A new correlation accurately predicts the flue-gas sulfuric-acid dewpoints to mitigate corrosion problems in process equipments and heat-recovery systems.
Acidic combustion gases can cause rapid corrosion when they condense on pollution control or energy-recovery equipment. This accurate and reliable correlation predicts sulfuric-acid dewpoints over wide ranges of sulfur trioxide and water vapor concentrations.
The correlation outperforms other available correlations both in accuracy and generality. The predicted flue-gas sulfuric-acid dewpoints are in excellent agreement with experimental data with the root mean square deviation of 0.73.
Flue-gas SOx
Sulfur compounds such as H2S, methyl mercaptan (CH4S), and sulfur (Sx) are present in industries chiefly as undesirable by-products of fossil fuel processing, including natural gas and petroleum. These pollutants burn in the incinerator to produce sulfur oxides in the flue combustion gas.1 2
Flue gas always contains substantial water vapor. Sulfur trioxide and water have a great affinity for each other: When temperatures are lowered to the dewpoint, the two combine rapidly to form sulfuric acid. The condensed sulfuric acid also has a powerful affinity for water to the extent that the concentrations of sulfuric acid occurring at elevated dewpoint temperatures are corrosive to steel and almost all plastics, as well as hydraulic cement composites.3
Further, if the gas is cooled below this dewpoint by radiation or convection, a mist of corrosive acid droplets forms that is highly detrimental to the stack and heat-recovery exchangers.3 4 Fig. 1 shows the trend of corrosion rate vs. wall temperature as the flue-gas temperature falls below the sulfuric-acid dewpoint.
Verhoff and Banchero provided the correlation (Equation 1 in the accompanying box) for predicting flue-gas sulfuric-acid dewpoint.5 It should be noted that there are some disagreements between experimental data and the Verhoff and Banchero correlation, especially at low SO3 concentration and high H2O content.
Dewpoints predicted in the range of 120-140° C. have a positive deviation of 4° C. and more. Also in the range of 100-121° C., the predicted dewpoints are 2.5-4° C. low.
Okkes proposed a correlation to overcome some of these shortcomings that can be written as Equation 2, in which the partial pressures are expressed in atmosphere and the dewpoint is in °C.6 Although this correlation is more accurate at H2O concentrations higher than 25%, but it significantly under predicts the acid dewpoints at low H2O concentrations prevailing in the oil and gas industry.
New correlation
This work proposes a new correlation (Equation 3), based on all verified experimental data, for accurate prediction of flue-gas sulfuric-acid dewpoints. A set of 188 validated data points7 has been used to derive the correlation.
Fig. 2 uses Equation 3 to predict flue-gas sulfuric-acid dewpoints and compares them with experimental data at different SO3 and H2O concentrations in Fig 2. As shown, the acid dewpoint is very sensitive to the flue-gas SO3 concentration so that a small increase in SO3 concentration leads to a large increase in acid dewpoint at a given H2O concentration.
The effect of moisture concentration on acid dewpoint is moderate, however, especially at high SO3 concentrations. As shown, the SO3 species has a very strong influence on sulfuric-acid dewpoint at SO3 concentrations of less than 100 ppm (vol). Thus accurate prediction of acid dewpoints, especially at low SO3 concentrations prevailing in petroleum industries, is very important to mitigate corrosion problems in process equipments.8
Flue gas always contains a substantial amount of water vapor. Also, the added moisture during turbine inlet-fogging operations affects the sulfuric-acid dewpoint temperature and may thus affect allowable metal operating temperature. At a low SO3 concentration of 1 ppm (vol), an increase of H2O concentration from 5 to 30 vol % leads to a 19.8% increase in acid dewpoint to 127° C. from 106° C., while at a high SO3 content of 500 ppm (vol), the same H2O increment results in a 7.5% increase in sulfuric-acid dewpoint to 186° C. from 173° C.
At moisture concentrations higher than 30 vol %, the H2O content does not have a major influence on the sulfuric-acid dewpoint, as shown in Fig 2. The acid dewpoints predicted by presented correlation are in good agreement with experimental data, as shown in this figure.
The acid dewpoints predicted by the presented correlation are compared with the predictions of Verhoff and Banchero and of Okkes correlations in Figs. 3 and 4, respectively. As shown there, the V.B. correlation overpredicts the experimental data, while the Okkes correlation under predicts the sulfuric-acid dewpoints significantly.
Using Verhoff and Banchero's correlation leads to a considerable acid dewpoint overprediction especially at moisture contents higher than 15 vol %. In such cases the designer may incorrectly increase the air preheating level to combat the cold-end corrosion problem so that excess energy is wasted.
The dewpoints predicted by Equation 3 are in excellent agreement with measured sulfuric-acid dewpoints at all ranges of SO3 and H2O concentrations, as shown in Figs. 3 and 4. The accurate prediction of flue-gas sulfuric-acid dewpoint is important for optimization of energy consumption in combustion devices.
For better comparison of the proposed equation with previous ones, the set of new data has been used for refitting the V.B. and Okkes correlations and the modified versions of two previously correlations are obtained as shown in Equations 4 and 5. The root mean square deviation, as defined by Equation 6, is calculated for Equation 3, V.B. (Equation 1), Okkes (Equation 2), modified V.B. (Equation 4), and modified Okkes (Equation 5) correlations (see table).
Due to the difficulty in functional form of V.B. and Okkes correlations, there are no significant improvements in dewpoint predictions even by using the modified versions of these two correlations, as shown in the table. According to this table, the root mean square deviation of the proposed correlation with respect to the experimental data is about 0.73, much smaller than those of the other correlations.
The predicted sulfuric-acid dewpoints according to Equation 3 are also compared with all available experimental data in Fig 5 at a relative deviation of 1%. As shown there, the proposed correlation can be used for accurate prediction of the flue-gas sulfuric-acid dewpoint temperature and evaluating the corrosion risk in process equipments and heat recovery systems.
Acknowledgment
The author thanks the National Iranian Gas Co. for its cooperation during the preparation of the manuscript.
References
- ZareNezhad, B., "Consider advanced catalysts for Claus units," Hydrocarbon Processing, October 2008, pp. 109-116.
- Lins, V.F.C., and Guimaraes, E.M., "Failure of a heat exchanger generated by an excess of SO2 and H2S in the Sulfur Recovery Unit of a petroleum refinery," Journal of Loss Prevention in the Process Industries, Vol. 20 (2007), pp. 91-97.
- Roberge, P.R., Handbook of Corrosion Engineering, New York: McGraw Hill, 1999.
- Blanco, J.M., and Pena, F., "Increase in the boiler's performance in terms of the acid dewpoint temperature," Applied Thermal Engineering, Vol. 28 (2008), No. 7, pp. 777-784.
- Verhoff, F.H., and Banchero, J.T., "Predicting dewpoints of flue gases," Chemical Engineering Progress, Vol. 70 (1974), pp. 71-72.
- Okkes, A.G., "Get acid dewpoint of flue gas," Hydrocarbon Processing, Vol. 7 (1987), pp. 53-55.
- Nippon Oil Co. Ltd., technical document, Tokyo, 2008.
- De Weijer, A., and Huijbregts, W.M.M., "Severe corrosion in a waste incinerator plant due to flue gas and steam leakage," Anti-Corrosion Methods and Materials, Vol. 50 (2003), No. 5, pp. 334-340.
The author
