基于实验数据的燃烧室火焰传递函数优化方法

Optimization method on flame transfer function of combustor based on experimental data

为简化火焰传递函数建模过程,在实验数据的基础上,采用随机采样和优化方法对迟滞时间与增益的计算模型进行了研究。首先,建立了贫油预混预蒸发(LPP)模型燃烧室实验系统的低阶热声网络(LOTAN)模型;随后,在该LOTAN模型的基础上,基于实验数据,采用Sobol采样构建n-τ的取值空间,利用优化的方法获得了不同工况条件下的火焰传递函数模型参数;最后,基于Kriging模型对n-τ进行了重构。研究结果表明:基于该方法构建的火焰传递函数能够较准确地反映各工况下的非稳态热释放特征,代入LOTAN模型中预测得到的振荡频率与实验结果吻合得较好,最大误差不超过5%,同时,利用该方法预测的振荡燃烧临界油气比(FAR)与实验结果保持一致。

In order to simplify the modeling process of flame transfer function,the calculation model of delay time and gain was studied by using random sampling and optimization method based on the experimental data.Firstly,the low order thermoacoustic network(LOTAN)model of lean premixing prevaporizing(LPP)model combustor test system was established.Then,based on the LOTAN model and test data,Sobol sampling was used to construct the sampling space of n-τ.Finally,n-τwas reconstructed based on the Kriging model.The results showed that the flame transfer function constructed based on this method can accurately express the unsteady heat release characteristics under various working conditions;if being substituted into LOTAN,the predicted results of oscillation frequency were in good agreement with the experimental results,the maximum error did not exceed 5%,at the same time,the critical fuel air ratio(FAR)predicted by this method was consistent with the experimental results.