论文部分内容阅读
在燃气轮机燃烧室内,压力脉动分布和放热率脉动分布的关系可以写成第一类Volterra积分方程的形式.通过此关系,可以利用动态压力信息,重建燃烧室内一维放热率脉动的分布.在重建过程中,将积分方程离散成矩阵方程进行求解.为了减小压力测量误差对热声反问题的影响,需对求解过程进行正则化处理.文中针对求解过程应用3种不同的正则化处理方法:Tikhonov正则化方法,TSVD正则化算和共轭梯度最小二乘法,并研究不同离散化算法、不同信噪比、不同离散步长以及不同正则化算法对放热率脉动重建结果的影响.综合考虑正则化算法的稳定性和适应性,在热声反问题中应采用Tikhonov或者TSVD正则化算法.“,”In the combustion chamber of Gas Turbine, the relationship between the oscillation heat release rate and the pressure can be represented as a Volterra integral equation of the first kind. The one-dimensional heat release rate distribution can be reconstructed with the pressure perturbation by this relationship. In the reconstruction, the integral equation was discretized to matrix to be solved. To decrease the influence of the error in pressure measurement to the inverse thermo-acoustic problem, the regularization method was proposed. This paper applied three different regularization methods to the solving process, including Tikhonov regularization, TSVD regularization and conjugate gradients for least squares (CGLS), and the influence of the different discretization methods, signal-to-noise ratio, discretization step and regularization methods to the reconstruction of the heat release rate was studied. Results show that considering the stability and adaptation, the Tikhonov or TSVD should be applied in the inverse thermo-acoustic problem.