Assessing the effect of primary air conditions on flame temperature zones during stacked biomass combustion using novel computer vision method

The flame formation is an indicative parameter during combustion and is used with imaging techniques to collect the parametric flame characteristics. Image processing and advanced imaging techniques have improved the flame data extraction and utilization. Flames, being a multi-temperature formation, affect the target load at different heat transfer rates, corresponding to different flame temperature regions. The multi-temperature zones of flame could never be assessed only through imaging techniques. Therefore, the present study presents a novel method of integrating computer vision with thermal imaging to deduce the variation in the in-situ flame area with changes in temperature within the flame. The technique has been depicted using Python and verified through multi-experimental studies with changes in primary air conditions for highly dynamic stacked biomass combustion. During natural draft (ND), forced draft (FD) and forced draft with pre-heated air (AFD) combustion, burn rates of 0.14 g/s, 0.22 g/s and 0.23 g/s, respectively, were obtained. Flame regions corresponding to 200–300 °C occupy the highest flame area of 2.23 × 10⁻² m² (300 s), 2.21 × 10⁻² m² (300 s) and 2.62 × 10⁻² m² (600 s) during ND, FD and AFD, respectively. Although cumulative flame area under ND remains highest (1.88 × 10⁻¹ m²), flame area during AFD condition has the highest peaks of 2.62 × 10⁻² m², 2.28 × 10⁻² m², 1.09 × 10⁻² m², 6.97 × 10⁻³ m², 5.01 × 10⁻³ m², 3.71 × 10⁻³ m², 1.36 × 10⁻⁴ m² and 3.60 × 10⁻⁷ m² corresponding to 200–300 °C, 301–400 °C, 401–500 °C, 501–600 °C, 601–700 °C, 701–800 °C, 801–900 °C and 901–1000 °C respectively. For all cases, the flame area decreases with increased flame temperature, and high-temperature zones remain closer to the combusting particles.