Torrefaction is a slow pyrolysis process that is carried out in the relatively low temperature range of 220-300°C. The influence of torrefaction as a pretreatment on biomass gasification technology was investigated using a bench-scale torrefaction unit, a bench-scale laminar entrained-flow gasifier, and the analysis techniques TGA-FTIR and low temperature nitrogen adsorption. A series of experiments were performed to examine the characteristics of the torrefaction process, the properties of torrefaction products, and the effects of torrefaction on gas composition, cold gas efficiency and gasification efficiency. The results showed that during the torrefaction process the moisture content of biomass were reduced, and the wood fiber structure of the material was destroyed. This was beneficial to storage, transport and subsequent treatments of biomass in large scale. For solid products, torrefaction increased the energy density, decreased the oxygen/carbon ratio, and created a more complex pore structure. These improved the syngas quality and cold gas efficiency. Combustible gases accounted for about 50% of non-condensable gaseous torrefaction products. Effective use of the torrefaction gases can save energy and improve efficiency. Overall, biomass torrefaction technology has good application prospects in gasification processes. Energy is the most important basis for economic and social development. With large-scale industrial development, the total exploitable amount of fossil fuel is declining, and environmental pollution is increasing. With the advantages of being clean and CO 2 neutral, biomass is the only renewable energy source that can fix carbon, and biomass has gradually won worldwide attention. However, as a result of its dispersion, low energy density, low bulk density and high moisture content, the costs of logistics and transport are increased. Those factors make large-scale utilization of biomass for bioenergy production inefficient and uneconomic. Consequently, it is necessary to enhance the characteristics of biomass feedstocks through pretreatment.At present, biomass pretreatments include drying, pelletisation, pyrolysis and torrefaction. While drying is a relatively *Corresponding author (email: zhoujs@cmee.zju.edu.cn) mature conventional technology, the moisture content of biomass is as high as typically about 10 wt% after drying [1]. Dried biomass will re-absorb water and start to decompose. In addition, drying has little benefit for the improvement on the properties such as low energy density and bulk density, high oxygen content and grindability. As a slow pyrolysis process at moderate temperatures under atmospheric pressure, torrefaction can solve these problems. Using torrefaction technology, the energy density and bulk density of biomass are increased, and the costs of transportation and storage reduced. Moreover, because of its high process efficiency (94%) compared with pelletisation (84%) and pyrolysis (64%), torrefaction is potentially the best method for improving the economics of the ...

Class imbalance ubiquitously exists in real life, which has attracted much interest from various domains. Direct learning from imbalanced dataset may pose unsatisfying results overfocusing on the accuracy of identification and deriving a suboptimal model. Various methodologies have been developed in tackling this problem including sampling, cost-sensitive, and other hybrid ones. However, the samples near the decision boundary which contain more discriminative information should be valued and the skew of the boundary would be corrected by constructing synthetic samples. Inspired by the truth and sense of geometry, we designed a new synthetic minority oversampling technique to incorporate the borderline information. What is more, ensemble model always tends to capture more complicated and robust decision boundary in practice. Taking these factors into considerations, a novel ensemble method, called Bagging of Extrapolation Borderline-SMOTE SVM (BEBS), has been proposed in dealing with imbalanced data learning (IDL) problems. Experiments on open access datasets showed significant superior performance using our model and a persuasive and intuitive explanation behind the method was illustrated. As far as we know, this is the first model combining ensemble of SVMs with borderline information for solving such condition.

The mechanisms of char gasification in the mixture of H 2 O and CO 2 are not clear, and some problems are yet to be solved. The Langmuir−Hinshelwood (L−H) model of char gasification in the mixture of H 2 O and CO 2 and the inhibition effect between char−H 2 O and char−CO 2 reactions are two controversial issues among these problems. This paper presented new data to elucidate these two issues. Experiments were carried out at atmospheric pressure using a modified thermogravimetric analyzer (TGA) system at various reactant partial pressures and within a temperature range of 1173−1273 K. The kinetic parameters in the L−H model were determined from pure H 2 O and CO 2 gasification (N 2 as a diluent). The experimental results showed that the L−H model based on common active site assumption is applicable to describe the experimental data and the pressure is not the reason leading to different results in validity experiments of common or separate active site assumptions. Including H 2 and CO in the reactant gas does not change the reaction mechanisms from common active sites to separate active sites either. It was also found that the reaction rate first decreases and then increases as the CO 2 concentration increases at a fixed H 2 O concentration, while the reaction rate continuously increases as the H 2 O concentration increases at a fixed CO 2 concentration. This means that the char−H 2 O reaction is inhibited by the char−CO 2 reaction. Finally, the specific surface area tests of char samples at 20% carbon conversion confirm the common active site assumption.

No abstract

In this paper, a two-echelon cooperated routing problem for the ground vehicle (GV) and its carried unmanned aerial vehicle (UAV) is investigated, where the GV travels on the road network and its UAV travels in areas beyond the road to visit a number of targets unreached by the GV. In contrast to the classical two-echelon routing problem, the UAV has to launch and land on the GV frequently to change or charge its battery while the GV is moving on the road network. A new 0–1 integer programming model is developed to formulate the problem, where the constraints on the spatial and temporal cooperation of GV and UAV routes are included. Two heuristics are proposed to solve the model: the first heuristic (H1) constructs a complete tour for all targets and splits it by GV routes, while the second heuristic (H2) constructs the GV tour and assigns UAV flights to it. Random instances with six different sizes (25–200 targets, 12–80 rendezvous nodes) are used to test the algorithms. Computational results show that H1 performs slightly better than H2, while H2 uses less time and is more stable.