Research On Tomography Algorithm Of Furnace Flame Temperature Field Based On Acoustic Method

Uneven combustion or flame center deviation in the furnace of boiler will lead to the decrease of furnace service life,furnace coking and furnace explosion.The measurement of furnace temperature distribution can effectively avoid these problems,so the real-time monitoring of furnace temperature field is of great significance for boiler operation.Acoustic temperature detection technology has the advantages of large measurement space,large measurement temperature range,principle of temperature measurement is simple,strong environmental adaptability and so on.In this paper,the reconstruction algorithm of acoustic temperature field is deeply analyzed and studied.The existing problems of the existing algorithms for reconstruction of temperature field are: the number of mesh division cannot be greater than the number of effective acoustic paths,the solution of regularization method is too smooth,and the difference of the maximum and minimum singular value of coefficient matrix is several orders of magnitude.Two improved regularization reconstruction algorithms are proposed.Both algorithms use radial basis function approximation to model positive problems.For the inverse problem,the first one uses the regularization method with singular value ratio method to solve inverse problem（mrtr algorithm）,and the second method uses fusion standard Tikhonov regularization The improved regularization（mttgr）algorithm for TSVD regularization fusion.The two algorithms can improve the mesh score and solve the problem of reconstruction.In order to verify the effectiveness of the two improved algorithms,five common models are simulated based on MATLAB,and four evaluation methods are used to evaluate.The results show that the two new algorithms can effectively describe the characteristics of temperature field under the same grid division,interpolation method and noise interference,and have excellent reconstruction ability and anti noise ability.Among them,MRTR algorithm is slightly better than the least square method and the singular value decomposition method based on radial basis function in reconstruction accuracy and anti noise ability,and MTTGR algorithm is obviously better than the least square method and the singular value decomposition method based on radial basis function.And in the field application,the working environment is bad and the background noise is large,so the MTTGR algorithm with stronger anti noise ability has better application value.