Tomography system for measurement of gas properties in combustion flow field

This paper describes a self-designed fiber-coupled tomography system and its application in combustion diagnostics. The tomographic technique, which combines tunable diode laser spectroscopy and algebraic reconstruction technique, enables the simultaneous reconstruction of temperature and gas concentration with both spatial and temporal resolutions. The system measures a maximum diameter of 35 cm in a circular area with a minimum spatial resolution of 1 mm×1 mm and temporal response of up to 1 kHz. Simulations validate the effects of the beam arrangement and discrete grid on reconstruction accuracy, and give the optimal beam arrangements. Experiments are made to demonstrate the tomography method, and systems are constructed in laboratory and on engineering test benches.

Challenges and perspectives of combustion chemistry research

In the past decades, combustion chemistry research grew rapidly due to the development of combustion diagnostic methods,quantum chemistry methods, kinetic theory, and computational techniques. A lot of kinetic models have been developed for fuels from hydrogen to transportation fuel surrogates. Besides, multi-scale research method has been widely adopted to develop comprehensive models, which are expected to cover combustion conditions in real combustion devices. However, critical gaps still remain between the laboratory research and real engine application due to the insufficient research work on high pressure and low temperature combustion chemistry. Besides, there is also a great need of predictive pollutant formation model. Further development of combustion chemistry research depends on a closer interaction of combustion diagnostics, theoretical calculation and kinetic model development. This paper summarizes the recent progress in combustion chemistry research briefly and outlines the challenges and perspectives.

Machine learning for combustion

Combustion science is an interdisciplinary study that involves nonlinear physical and chemical phenomena in time and length scales,including complex chemical reactions and fluid flows.Combustion widely supplies energy for powering vehicles,heating houses,generating electricity,cooking food,etc.The key to study combustion is to improve the combustion efficiency with minimum emission of pollutants.Machine learning facilitates datadriven techniques for handling large amounts of combustion data,either obtained through experiments or simulations under multiple spatiotemporal scales,thereby finding the hidden patterns underlying these data and promoting combustion research.This work presents an overview of studies on the applications of machine learning in combustion science fields over the past several decades.We introduce the fundamentals of machine learning and its usage in aiding chemical reactions,combustion modeling,combustion measurement,engine performance prediction and optimization,and fuel design.The opportunities and limitations of using machine learning in combustion studies are also discussed.This paper aims to provide readers with a portrait of what and how machine learning can be used in combustion research and to inspire researchers in their ongoing studies.Machine learning techniques are rapidly advancing in this era of big data,and there is high potential for exploring the combination between machine learning and combustion research and achieving remarkable results.

Emission factors of polycyclic aromatic hydrocarbons from domestic coal combustion in China

Domestic coal stove is widely used in China, especially for countryside during heating period of winter, and polycyclic aromatic hydrocarbons （PAHs） are important in flue gas of the stove. By using dilution tunnel system, samples of both gaseous and particulate phases from domestic coal combustion were collected and 18 PAH species were analyzed by GC-MS. The average emission factors of total 18 PAH species was 171.73 mg/kg, ranging from 140.75 to 229.11 mg/kg for bituminous coals, while was 93.98 mg/kg, ranging from 58.48 to 129.47 mg/kg for anthracite coals. PAHs in gaseous phases occupied 95% of the total of PAHs emission of coal combustion. In particulate phase, 3-ring and 4- ring PAHs were the main components, accounting for 80% of the total particulate PAHs. The total toxicity potency evaluated by benzo[a]pyrene-equivalent carcinogenic power, sum of 7 carcinogenic PAH components and 2,3,7,8-tetrachlorodibenzodioxin had a similar tendency. And as a result, the toxic potential of bituminous coal was higher than that of anthracite coal. Efficient emission control should be conducted to reduce PAH emissions in order to protect ecosystem and human health.

Untrained neural network for linear tomographic absorption spectroscopy

Linear tomographic absorption spectroscopy(LTAS) is a non-destructive diagnostic technique widely employed for gas sensing.The inverse problem of LTAS represents a classic example of an ill-posed problem. Linear iterative algorithms are commonly employed to address such problems, yielding generally poor reconstruction results due to the incapability to incorporate suitable prior conditions within the reconstruction process. Data-driven deep neural networks(DNN) have shown the potential to yield superior reconstruction results;however, they demand a substantial amount of measurement data that is challenging to acquire.To surmount this limitation, we proposed an untrained neural network(UNN) to tackle the inverse problem of LTAS. In conjunction with an early-stopping method based on running variance, UNN achieves improved reconstruction accuracy without supplementary training data. Numerical studies are conducted to explore the optimal network architecture of UNN and to assess the reliability of the early-stopping method. A comparison between UNN and superiorized ART(SUP-ART) substantiates the exceptional performance of UNN.

Volumetric imaging of flame refractive index, density, and temperature using background-oriented Schlieren tomography

Volumetric imaging represents one of the major development trends of flow diagnostics due to both the advancement in hardware and the requirement for more information to further understand complicated turbulent and/or reactive flows. Backgroundoriented Schlieren tomography(BOST) has become increasingly popular due to its experimental simplicity. It has been demonstrated to be capable of simultaneously recovering the distributions of refractive index, density, and temperature of flows.However, its capability in thermometry has only been demonstrated under the axisymmetric assumption, which greatly limits its applicability. In this work, we dedicated to developing a cost-effective BOST system for the simultaneous retrieval of refractive index, density, and temperature distributions for the asymmetric flame. A few representative tomographic inversion algorithms were assessed as well. Both numerical and experimental demonstrations were conducted and the results show that our implemented BOST can successfully reconstruct the three-dimensional temperature distribution with a satisfactory accuracy.