Large-eddy simulation (LES) is under its rapid development and is recognized as a possible second generation of CFD methods used in engineering. Large-eddy simulation of two-phase flows and combustion is particularly important for engineering applications. Some investigators, including the present authors, give their review on LES of spray combustion in gas-turbine combustors and internal combustion engines. However, up to now only a few papers are related to the state-of-the-art on LES of gas-particle flows and combustion. In this paper a review of the advances in LES of complex gas-particle flows and coal combustion is presented. Different sub-grid scale (SGS) stress models and combustion models are described, some of the main results are summarized, and some research needs are discussed.

This paper reports experimental and computational fluid dynamics (CFD) studies on the performance of microfiltration enhanced by a helical screw insert. The experimental results show that the use of turbulence pro-moter can improve the permeate flux of membrane in the crossflow microfiltration of calcium carbonate suspension, and flux improvement efficiency is strongly influenced by operation conditions. The energy consumption analysis indicates that the enhanced membrane system is more energy saving at higher feed concentrations. To explore the intrinsic mechanism of flux enhancement by a helical screw insert, three-dimensional CFD simulation of fluid flow was implemented. It reveals that hydrodynamic characteristics of fluid flow inside the channel are entirely changed by the turbulence promoter. The rotational flow pattern increases the scouring effect on the tube wall, reducing the particle deposition on the membrane surface. The absence of stagnant regions and high wall shear stress are respon-sible for the enhanced filtration performance. No secondary flow is generated in the channel, owing to the streamline shape of helical screw insert, so that the enhanced performance is achieved at relatively low energy consumption.

Hydrophobic ZSM-5 zeolite filled polydimethylsiloxane (PDMS) composite membranes with Nylon micro-filtration membrane as the support layer were prepared to separate acetaldehyde from its aqueous solution. The composite membranes were characterized by Fourier transform infrared spectroscopy and X-ray diffraction. Their structural morphology and thermal stability were also examined. The swelling study showed that the composite membranes presented higher degree of swelling in aqueous solution of acetaldehyde than in pure water at 25?C, suggesting that the membranes have stronger sorption capacity in acetaldehyde solution. The effects of ZSM-5 fill-ing content and acetaldehyde concentration on pervaporation performance of composite membranes were investi-gated. The permeation experiments at different temperatures showed that both selectivity and permeation flux of composite membranes increased with temperature. With 5%ZSM-5-PDMS/Nylon membrane at acetaldehyde mass concentration of 8% and 25?C, the separation factor of acetaldehyde/water achieved 35 and the permeation flux was 233.3 g•m?2•h?1.

A novel disphase supplying supported liquid membrane (DSSLM), containing supplying feed phase and supplying stripping phase for transport behavior of Ni(II), have been studied. The supplying supported feed phase included feed solution and di(2-ethyhexyl) phosphoric acid (HDEHP) as the carrier in kerosene, and supplying stripping phase included HDEHP as the carrier in kerosene and HCl as the stripping agent. The effects of volume ratio of membrane solution to feed solution (O/F), pH, initial concentration of Ni(II) and ionic strength in the feed solution, volume ratio of membrane solution to stripping solution (O/S), concentration of H2SO4 solution, HDEHP concentration in the supplying stripping phase on transport of Ni(II), the advantages of DSSLM compared to the traditional supported liquid membrane (SLM), the system stability, the reuse of membrane solution and the reten-tion of membrane phase were studied. Experimental results indicated that the optimum transport of Ni(II) was ob-tained when H2SO4 concentration was 2.00 mol•L?1, HDEHP concentration was 0.120 mol•L?1, and O/S was 4︰1 in the supplying stripping phase, O/F was 1︰10 and pH was 5.20 in the supplying feed phase. The ionic strength in supplying feed phase had no obvious effect on transport of Ni(II). When initial Ni(II) concentration was 2.00×10?4 mol/L, the transport percentage of Ni(II) was up to 93.1 % in 250 min. The kinetic equation was deduced in terms of the law of mass diffusion and the interface chemistry.

A novel biosorbent was developed by coating chitosan, a naturally and abundantly available biopolymer, on to activated alumina based on oil shale ash via crosslinking. The adsorbent was characterized by various techniques, such as Fourier transform infrared spectroscopy, scanning electron microscopy, thermogravimetric-differential thermal analysis, and X-ray photoelectron spectroscope. Batch isothermal equilibrium adsorption experiments were condcted to evaluate the adsorbent for the removal of Cu(II) from wastewater. The effect of pH and agitation time on the adsorption capacity was also investigated, indicating that the optimum pH was 6.0. The equilibrium adsorp-tion data were correlated with Langmuir and Freundlich models. The maximum monolayer adsorption capacity of chitosan coated alumina sorbent as obtained from Langmuir adsorption isotherm was found to be 315.46 mg•g?1 for Cu(II). The adsorbent loaded with Cu(II) was readily regenerated using 0.1 mol•L?1 sodium hydroxide solution. All these indicated that chitosan coated alumina adsorbent not only have high adsorption activity, but also had good stability in the wastewater treatment process.

Laboratory-scale experiments were carried out to evaluate the influences of operational parameters on the melt crystallization efficiency for p-cresol purification. The optimal crystallization conditions were determined: dynamic pulsed aeration at 90 L•h?1 and the cooling rate of 0.6-0.8 °C•min?1, followed by sweating at 0.2-0.3 °C•min?1 for 40 min. Results also demonstrate that the melt crystallization efficiency is sensitive to feed concentration, which highlights this technology for separation and purification of high purity products.

Oil droplets in nanometer scale which are dispersed in water cannot be separated easily. An attractive technique is carried out by electrical phenomena to demulsify oil in water emulsion. In this research, non-uniform electric field or dielectrophoresis (DEP) is applied to remove sunflower oil (which is dispersed in the water). Effects of temperature, time and voltage (using AC-electric field) were considered to get the highest DEP-force (Fdi) and the best results. The oil particles sizes with average of approximately 76 nm have been shown using a ZetaSizer Nano ZS, Model ZEN 1600 (Malvern Instrument Ltd.). The maximum separation efficiency of 85% is obtained at the optimum temperature of 38 °C and voltage of 3000 V.

In this study, spent activated carbon (AC) saturated with caramel was regenerated by using yeast and NaOH. The efficiency of regeneration was evaluated under parameters such as amount, treatment time, temperature, pH value, stirring temperature of yeast and NaOH concentration. The optimum condition for AC regeneration was 8 h for yeast treatment time, 35 °C for 0.075% yeast culture temperature, a pH value of 6 for the yeast dealing with the spent AC, 90 °C for NaOH stirring temperature of AC and 6% NaOH for washing after the spent AC was treated by yeast. Under these conditions, methylene blue (MB) adsorption was 213 mg•g?1 in comparison with 60 mg•g?1 of spent AC. The micro structure and surface area of the regenerated AC were characterized by scanning electron mi-croscope (SEM) and N2 sorption, respectively. The pore size distributions of virgin and regenerated AC were ana-lyzed by means of H-K equation, resulting in a mean pore diameter of 1.28 nm and a pore volume of 1.13 cm3•g?1. This study provides data for theoretical support of the AC regeneration technology.

Egg yolk phosphatidylcholine (EYPC) is being widely used in food and pharmaceutical industries nowadays owing to its surface activity, pharmaceutical usefulness, and so on. Common determination methods of phospholipids were based on the American Oil Chemists’ Society (AOCS) Official Method Ja7b-91, in which n-hexane/2-propanol/acetate buffer was used as the mobile phase. In order to achieve desired results, gradient elu-tion or buffer solution was used, which made the detection process more complicated. Moreover, water or buffer solution could affect the silica gel column both on its lifespan and the separation efficiency significantly. In this study, different mobile phase and detector were used to simplify EYPC analyzing process instead of using water within the mobile phase. The optimized HPLC operating conditions are as follows: pure methanol as a mobile phase, flow rate of 1.0 ml•min?1, silica gel column (250 mm×4.6 mm, 5 μm, Inertsil GLTM), column temperature 30 °C and low temperature evaporative light scattering detector (40 °C, 0.35 MPa) as used. Under this optimal condition, the linear relative coefficient of the standard curve is 0.998 and the recovery was in the range of 96.83%-101.58% with a relative standard deviation of 1.79% (n 6).

A two-feed simulated moving bed (SMB) was proposed to utilize the stationary phase more effectively. Zone I is disconnected with zone II and zone III is disconnected with zone IV. Two feeds are added to zones I and III, while two desorbents are added to zones II and IV separately. The investigation on the movement of concentration bands indicated that the tail of the strong retained solute and the front of the weak retained solute were eluted from zones I and III orderly, so the streams from zones I and III could be cut into different fractions to give the products with high purity. The two-feed SMB was used to separate α-tocopherol from its homologue mixture successfully, and could double the productivity if both feeds were identical to the feed in the conventional four-zone SMB. The solvent consumption could be reduced by reusing the middle fraction without solutes.

Numerical simulations of nanoparticle migration in a fully developed turbulent pipe flow are performed. The evolution of particle number concentration, total particle mass, polydispersity, particle diameter and geometric standard deviation is obtained by using a moment method to approximate the particle general dynamic equation. The effects of Schmidt number and Damköhler number on the evolution of the particle parameters are analyzed. The results show that nanoparticles move to the pipe center. The particle number concentration and total particle mass are distributed non-uniformly along the radial direction. In an initially monodisperse particle field, the particle clusters with various sizes will be produced because of coagulation. As time progresses, the particle cluster diameter grows from an initial value at different rates depending on the radial position. The largest particle clusters are found in the pipe center. The particle cluster number concentration and total particle mass decrease with the increase of Schmidt number in the region near the pipe center, and the particles with lower Schmidt number are of many dif-ferent sizes, i.e. more polydispersity. The particle cluster diameter and geometric standard deviation increase with the increase of Damköhler number at the same radial position. The migration properties for nano-sized particles are different from that for micro-sized particles.

A three-dimensional numerical model is presented for studying the convection-condensation of mixture with vapor in a tube with edgefold-twisted-tape inserts under transition flow. According to the diffusion layer theory and laminar species transport, a condensation model with user defined function is proposed and compared with heat and mass transfer analogy and experimental test. With the condensation model, the influences of gap width and op-erating parameters on thermal-hydrodynamics performance are simulated. As the gap width increases, convection and condensation heat transfer increase initially and then decrease, while convection heat transfer increases sharply and then decreases slightly. Increasing vapor fraction has a significant effect on condensation heat transfer but it has little effect on convective heat transfer. With the increase of inner wall temperature both convection and condensa-tion heat transfer all decrease and the ratio of condensation to total heat decrease dramatically. Increases inlet tem-perature mainly affects convection heat transfer.

Experiments were carried out to investigate the liquid flow distribution at high gas/liquid ratios in a cold model monolith bed of a 0.048 m diameter with 62 cells per cm2. Three types of distributor for the liquid distribution were used to evaluate their distribution performance. Local liquid saturation in individual channels was measured using 16 single-point optical fiber probes mounted inside the channels. The results indicate that 1) The optical fiber probe technique can measure phase distribution in the monolith bed; 2) Liquid saturation distribution along the radial direction of the monolith bed is not uniform and the extent of non-uniformity depends on the distributor design and phase velocities; and 3) The tube array distributor provides superior liquid distribution performance over the showerhead and nozzle distributors.

Homogeneous mixtures of CH4/air under moderate or intense low-oxygen dilution (MILD) combustion conditions were numerically studied to clarify the fundamental effects of exhaust gas recirculation (EGR), especially CO2 in EGR gases, on ignition characteristics. Specifically, effects of CO2 addition on autoignition delay time were emphasized at temperature between 1200 K and 1600 K for a wide range of the lean-to-rich equivalence ratio (0.2~2). The results showed that the ignition delay time increased with equivalence ratio or CO2 dilution ratio. Fur-thermore, ignition delay time was seen to be exponentially related with the reciprocal of initial temperature. Special concern was given to the chemical effects of CO2 on the ignition delay time. The enhancement of ignition delay time with CO2 addition can be mainly ascribed to the decrease of H, O and OH radicals. The predictions of tem-perature profiles and mole fractions of CO and CO2 were strongly related to the chemical effects of CO2. A single ignition time correlation was obtained in form of Arrhenius-type for the entire range of conditions as a function of temperature, CH4 mole fraction and O2 mole fraction. This correlation could successfully capture the complex be-haviors of ignition of CH4/air/CO2 mixture. The results can be applied to MILD combustion as “reference time”, for example, to predict ignition delay time in turbulent reacting flow.

In order to design the relief system size of di-tert-butyl peroxide (DTBP) storage tanks, the runaway reaction of DTBP was simulated by accelerating rate calorimeter (ARC). The results indicated that under adiabatic conditions the initial exothermic temperature was 102.6 ?C, the maximum self-heating rate was 3.095×107 ?C•min?1, the maximum self-heating temperature was 375.9 ?C, and the pressure produced by unit mass was 4.512 MPa•g?1. Judged by ARC test, the emergency relief system for DTBP was a hybrid system. Based on Design Institute for Emergency Relief System (DIERS) method, the releasing mass flow rate W was determined by Leung methods, and the mass velocity G was calculated by two modified Omega methods. The two relief sizes calculated by monograph Omega method and arithmetic Omega method are close, with only 0.63% relative error. The monograph Omega method is more convenient to apply.

Support vector machine (SVM) has shown great potential in pattern recognition and regressive estimation. Due to the industrial development demands, such as the fermentation process modeling, improving the training performance on increasingly large sample sets is an important problem. However, solving a large optimization problem is computationally intensive and memory intensive. In this paper, a geometric interpretation of SVM re-gression (SVR) is derived, and μ-SVM is extended for both L1-norm and L2-norm penalty SVR. Further, Gilbert al-gorithm, a well-known geometric algorithm, is modified to solve SVR problems. Theoretical analysis indicates that the presented SVR training geometric algorithms have the same convergence and almost identical cost of computa-tion as their corresponding algorithms for SVM classification. Experimental results show that the geometric meth-ods are more efficient than conventional methods using quadratic programming and require much less memory.

Vapor-liquid equilibrium data for the binary systems of ethyl acetate + methanol, ethyl acetate + ethanol and ethyl acetate + 1-propanol at 0.3 MPa were determined. The experimental data were verified with the point-to-point test of van Ness (1973). All these systems present an azeotropic point at 0.3 MPa that increases in ester composition for longer alcohol chains. The UNIFAC in different versions and ASOG prediction models were applied.

Engineered Corynebacterium glutamicum was constructed for L-ornithine production by disrupting genes of argF and proB to prevent the flux away from L-ornithine. Effect of the inactivation of 2-oxoglutarate dehydrogenase complex (ODHC) on L-ornithine production was also investigated. It was found that the inactivation of ODHC by knockout of the kgd gene enhanced L-ornithine production. The engineered C. glutamicum ATCC13032 (ΔargFΔproBΔkgd) produced L-ornithine up to 4.78 g•L?1 from 0.24 g•L?1 of the wild-type strain. In order to understand the mechanism of L-ornithine production in C. glutamicum ATCC13032 (ΔargFΔproBΔkgd) and find out new strategies for further enhancing L-ornithine production, the comparative proteome between the wild-type and the engineered strain was analyzed. L-Ornithine overproduction in the engineered strain was related to the up-regulation of the expression levels of enzymes involved in L-ornithine biosynthesis pathway and down-regulation of the expression levels of proteins involved in pentose phosphate pathway. The overexpression of genes in the upstream pathway of glutamate to increase the availability of endogenous glutamate may further in-crease ornithine production in the engineered C. glutamicum and the ornithine synthesis enzymes (ArgCJBD) may not be the limiting enzymes in the engineered C. glutamicum.

Biodiesel produced from crude Jatropha curcas L. oil with trace sulfuric acid catalyst (0.02%-0.08% oil) was investigated at 135-184 °C. Both esterification and transesterification can be well carried out simultaneously. Factors affecting the process were investigated, which included the reaction temperature, reaction time, the molar ratio of alcohol to oil, catalyst amount, water content, free fatty acid (FFA) and fatty acid methyl ester (FAME) content. Under the conditions at 165 °C, 0.06% (by mass) H2SO4 of the oil mass, 1.6 MPa and 20︰1 methanol/oil ratio, the yield of glycerol reached 84.8% in 2 hours. FFA and FAME showed positive effect on the transesterification in certain extent. The water mass content below 1.0% did not show a noticeable effect on trans-esterification. Reaction kinetics in the range of 155 °C to 175 °C was also measured.

A pilot-scale, pre-anoxic-anaerobic oxidation ditch was used in this study to treat municipal wastewater with limited carbon source. A novel return activated sludge (RAS) pre-concentration tank was adopted for improving the phosphorus removal efficiency and the effects of RAS pre-concentration ratio were studied. Under the optimal operational condition, the suspended total phosphorus (STP) and the total phosphorus (TP) removal efficiencies were around 58.9% and 63.9% respectively and the effluent -P was lower than 0.8 mg•L?1. The reason is that with the optimal RAS pre-concentration ratio, nitrate is completely removed with endogenous carbon source and the secondary phosphorus release is strictly restrained in the pre-anoxic tank. Therefore, the anaerobic phosphorus release and the carbon source uptake by phosphorus accumulation organisms (PAOs) in the sludge, which are ex-tremely important to the phosphorus removal process, can be fully satisfied. Furthermore, the oxidation-reduction potential is proved to be suitable for controlling the RAS pre-concentration ratio due to influent fluctuation and varied conditions. The novel modified system is also beneficial for PAO accumulation.

A simple and efficient sonochemical method was developed for the degradation of organic matter and ammonia nitrogen in azodicarbonamide wastewater. The effects of initial pH, ultrasound format and peripheral water level on the sonolysis of hydrazine, urea, COD and ammonia nitrogen were investigated. It is found that the initial pH has a significant influence on the degradation of hydrazine and ammonia nitrogen, whereas this impact to urea is relatively small. It also shows that a noticeable enhancement of ammonia nitrogen removal could be achieved in a proper intermittent ultrasound operation mode, i.e., 1/1 min on/off mode. The height difference between the periph-eral water level and the inner water level of the flask affects the efficiency of ultrasonic treatment as well.

The paper presents results of phenol oxidized under the conditions of high temperature created during collapse of cavitation bubbles. The degradation efficiency has been greatly improved by using cavitation water jets combined with H2O2 as demonstrated in laboratory tests. Various factors affecting phenol removal ratio were examined and the degradation mechanism was revealed by high performance liquid chromatography (HPLC). The results showed that 99.85% of phenol was mineralized when phenol concentration was 100 mg•L?1 with pH value of 3.0, H2O2 concentration of 300 mg•L?1, confining pressure of 0.5 MPa, and pumping pressure of 20 MPa. The intermediate products after phenol oxidation were composed of catechol, hydroquinone and p-benzoquinone. Finally, phenol was degraded into maleic acid and acetic acid. Furthermore, a dynamic model of phenol oxidation via cavitation water jets combined with H2O2 has been developed.

In present study, we report the preparation of coconut shell activated carbon as adsorbent and its application for Bi (III) removal from aqueous solutions. The developed adsorbent was characterized with scanning electron microscope (SEM), Fourier Transform Infrared (FTIR), C, H, N, S analyzer, and BET surface area analyzer. The parameters examined include agitation time, initial concentration of Bi (III), adsorbent dose and temperature. The maximum adsorption of Bi (III) (98.72%) was observed at 250 mg•L?1 of Bi (III) and adsorbent dose of 0.7 g when agitation was at 160 r•min?1 for 240 min at (299±2) K. The thermodynamic parameters such as Gibb’s free energy (?G?), enthalpy (?H?) and entropy (?S?) were evaluated. For the isotherm models applied to adsorption study, the Langmuir isotherm model fits better than the Freundlich isotherm. The maximum adsorption capacity from the Langmuir isotherm was 54.35 mg•g?1 of Bi (III). The kinetic study of the adsorption shows that the pseudo second order model is more appropriate than the pseudo first order model. The result shows that, coconut shell activated carbon is an effective adsorbent to remove Bi (III) from aqueous solutions with good adsorption capacity.

The present study reports a simple, effective and energy-efficient method to prepare γ-LiAlO2 powder as a matrix in a molten carbonate fuel cell (MCFC). In our experiments, aqueous solution based sol-gel technique was used to synthesize γ-LiAlO2. Highly dispersed AlOOH•nH2O and LiOH•H2O aqueous solutions were mixed to form a colloid mixture, which was calcined to synthesize γ-LiAlO2. Thermogravimetric analysis (TGA), X-ray diffraction (XRD), and scanning electron microscopy (SEM) were used to study the composition and morphology of the intermediate and final products. The analysis results showed that an intermediate product Li2Al4CO3(OH)12 was produced after the colloid mixture was dried at 80 °C, and highly purified γ-LiAlO2 powder with fine particle size was resulted from the subsequent calcinations. A single MCFC was assembled with a matrix of the γ-LiAlO2 pow-der. The testing results showed that the matrix performed well in preventing gas leakage.

Insoluble β-cyclodextrin polymers were prepared from β-cyclodextrin (β-CD) using epichlorohydrin (EPI) as crosslinking agent under basic conditions. The polymers were characterized by Fourier Transform Infrared (FTIR), Thermogravimetry (TG), X-ray diffraction (XRD) and TG-FTIR. The results demonstrated that the polymerization between EPI and β-CD indeed occurred, and a number of CD rings were interconnected to form a three-dimensional network. Moreover, different factors influencing the polymerization, e.g. molar ratio of EPI to β-CD, the concentration of NaOH and reaction temperature, have been investigated. The polymer prepared under the optimal conditions (the molar ratio EPI: β-CD of 44, the NaOH concentration 50% in mass, and the temperature at 65 °C) showed excellent thermal stability and insolubility in organic solvents or strong acid/base. In addition, the β-cyclodextrin polymers also presented high catalytic activity for aqueous oxidation of benzyl alcohol with hy-pochlorite as oxidant.

A statistically based optimization strategy is used to optimize the carbothermal reduction technology for the synthesis of LiFePO4/C using LiOH, FePO4 and sucrose as raw materials. The experimental data for fitting the response are collected by the central composite rotatable design (CCD). A second order model for the discharge ca-pacity of LiFePO4/C is expressed as a function of sintering temperature, sintering time and carbon content. The ef-fects of individual variables and their interactions are studied by a statistical analysis (ANOVA). The results show that the linear effects and the quadratic effects of sintering temperature, carbon content and the interactions among these variables are statistically significant, while those effects of sintering time are insignificant. Response surface plots for spatial representation of the model illustrate that the discharge capacity depends on sintering temperature and carbon content more than sintering time. The model obtained gives the optimized reaction parameters of sinter-ing temperature at 652.0 ?C, carbon content of 34.33 g•mol?1 and 8.48 h sintering time, corresponding to a dis-charge capacity of 150.8 mA•h•g?1. The confirmatory test with these optimum parameters gives the discharge ca-pacity of 147.2 and 105.1 mA•h•g?1 at 0.5 and 5 C, respectively.

Paclitaxel (PTX) is an effective anticancer drug with poor solubility in water. Recently, much effort has been devoted into alternative formulations of PTX for improving its aqueous solubility. In this study, PTX and poly(L-lactic acid) (PLLA) were co-precipitated by a supercritical antisolvent (SAS) process using dichloromethane (DCM) and the mixtures of DCM/ethanol (EtOH) or DCM/dimethyl sulfoxide (DMSO) as the solvent, with super-critical carbon dioxide as the antisolvent. The effects of solvent, solvent ratio, temperature, pressure, polymer con-centration and solution flow rate on particle morphology, mass median diameter (Dp50) and PTX loading were in-vestigated using single-factor method. The particle samples were characterized using X-ray diffraction (XRD), scanning electron microscopy (SEM), laser diffraction particle size analyzer and high pressure liquid chromatogra-phy (HPLC). XRD results indicate that the micronized PTX is dispersed into the PLLA matrix in an amorphous form. SEM indicates that the solvent and the solvent ratio have great effect on the particle morphologies, and particle morphology is good at the volume ratio of DCM/EtOH of 50/50. For the mixed DCM/EtOH solvent, Dp50 increases with the increase of the temperature, pressure, PLLA concentration and solution flow rate, and PTX loading in-creases with pressure. Suitable operating conditions for the experimental system are as follows: DCM/EtOH 50/50 (by volume), 35 ?C, 10-12 MPa, PLLA concentration of 5 g•L?1 and solution flow rate of 0.5 ml•min?1.

Candidate materials for water wall of supercritical and ultra-supercritical utility boilers, T23 and T24, were chosen as the experimental samples and exposed to oxidizing atmosphere, reducing atmosphere and oxidizing/ reducing alternating atmosphere separately. The corrosion temperature was 450-550?C. The effects of oxygen content and temperature on the corrosion in reducing atmosphere and alternating atmosphere were investigated. The scanning electron microscope (SEM) and energy dispersive spectrometer (EDS) were used to examine the corroded samples. The results show that the corrosion kinetics of T23 and T24 can be described by the double logarithmic equation and parabolic equation respectively. To describe the corrosion of materials accurately it is not sufficient to analyze the macro-mass gain and the macro-thickness of the corroded layer only, but the EDS should be applied to examine the migration depth of corrosive elements O and S. It is revealed that the corrosion becomes more severe when H2S is present in the corrosive gas. S is more active than O, and Cr can reduce the migration of oxygen but not S. The combination corrosion of S and O and pure [S] has a stronger corrodibility than pure H2S. T24 suffers the most severe corrosion at oxygen content of 0.8%. Corrosion is aggravated when the corrosion temperature is above 450 °C in the alternating atmosphere. T23 has better corrosion resistance than T24 and W contributes a lot to the corrosion resistance of T23.