基于支持向量机理论对燃气火焰稳定性的判别

Discrimination of gas flame stability based on support vector machine

针对燃气火焰稳定性的判别,搭建相应的燃气燃烧试验台,使用FX2000光谱仪采集数字信号并使用数字信号处理的相关知识和方法,通过分析燃气火焰所产生的特定自由基辐射强度的时域与频域,从中提炼出五种特征值,综合试验结果对不同工况条件下各特征值的变化进行分析。在充分考虑各特征值的相关性之后,采用支持向量机的分类方法对火焰信号进行分类与预测,结果表明使用支持向量机原理判别火焰燃烧的稳定性具有较高准确性。

In order to distinguish the stability of gas flame effectively,the corresponding gas combustion test-bed was built through theoretical analysis and experimental summary.Gas flame radiation data was collected by FX2000 Spectrometer and the radiation intensities of specific free radicals in time and frequency domain were digitally extracted five eigenvalues.Various eigenvalues under different working conditions were analyzed by the experimental results.Support vector machine method was used to classify and predict flame signals after fully considering the correlation of eigenvalues.Results showed that the principle of support vector machine had high accuracy for the stability of flame combustion.