DEPENDENCE OF PYROLYSIS RATE OF COAL ON TEMPERATURE

Pyrolysis process of coal has been researched to define kinetic constants which can be used for design and optimization of different processes of fuel transformation. The article considers anthracite powders and bituminous coal of Krasnogorsky mine with the use of nonisothermal thermogravimetric analysis with mass spectrometry. Spectroscanning microscopy and laser diffraction for definition of sizes and forms of particles distribution has been done. Other parameters – carbon content, ash and volatiles, density and moisture have been defined by standard methods. Energy of activation and pre-exponent with the use of models of Freedman, Starink and distributed activation energy model (DAEM), and also relative deviation of design data from experimental ones have been designed. The results of the analysis have shown the important influence of volatiles content and coal transformation degree on maximum reaction rate. Energy activation values received with the help of DAEM model are higher than with Freedman and Starink models. Process of pyrolysis of bituminous coal has a big rate in comparison with anthracite, and is better described by the abovementioned models.


Introduction
World resources of carbonaceous gas and liquid fuels are limited.With the same existing speed of energy consumption growth and ease of access deposits depletion oil and gas extraction will become more and more expensive: values EROI (Energy Return On Investmentratio of received energy to investments) of oil and natural gas extraction decreased from 26 to 18 for the last 6 years in the world, while it almost did not change for coalnot less than 50 [1].This makes use of technologies of deep coal transformation more and more attractive.Accordingly for the last time interest to technologies using solid fuel has increased in scientific literature.Ecological problems and growth of costs on solid fuel led to necessity of its utilization efficiency growth [2], with less amounts of pollution and low disposal costs [3].Understanding of processes which occur during fuel transformation is exceptionally important for fulfillment of these conditions.
Pyrolysis is thermochemical process of volatiles deletion from coal and partial oxidation which strongly influences final value of coal transformation degree and mass decrease rate [4].A lot of chemical and physical processes occur at this stage: drying, volatiles output, cracking and burning.The same factorstemperature, pressure, oxidation environment content influence their efficiency differently, which makes the research process more difficult.Other parameters of pyrolysis are a Corresponding author : madiyar.beysekov@mail.rustrongly influenced by parameters of burning, gasification and other transformational processes [5].Nowadays apparatuses with chamber division for different processes have being worked out which makes study of these processes more important [6]: pyrolysis, gasification and burning.Thermogravimentic analysis is one of the most spread methods of receiving pyrolysis process parameters.It allows receiving necessary data with precision and accuracy.There are a lot of mathematical models for description of such processes.The most famous models are Freedman and Staring models and distributed activation energy model (DAEM) based on Arrhenius equation: where vmass loss rate, mass.%/s;Аpre-exponent (frequency factor), mass.%/s;Е -energy activation, J/mole; Rgas coefficient, J/(mole•К); Тtemperature, К.
Analysis of pyrolysis processes of Kuznetsky anthracite and bituminous coal in argon atmosphere was made with the help of differential thermal analysis.Three groups of kinetic constants have been defined for three non-isothermal models -Freedman, Starink and DAEM.Conclusions on models adequacy have been made on the basis of comparison of values presented in literature.

Experimental part 2.1 Preparation of samples and characteristics
Coal powders of Krasnogorsky mine, Kemerovo region have been researched.Powders have been received as a result of coal grinding and its screening through screen plate with the size of cells 315, 150 and 80 mkm.Bulk and volumetric densities have been defined with the help of pycnomentry with five parallel changes as long as it strongly influences pyrolysis rate [7,8].Relative measuring inaccuracy of density did not exceed 0.5%.Moisture, ash and volatiles output have been defined by standard methods according to ISO 1171-81, 602-83 and 562-81 (table 1).Values of both ash content and moisture are relatively low while volatiles content in bituminous coal is higher than in anthracite.Most models have hypotheses in relation to the form of particle and powder.Stereo scanning microscopy has been made for definition of this form with the help of raster type scanning microscope JEOL 6000 С (fig.1).Large particles of bituminous coal powder do not have a correct form, anthracite spherical.The form of small particles is spherical for both samples.Analysis of particles distribution has been made with the help of laser diffraction HELOS in accordance with ISO 13320.Average size of particles and some other values -Х 10 , Х 50 , Х 90are presented in table 1.

TG-analysis and mass spectrometry
TG-analysis was made with the help of synchronous differential thermal analyzer Netzsch STA 449 C (Germany) at heating rate 10К/min in the range of temperatures from 323 to 1273 К. Argon of high refinement degree В was used as blowout gas for the process of pyrolysis.During the experiment for excluding the possibility of chemical reactions appearance with residual air in chamber TG, it was blown out with argon with volume flow rate 250 ml/min during 1 hour.After the beginning of the experiment volume flow rate made 50 ml/min.Mass spectrometry was made with Netzsch QMS 403 D Aëolos.Inaccuracy of mass decrease and temperature made less than 0.5 mass.% and 1 K, accordingly.

Kinetic models
Non-isothermal kinetic research of burning and pyrolysis processes problems is extremely difficult because of a large amount of components and their parallel and successive reactions [9].At kinetic analysis of solid fuels process rate is usually expressed as follows: where α=(m(t)-m a )/(m 0 -m a )coal transformation degree; m(t), m 0 and m amass of samples in a moment of time t, initial moment of time and after the pyrolysis process, g; βsample heating rate, K/min; f(α)dependence of transformation rate on transformation degree (function of reaction rate distribution considering change of properties and structure of fuel particles during the process).
Energy activation (Е), pre-exponent (А) and function of reaction rate distribution [f(α)]kinetic triplet.Non-isothermal kinetic methods of Freedman, Starink and DAEM were used for getting energy activation of burning processes and pyrolysis reaction.
Freedman model can be expressed with the equation [9]: Starink equation is as follows: ln(β/T At similar transformation degree at different heating rates, energy activation can be defined according to decline of dependence curve ln(β dα/dT) for Freedman ln(β/T 1.92 ) for Starink and for DAEM method ln(β/T 2 ) from 1/T.DAEM model is thought to be the most accurate among the abovementioned ones but use of simplified integral approximation for temperature can cause a substantial mistake.

Parameters of pyrolysis
Dependences of mass and its decrease rate on ambient temperature are in fig. 2 а, b.Some peaks can be noted in the graphs: in the range of temperatures 330 -375 K with maximum rate 0.138 mass.%/s and 0.868 mass.%/s for anthracite and bituminous coal accordingly and at 750-780 К with 0.132 mass.%/s and 0.362 mass.%/s.The first peak can be connected with adsorbed moisture deletionthis can be proved with corresponding peaks in graphs of mass spectroscopy.Mass decrease rate increase does not have a sharply defined peak for the second time and has a bimodal character for thin coal.It is probably connected with some chemical reactions behavior such as hydrogen containing elements output (hydrogen, methane, hydrocarbon etc.), and also compounds of full and partial oxidation of sulfur, nitrogen and carbon.Oxidizing of substances of solid phase arises as a result of elements interaction in solid phasesulfur, carbonwith oxygen in composition of adsorbed gases, and also air inflow is also possible into the chamber of differential chamber analysis analyzer.Mass spectrometric analysis shows the presence of a lot of compounds which can be the result of oxidation reactions or adsorbed gas output.At design with formula 1, fixed value of energy activation reaction is often used E=const.In this case the most accurate result will be achieved at reaction constant use defined by Starink model as its change at different values α is the lowest one for both anthracite and thin coal.Average values of energy activation and pre-exponents are presented in table 2. Pre-exponent values for each bituminous coal increase at energy activation growth, and for anthracitedecrease.It is connected with both a bigger intensiveness on the whole and with influence increase of porous structure of fuel particles on process kinetics.The received data are coordinated with values presented in literature for analogous coals [4][5][6].

Conclusion
Energy activation values for pyrolysis process of anthracite and bituminous coal of Krasnoborodsky mine have been defined as a result of the research, two main phases of coal pyrolysis process and maximum values of mass decrease for each of them have been defined.Energy activation values of two coals with the use of three kinetic models have been defined.For DAEM model decline angle is far bigger than for Freedman and Starink models.Starink model is the most preferable for design with the use of energy activation constant value.

C
Owned by the authors, published by EDP Sciences,

Figure 2 .
Figure 2. TG data (а) and dependence of coal samples pyrolysis rate (b) on time in argon atmosphere at temperature 1000 °С

Fig. 3 Figure 3 .
Fig.3illustrates dependence of energy activation on coal transformation for different samples and kinetic models.Let us consider that energy activation calculated with DAEM model is linearly dependent on transformation while it is close to constant for equations of Freedman and Starink.

Table 1 .
Main characteristics of researched samples of solid fuel 01053-p.2

Table 2 .
Average values of kinetic constants of pyrolysis process samples It is clear from the table that use of DAEM model gives the highest values of energy activation, and Starink modelthe lowest ones.