Inverse diffusion
flame (IDF) is an effective and widely used reaction
form in the process of noncatalytic partial oxidation (NC-POX) of
gaseous hydrocarbons (such as natural gas and coke oven gas). However,
soot is generated in the combustion chamber in the case of unreasonable
feeding conditions, and thus causes serious damage to the wall and
nozzle. In this study, the effects of the equivalence ratio ([O/C]
e
), the oxygen flow rate, and the Reynolds number on the soot
and CH* emission characteristics of CH
4
/O
2
inverse
diffusion flame were comprehensively analyzed based on a hyperspectral
imaging system. In addition, the relationship between CH* and soot
is explored using Ansys Fluent simulation. The experimental results
show that the soot radiation core generation area is located in the
outer ring of the flame, and the radial distribution of the radiation
intensity is bimodal. With the increase in [O/C]
e
, the
initial position for soot radiation and the overall radiation intensity
of soot decrease. In addition, the CH* radiation intensity decreases
as [O/C]
e
increases, and CH* exists in the whole flame.
The simulation results clearly show that the existence of CH* is conducive
to soot production. The emission intensity and the core area of soot
formation increase with the increase in the oxygen velocity. Additionally,
the soot emission region increases and the flame tip changes from
a round blunt to symmetrical tip with the increase in the Reynolds
number.
This
work studies the single coal particle combustion process in
an O2/CO2 atmosphere based on spectral diagnostics
technology in a visual drop tube furnace (VDTF). The chemiluminescence
characteristics of OH*, CH*, Na*, and K* in single coal particles
at two reaction stages and different oxygen fractions (X
i,O2
) are investigated. The results show that
both flame temperature and alkali metal spectral intensity increase
first and then decrease, and the trend of temperature variation is
consistent with that of alkali metal spectral intensity. Meanwhile,
with X
i,O2
increases, Na* and
K* peak intensities are enhanced because of the improvement in the
oxygen flow rate. In the volatile reaction stage, the coal particle
shows a typical envelope flame with high soot generation and luminosity.
As the reaction time increases from 10 to 30 ms, the flame size increases
and the flame temperature increases from 1550 to 1950 K. The spectroscopic
results corresponding to the volatile reaction stage show that, as X
i,O2
increases from 30 to 50, the
OH* and CH* intensity peaks increase linearly. Moreover, the intensities
of the Na* and K* peaks increase by approximately 50.2 and 89.2%,
respectively. During the volatile–char reaction stage, the
coal particle exhibits a brighter luminous characteristic. As the
reaction time increases from 40 to 70 ms, the flame size reduces and
the flame temperature decreases from 1900 to 1700 K. The spectroscopic
results corresponding to the volatile–char reaction stage indicate
that, with the increase in X
i,O2
, the positions of the OH* and CH* peaks change little and the intensities
of the Na* and K* peaks increase by 21.3 and 75.1%, respectively.
Our results prove that the flame temperature and alkali metal atomic
emission spectroscopy exhibit the same trends as a function of the
reaction time, and the alkali metal atomic emission spectroscopy can
be used to characterize the flame temperature.
In order to clean the mislabeled images in the esophageal endoscopy image data set, we designed a new neural network VGG_NIN. Based on the new neural network structure, we developed a method to clean the mislabeled images in the esophageal endoscopy image data set. To verify the effectiveness of the proposed method, we designed two experiments using 3835 esophageal endoscopy images provided by West China Hospital of Sichuan University. The experimental results showed that the proposed method could clean about 93% of the mislabeled images in the data set, which was the first time in the cleaning of esophageal endoscopy image data set. Finally, in order to verify the generalization ability of this method, we cleaned the Kaggle open cat and dog data set, and cleaned out about 167 mislabeled images. Therefore, the proposed method can effectively screen the mislabeled images in the esophageal endoscopy image data set and has good generalization ability, which can provide great help for the development of high-performance gastrointestinal endoscopy image analysis model.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.