical research methods of inverse problems of conjugate gradient method and sequential function method,the method of initial guess value,the temperature of temperature measuring points and the measurement error has a strong dependence,it is difficult to obtain good inversion results.Based on the basic theory of fuzzy logic,decentralized fuzzy inference method has a good robustness and fault-tolerant,which can make effective decision based on imprecise and incomplete input information.DFIM provides an effective theoretical method for the study of inverse heat conduction problem.In this dissertation,the inverse problem of three-dimensional heat transfer system based on DFIM is studied:(1)The finite difference method for solving three-dimensional steady state heat transfer system.The mathematical model of heat transfer with each end face of the given boundary conditions by using the Gauss iterative method for solving the three-dimensional region is the match was the problem,determine the three-dimensional temperature field of the whole region,so as to obtain the measurement information of the boundary node temperature.(2)The overview of the basic principle and process of using CGM to solve the inverse problem,the boundary temperature distribution inversion problem in three-dimensional heat transfer plate system as an example,is established based on the conjugate gradient method heat transfer optimization algorithm of inverse problem model.By means of numerical simulation,discussed the initial boundary temperature inversion to guess value,temperature measuring points and temperature measurement error conditions on the inversion results,on this basis,summarizes the limitations of the CGM inverse problem solving three-dimensional heat transfer.(3)Based on the problems existing in the inverse heat transfer problem and the characteristics of the fuzzy inference method,a new decentralized fuzzy inference method for solving the inverse problem of the three-dimensional steady state heat transfer system is established.Based on the inversion of the boundary temperature of three dimensional plate heat transfer system.Compared with CGM,NDFI significantly reduces the requirement of temperature measurement points in the inversion process,enhances the anti-interference ability of the measurement error,and obtains a better inversion effect.(4)Based on the above mentioned NDFI method,the inversion of the temperature field of the MEA/carbon plate interface is studied.Through numerical experiments,the influence of the initial value,the number of measured points and the measurement error on the inversion results is discussed.The results show that NDFI has a lower requirement on the initial value of temperature and the number of temperature measurement points,and has strong anti-interference ability to the measurement error.The effectiveness of NDFI inversion for the temperature distribution in the MEA/carbon plate interface of PEMFC is verified.
Keywords/Search Tags:Decentralized fuzzy inference, three-dimensional steady state, inverse heat transfer problem, two-dimensional normal distribution
PDF Full Text Request
Related items