Horng-Jang Liaw

A mathematical model for predicting the flash point of binary solutions

Abstract A mathematical model which may be used for predicting the flash point of binary solutions has been proposed and subsequently verified by experimentally-derived data, such data pertaining to an almost-ideal solution as also to highly non-ideal solutions. The results reveal that the model is able to precisely predict the flash point over the entire composition range of binary solutions for both ideal solutions and non-ideal solutions by way of utilizing the flash point of the individual components. The highly non-ideal solution like octane+ethanol exhibits the minimum flash-point behavior, which leads to the minimum on the flash point vs composition curve.

The prediction of the flash point for binary aqueous-organic solutions

A mathematical model, which may be used for predicting the flash point of aqueous-organic solutions, has been proposed and subsequently verified by experimentally-derived data. The results reveal that this model is able to precisely predict the flash point over the entire composition range of binary aqueous-organic solutions by way of utilizing the flash point data pertaining to the flammable component. The derivative of flash point with respect to composition (solution composition effect upon flash point) can be applied to process safety design/operation in order to identify as to whether the dilution of a flammable liquid solution with water is effective in reducing the fire and explosion hazard of the solution at a specified composition. Such a derivative equation was thus derived based upon the flash point prediction model referred to above and then verified by the application of experimentally-derived data.

Carbon dioxide dilution effect on flammability limits for hydrocarbons

Theoretical models to predict the upper/lower flammability limits of a mixture composed of hydrocarbon and inert carbon dioxide are proposed in this study. It is found theoretically that there are linear relations between the reciprocal of the upper/lower flammability limits and the reciprocal of the molar fraction of hydrocarbon in the hydrocarbon/inert gas mixture. These theoretical linear relations are examined by existing experimental results reported in the literature, which include the cases of methane, propane, ethylene, and propylene. The coefficients of determination (R(2)) of the regression lines are found to be larger than 0.959 for all aforementioned cases. Thus, the proposed models are highly supported by existing experimental results. A preliminary study also shows the conclusions in present work have the possibility to extend to non-hydrocarbon flammable materials or to inert gas other than carbon dioxide. It is coincident that the theoretical model for the lower flammability limit (LFL) in present work is the same as the empirical model conjectured by Kondo et al.

Relationship between flash point of ionic liquids and their thermal decomposition

Recently, ionic liquids were verified to be combustible instead of nonflammable; the contrary was thought to be true due their extremely low vapor pressure. Flash point is one of the most important variables used to characterize the fire and explosion hazards of liquids. Because of extremely low vapor pressure and decomposition at elevated temperatures, the reason for ionic liquids to be combustible should be different from that of traditionally defined liquids. The flash point of ionic liquids in relation to their decomposition was investigated in this study by the estimation of vapor pressure and by use of thermogravimetric analysis (TGA), differential scanning calorimetry (DSC), Fourier transform infrared spectroscopy (FTIR), and flash point analyzer apparatus. The ionic liquids 1-ethyl-3-methylimidazolium ethylsulfate ([Emim][EtSO4]), 1-hexyl-3-methylimidazolium chloride ([C6mim][Cl]) and 1-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ([Bmim][NTf2]), were selected as test examples. Results revealed that the flammability of ionic liquids was mainly attributed to the decomposition of the ionic liquids generating flammable substances instead of themselves vaporizing, as do traditionally defined combustible/flammable liquids. Lyons method, applied by Fox et al. to estimate the flash point of ionic liquids from the TGA decomposition temperature, was assessed using our experimental data and the data from the published literature and resulted in substantial overestimation of the flash-point values of ionic liquids, which underestimates the fire and explosion hazards of ionic liquids. This deviation is attributed to flash-point values of ionic liquids located in the second temperature range of the TGA tracing, rather than in the third analogue, as predicted by Lyons method.

Nitrogen dilution effect on the flammability limits for hydrocarbons.

Theoretical models to predict the upper/lower flammability limits of hydrocarbons diluted with inert nitrogen gas are proposed in this study. It is found that there are linear relations between the reciprocal of the upper/lower flammability limits and the reciprocal of the molar fraction of hydrocarbon in the hydrocarbon/inert nitrogen mixture. Such linearity is examined by experimental data reported in the literature, which include the cases of methane, propane, ethylene and propylene. The R-squared values (R(2)) of the regression lines of the cases explored are all greater than 0.989 for upper flammability limit (UFL). The theoretical slope of the predictive line for lower flammability limit (LFL) is found to be very close to zero for all explored cases; and this result successfully explains the experimental fact that adding inert nitrogen to a flammable material has very limited effect on LFL. Because limit oxygen concentration (LOC) could be taken as the intersectional point of the UFL curve and LFL curve, a LOC-based method is proposed to predict the slope of the UFL curve when experimental data of UFL are not available. This LOC-based method predicts the UFL with average error ranging from 2.17% to 5.84% and maximum error ranging from 8.58% to 12.18% for the cases explored. The predictive models for inert gas of nitrogen are also extended to the case of inert gas other than nitrogen. Through the extended models, it was found that the inert ability of an inert gas depends on its mean molar heat capacity at the adiabatic flame temperature. Theoretical calculation shows that the inert abilities of carbon dioxide, steam, nitrogen and helium are in the following order: carbon dioxide>steam>nitrogen>helium; and this sequence conforms to the existing experimental data reported in the literature.

Prediction of autoignition temperatures of organic compounds by the structural group contribution approach

A model to predict the autoignition temperatures (AIT) of organic compounds is proposed based on the structural group contribution (SGC) approach. This model has been built up using a 400-compound training set; the fitting ability for these training data is 0.8474, with an average error of 32K and an average error percentage of 4.9%. The predictive capability of the proposed model has been demonstrated on an 83-compound validation set; the predictive capability for these validation data is about 0.5361, with an average error of 70K and an average error percentage of 11.0%. The proposed model is shown to be more accurate than those of other published works. This improvement is largely attributed to the modifications of the group definitions for estimating the AIT instead of the type of empirical model chosen. Through the Q(2) value and hypothesis testing, it was found that the empirical model should be chosen as a polynomial of degree 3. As compared to the known errors in experimentally determining the AIT, the proposed method offers a reasonable estimate of the AIT for the organic compounds in the training set, and can also approximate the AIT for compounds whose AIT is as yet unknown or not readily available to within a reasonable accuracy.

Effect of stirring on the safety of flammable liquid mixtures

Flash point is the most important variable employed to characterize fire and explosion hazard of liquids. The models developed for predicting the flash point of partially miscible mixtures in the literature to date are all based on the assumption of liquid-liquid equilibrium. In real-world environments, however, the liquid-liquid equilibrium assumption does not always hold, such as the collection or accumulation of waste solvents without stirring, where complete stirring for a period of time is usually used to ensure the liquid phases being in equilibrium. This study investigated the effect of stirring on the flash-point behavior of binary partially miscible mixtures. Two series of partially miscible binary mixtures were employed to elucidate the effect of stirring. The first series was aqueous-organic mixtures, including water+1-butanol, water+2-butanol, water+isobutanol, water+1-pentanol, and water+octane; the second series was the mixtures of two flammable solvents, which included methanol+decane, methanol+2,2,4-trimethylpentane, and methanol+octane. Results reveal that for binary aqueous-organic solutions the flash-point values of unstirred mixtures were located between those of the completely stirred mixtures and those of the flammable component. Therefore, risk assessment could be done based on the flammable component flash-point value. However, for the assurance of safety, it is suggested to completely stir those mixtures before handling to reduce the risk.

A mathematical model for predicting thermal hazard data

Abstract A systematic procedure and mathematical model for predicting thermal behavior is proposed. This model has been verified by experimental data. The results show that the model will predict thermal hazard behavior precisely. A procedure for predicting thermal hazard data is also developed. Some examples of predicting real behavior are simulated.

The multiple runaway-reaction behavior prediction of MEK- oxidation reactions

Abstract The behavior of multiple runaway reactions is more complex and difficult to predict compared with that of a (more conventional) single runaway reaction. A model to predict the multiple runaway reactions associated with MEK oxidation is proposed here, and the prediction curves are compared with experimental data. Our study includes a systematic procedure to determine the thermodynamic and kinetic parameters of individual reactions and a kinetic model to describe multiple runaway-reaction behaviors. The results revealed that this model for the multiple runaway-reaction system can be used to predict both the time-temperature profile precisely, and the temperature-self heat rate curve approximately.